Fix async generator type mismatch in Docker Client streaming

- Fixed single_result_generator to properly handle async generators from deep crawl strategies
- Added proper __aiter__ checking to distinguish between CrawlResult and async generators
- Await and yield individual results from nested async generators
- Streaming functionality now works correctly for all patterns (SDK, Direct API, Docker Client)
- All 22 comprehensive tests passing with 100% success rate
- Live streaming test confirmed working end-to-end

.gitignore

@@ -1,6 +1,11 @@
 # Scripts folder (private tools)
 .scripts/
 
+# Local development CLI (private)
+local_dev.py
+dev
+DEV_CLI_README.md
+
 # Byte-compiled / optimized / DLL files
 __pycache__/
 *.py[cod]
@@ -270,4 +275,7 @@ docs/**/data
 .codecat/
 
 docs/apps/linkdin/debug*/
-docs/apps/linkdin/samples/insights/*
+docs/apps/linkdin/samples/insights/*
+
+# Production checklist (local, not for version control)
+PRODUCTION_CHECKLIST.md

README.md

@@ -27,11 +27,9 @@
 
 Crawl4AI turns the web into clean, LLM ready Markdown for RAG, agents, and data pipelines. Fast, controllable, battle tested by a 50k+ star community.
 
-[✨ Check out latest update v0.7.4](#-recent-updates)
+[✨ Check out latest update v0.7.3](#-recent-updates)
 
-✨ New in v0.7.4: Revolutionary LLM Table Extraction with intelligent chunking, enhanced concurrency fixes, memory management refactor, and critical stability improvements. [Release notes →](https://github.com/unclecode/crawl4ai/blob/main/docs/blog/release-v0.7.4.md)
-
-✨ Recent v0.7.3: Undetected Browser Support, Multi-URL Configurations, Memory Monitoring, Enhanced Table Extraction, GitHub Sponsors. [Release notes →](https://github.com/unclecode/crawl4ai/blob/main/docs/blog/release-v0.7.3.md)
+✨ New in v0.7.3: Undetected Browser Support, Multi-URL Configurations, Memory Monitoring, Enhanced Table Extraction, GitHub Sponsors. [Release notes →](https://github.com/unclecode/crawl4ai/blob/main/docs/blog/release-v0.7.3.md)
 
 <details>
   <summary>🤓 <strong>My Personal Story</strong></summary>
@@ -544,40 +542,6 @@ async def test_news_crawl():
 
 ## ✨ Recent Updates
 
-<details>
-<summary><strong>Version 0.7.4 Release Highlights - The Intelligent Table Extraction & Performance Update</strong></summary>
-
-- **🚀 LLMTableExtraction**: Revolutionary table extraction with intelligent chunking for massive tables:
-  ```python
-  from crawl4ai import LLMTableExtraction, LLMConfig
-  
-  # Configure intelligent table extraction
-  table_strategy = LLMTableExtraction(
-      llm_config=LLMConfig(provider="openai/gpt-4.1-mini"),
-      enable_chunking=True,           # Handle massive tables
-      chunk_token_threshold=5000,     # Smart chunking threshold
-      overlap_threshold=100,          # Maintain context between chunks
-      extraction_type="structured"    # Get structured data output
-  )
-  
-  config = CrawlerRunConfig(table_extraction_strategy=table_strategy)
-  result = await crawler.arun("https://complex-tables-site.com", config=config)
-  
-  # Tables are automatically chunked, processed, and merged
-  for table in result.tables:
-      print(f"Extracted table: {len(table['data'])} rows")
-  ```
-
-- **⚡ Dispatcher Bug Fix**: Fixed sequential processing bottleneck in arun_many for fast-completing tasks
-- **🧹 Memory Management Refactor**: Consolidated memory utilities into main utils module for cleaner architecture
-- **🔧 Browser Manager Fixes**: Resolved race conditions in concurrent page creation with thread-safe locking
-- **🔗 Advanced URL Processing**: Better handling of raw:// URLs and base tag link resolution
-- **🛡️ Enhanced Proxy Support**: Flexible proxy configuration supporting both dict and string formats
-
-[Full v0.7.4 Release Notes →](https://github.com/unclecode/crawl4ai/blob/main/docs/blog/release-v0.7.4.md)
-
-</details>
-
 <details>
 <summary><strong>Version 0.7.3 Release Highlights - The Multi-Config Intelligence Update</strong></summary>
 

crawl4ai/__init__.py

@@ -29,12 +29,6 @@ from .extraction_strategy import (
 )
 from .chunking_strategy import ChunkingStrategy, RegexChunking
 from .markdown_generation_strategy import DefaultMarkdownGenerator
-from .table_extraction import (
-    TableExtractionStrategy,
-    DefaultTableExtraction,
-    NoTableExtraction,
-    LLMTableExtraction,
-)
 from .content_filter_strategy import (
     PruningContentFilter,
     BM25ContentFilter,
@@ -162,9 +156,6 @@ __all__ = [
     "ChunkingStrategy",
     "RegexChunking",
     "DefaultMarkdownGenerator",
-    "TableExtractionStrategy",
-    "DefaultTableExtraction",
-    "NoTableExtraction",
     "RelevantContentFilter",
     "PruningContentFilter",
     "BM25ContentFilter",

crawl4ai/__version__.py

@@ -1,7 +1,7 @@
 # crawl4ai/__version__.py
 
 # This is the version that will be used for stable releases
-__version__ = "0.7.4"
+__version__ = "0.7.3"
 
 # For nightly builds, this gets set during build process
 __nightly_version__ = None

crawl4ai/async_configs.py

@@ -20,7 +20,6 @@ from .chunking_strategy import ChunkingStrategy, RegexChunking
 from .markdown_generation_strategy import MarkdownGenerationStrategy, DefaultMarkdownGenerator
 from .content_scraping_strategy import ContentScrapingStrategy, LXMLWebScrapingStrategy
 from .deep_crawling import DeepCrawlStrategy
-from .table_extraction import TableExtractionStrategy, DefaultTableExtraction
 
 from .cache_context import CacheMode
 from .proxy_strategy import ProxyRotationStrategy
@@ -983,8 +982,6 @@ class CrawlerRunConfig():
                                          Default: False.
         table_score_threshold (int): Minimum score threshold for processing a table.
                                      Default: 7.
-        table_extraction (TableExtractionStrategy): Strategy to use for table extraction.
-                                     Default: DefaultTableExtraction with table_score_threshold.
 
         # Virtual Scroll Parameters
         virtual_scroll_config (VirtualScrollConfig or dict or None): Configuration for handling virtual scroll containers.
@@ -1111,7 +1108,6 @@ class CrawlerRunConfig():
         image_description_min_word_threshold: int = IMAGE_DESCRIPTION_MIN_WORD_THRESHOLD,
         image_score_threshold: int = IMAGE_SCORE_THRESHOLD,
         table_score_threshold: int = 7,
-        table_extraction: TableExtractionStrategy = None,
         exclude_external_images: bool = False,
         exclude_all_images: bool = False,
         # Link and Domain Handling Parameters
@@ -1228,12 +1224,6 @@ class CrawlerRunConfig():
         self.exclude_external_images = exclude_external_images
         self.exclude_all_images = exclude_all_images
         self.table_score_threshold = table_score_threshold
-        
-        # Table extraction strategy (default to DefaultTableExtraction if not specified)
-        if table_extraction is None:
-            self.table_extraction = DefaultTableExtraction(table_score_threshold=table_score_threshold)
-        else:
-            self.table_extraction = table_extraction
 
         # Link and Domain Handling Parameters
         self.exclude_social_media_domains = (
@@ -1505,7 +1495,6 @@ class CrawlerRunConfig():
                 "image_score_threshold", IMAGE_SCORE_THRESHOLD
             ),
             table_score_threshold=kwargs.get("table_score_threshold", 7),
-            table_extraction=kwargs.get("table_extraction", None),
             exclude_all_images=kwargs.get("exclude_all_images", False),
             exclude_external_images=kwargs.get("exclude_external_images", False),
             # Link and Domain Handling Parameters
@@ -1614,7 +1603,6 @@ class CrawlerRunConfig():
             "image_description_min_word_threshold": self.image_description_min_word_threshold,
             "image_score_threshold": self.image_score_threshold,
             "table_score_threshold": self.table_score_threshold,
-            "table_extraction": self.table_extraction,
             "exclude_all_images": self.exclude_all_images,
             "exclude_external_images": self.exclude_external_images,
             "exclude_social_media_domains": self.exclude_social_media_domains,

crawl4ai/async_crawler_strategy.py

@@ -2129,265 +2129,3 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
             return True  # Default to scrolling if check fails
 
 
-####################################################################################################
-# HTTP Crawler Strategy
-####################################################################################################
-
-class HTTPCrawlerError(Exception):
-    """Base error class for HTTP crawler specific exceptions"""
-    pass
-
-
-class ConnectionTimeoutError(HTTPCrawlerError):
-    """Raised when connection timeout occurs"""
-    pass
-
-
-class HTTPStatusError(HTTPCrawlerError):
-    """Raised for unexpected status codes"""
-    def __init__(self, status_code: int, message: str):
-        self.status_code = status_code
-        super().__init__(f"HTTP {status_code}: {message}")
-
-
-class AsyncHTTPCrawlerStrategy(AsyncCrawlerStrategy):
-    """
-    Fast, lightweight HTTP-only crawler strategy optimized for memory efficiency.
-    """
-    
-    __slots__ = ('logger', 'max_connections', 'dns_cache_ttl', 'chunk_size', '_session', 'hooks', 'browser_config')
-
-    DEFAULT_TIMEOUT: Final[int] = 30
-    DEFAULT_CHUNK_SIZE: Final[int] = 64 * 1024  
-    DEFAULT_MAX_CONNECTIONS: Final[int] = min(32, (os.cpu_count() or 1) * 4)
-    DEFAULT_DNS_CACHE_TTL: Final[int] = 300
-    VALID_SCHEMES: Final = frozenset({'http', 'https', 'file', 'raw'})
-
-    _BASE_HEADERS: Final = MappingProxyType({
-        'Accept': 'text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8',
-        'Accept-Language': 'en-US,en;q=0.5',
-        'Accept-Encoding': 'gzip, deflate, br',
-        'Connection': 'keep-alive',
-        'Upgrade-Insecure-Requests': '1',
-        'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36'
-    })
-    
-    def __init__(
-        self, 
-        browser_config: Optional[HTTPCrawlerConfig] = None,
-        logger: Optional[AsyncLogger] = None,
-        max_connections: int = DEFAULT_MAX_CONNECTIONS,
-        dns_cache_ttl: int = DEFAULT_DNS_CACHE_TTL,
-        chunk_size: int = DEFAULT_CHUNK_SIZE
-    ):
-        """Initialize the HTTP crawler with config"""
-        self.browser_config = browser_config or HTTPCrawlerConfig()
-        self.logger = logger
-        self.max_connections = max_connections
-        self.dns_cache_ttl = dns_cache_ttl
-        self.chunk_size = chunk_size
-        self._session: Optional[aiohttp.ClientSession] = None
-        
-        self.hooks = {
-            k: partial(self._execute_hook, k) 
-            for k in ('before_request', 'after_request', 'on_error')
-        }
-
-        # Set default hooks
-        self.set_hook('before_request', lambda *args, **kwargs: None)
-        self.set_hook('after_request', lambda *args, **kwargs: None)
-        self.set_hook('on_error', lambda *args, **kwargs: None)
-                      
-
-    async def __aenter__(self) -> AsyncHTTPCrawlerStrategy:
-        await self.start()
-        return self
-        
-    async def __aexit__(self, exc_type, exc_val, exc_tb) -> None:
-        await self.close()
-
-    @contextlib.asynccontextmanager
-    async def _session_context(self):
-        try:
-            if not self._session:
-                await self.start()
-            yield self._session
-        finally:
-            pass
-
-    def set_hook(self, hook_type: str, hook_func: Callable) -> None:
-        if hook_type in self.hooks:
-            self.hooks[hook_type] = partial(self._execute_hook, hook_type, hook_func)
-        else:
-            raise ValueError(f"Invalid hook type: {hook_type}")
-
-    async def _execute_hook(
-        self, 
-        hook_type: str, 
-        hook_func: Callable,
-        *args: Any, 
-        **kwargs: Any
-    ) -> Any:
-        if asyncio.iscoroutinefunction(hook_func):
-            return await hook_func(*args, **kwargs)
-        return hook_func(*args, **kwargs)
-
-    async def start(self) -> None:
-        if not self._session:
-            connector = aiohttp.TCPConnector(
-                limit=self.max_connections,
-                ttl_dns_cache=self.dns_cache_ttl,
-                use_dns_cache=True,
-                force_close=False
-            )
-            self._session = aiohttp.ClientSession(
-                headers=dict(self._BASE_HEADERS),
-                connector=connector,
-                timeout=ClientTimeout(total=self.DEFAULT_TIMEOUT)
-            )
-
-    async def close(self) -> None:
-        if self._session and not self._session.closed:
-            try:
-                await asyncio.wait_for(self._session.close(), timeout=5.0)
-            except asyncio.TimeoutError:
-                if self.logger:
-                    self.logger.warning(
-                        message="Session cleanup timed out",
-                        tag="CLEANUP"
-                    )
-            finally:
-                self._session = None
-
-    async def _stream_file(self, path: str) -> AsyncGenerator[memoryview, None]:
-        async with aiofiles.open(path, mode='rb') as f:
-            while chunk := await f.read(self.chunk_size):
-                yield memoryview(chunk)
-
-    async def _handle_file(self, path: str) -> AsyncCrawlResponse:
-        if not os.path.exists(path):
-            raise FileNotFoundError(f"Local file not found: {path}")
-            
-        chunks = []
-        async for chunk in self._stream_file(path):
-            chunks.append(chunk.tobytes().decode('utf-8', errors='replace'))
-            
-        return AsyncCrawlResponse(
-            html=''.join(chunks),
-            response_headers={},
-            status_code=200
-        )
-
-    async def _handle_raw(self, content: str) -> AsyncCrawlResponse:
-        return AsyncCrawlResponse(
-            html=content,
-            response_headers={},
-            status_code=200
-        )
-
-
-    async def _handle_http(
-        self, 
-        url: str, 
-        config: CrawlerRunConfig
-    ) -> AsyncCrawlResponse:
-        async with self._session_context() as session:
-            timeout = ClientTimeout(
-                total=config.page_timeout or self.DEFAULT_TIMEOUT,
-                connect=10,
-                sock_read=30
-            )
-            
-            headers = dict(self._BASE_HEADERS)
-            if self.browser_config.headers:
-                headers.update(self.browser_config.headers)
-
-            request_kwargs = {
-                'timeout': timeout,
-                'allow_redirects': self.browser_config.follow_redirects,
-                'ssl': self.browser_config.verify_ssl,
-                'headers': headers
-            }
-
-            if self.browser_config.method == "POST":
-                if self.browser_config.data:
-                    request_kwargs['data'] = self.browser_config.data
-                if self.browser_config.json:
-                    request_kwargs['json'] = self.browser_config.json
-
-            await self.hooks['before_request'](url, request_kwargs)
-
-            try:
-                async with session.request(self.browser_config.method, url, **request_kwargs) as response:
-                    content = memoryview(await response.read())
-                    
-                    if not (200 <= response.status < 300):
-                        raise HTTPStatusError(
-                            response.status,
-                            f"Unexpected status code for {url}"
-                        )
-                    
-                    encoding = response.charset
-                    if not encoding:
-                        encoding = chardet.detect(content.tobytes())['encoding'] or 'utf-8'                    
-                    
-                    result = AsyncCrawlResponse(
-                        html=content.tobytes().decode(encoding, errors='replace'),
-                        response_headers=dict(response.headers),
-                        status_code=response.status,
-                        redirected_url=str(response.url)
-                    )
-                    
-                    await self.hooks['after_request'](result)
-                    return result
-
-            except aiohttp.ServerTimeoutError as e:
-                await self.hooks['on_error'](e)
-                raise ConnectionTimeoutError(f"Request timed out: {str(e)}")
-                
-            except aiohttp.ClientConnectorError as e:
-                await self.hooks['on_error'](e)
-                raise ConnectionError(f"Connection failed: {str(e)}")
-                
-            except aiohttp.ClientError as e:
-                await self.hooks['on_error'](e)
-                raise HTTPCrawlerError(f"HTTP client error: {str(e)}")
-            
-            except asyncio.exceptions.TimeoutError as e:
-                await self.hooks['on_error'](e)
-                raise ConnectionTimeoutError(f"Request timed out: {str(e)}")
-            
-            except Exception as e:
-                await self.hooks['on_error'](e)
-                raise HTTPCrawlerError(f"HTTP request failed: {str(e)}")
-
-    async def crawl(
-        self, 
-        url: str, 
-        config: Optional[CrawlerRunConfig] = None, 
-        **kwargs
-    ) -> AsyncCrawlResponse:
-        config = config or CrawlerRunConfig.from_kwargs(kwargs)
-        
-        parsed = urlparse(url)
-        scheme = parsed.scheme.rstrip('/')
-        
-        if scheme not in self.VALID_SCHEMES:
-            raise ValueError(f"Unsupported URL scheme: {scheme}")
-            
-        try:
-            if scheme == 'file':
-                return await self._handle_file(parsed.path)
-            elif scheme == 'raw':
-                return await self._handle_raw(parsed.path)
-            else:  # http or https
-                return await self._handle_http(url, config)
-                
-        except Exception as e:
-            if self.logger:
-                self.logger.error(
-                    message="Crawl failed: {error}",
-                    tag="CRAWL",
-                    params={"error": str(e), "url": url}
-                )
-            raise

crawl4ai/async_dispatcher.py

@@ -22,7 +22,7 @@ from urllib.parse import urlparse
 import random
 from abc import ABC, abstractmethod
 
-from .utils import get_true_memory_usage_percent
+from .memory_utils import get_true_memory_usage_percent
 
 
 class RateLimiter:

crawl4ai/content_scraping_strategy.py

@@ -586,6 +586,117 @@ class LXMLWebScrapingStrategy(ContentScrapingStrategy):
 
         return root
 
+    def is_data_table(self, table: etree.Element, **kwargs) -> bool:
+        score = 0
+        # Check for thead and tbody
+        has_thead = len(table.xpath(".//thead")) > 0
+        has_tbody = len(table.xpath(".//tbody")) > 0
+        if has_thead:
+            score += 2
+        if has_tbody:
+            score += 1
+
+        # Check for th elements
+        th_count = len(table.xpath(".//th"))
+        if th_count > 0:
+            score += 2
+            if has_thead or table.xpath(".//tr[1]/th"):
+                score += 1
+
+        # Check for nested tables
+        if len(table.xpath(".//table")) > 0:
+            score -= 3
+
+        # Role attribute check
+        role = table.get("role", "").lower()
+        if role in {"presentation", "none"}:
+            score -= 3
+
+        # Column consistency
+        rows = table.xpath(".//tr")
+        if not rows:
+            return False
+        col_counts = [len(row.xpath(".//td|.//th")) for row in rows]
+        avg_cols = sum(col_counts) / len(col_counts)
+        variance = sum((c - avg_cols)**2 for c in col_counts) / len(col_counts)
+        if variance < 1:
+            score += 2
+
+        # Caption and summary
+        if table.xpath(".//caption"):
+            score += 2
+        if table.get("summary"):
+            score += 1
+
+        # Text density
+        total_text = sum(len(''.join(cell.itertext()).strip()) for row in rows for cell in row.xpath(".//td|.//th"))
+        total_tags = sum(1 for _ in table.iterdescendants())
+        text_ratio = total_text / (total_tags + 1e-5)
+        if text_ratio > 20:
+            score += 3
+        elif text_ratio > 10:
+            score += 2
+
+        # Data attributes
+        data_attrs = sum(1 for attr in table.attrib if attr.startswith('data-'))
+        score += data_attrs * 0.5
+
+        # Size check
+        if avg_cols >= 2 and len(rows) >= 2:
+            score += 2
+
+        threshold = kwargs.get("table_score_threshold", 7)
+        return score >= threshold
+
+    def extract_table_data(self, table: etree.Element) -> dict:
+        caption = table.xpath(".//caption/text()")
+        caption = caption[0].strip() if caption else ""
+        summary = table.get("summary", "").strip()
+
+        # Extract headers with colspan handling
+        headers = []
+        thead_rows = table.xpath(".//thead/tr")
+        if thead_rows:
+            header_cells = thead_rows[0].xpath(".//th")
+            for cell in header_cells:
+                text = cell.text_content().strip()
+                colspan = int(cell.get("colspan", 1))
+                headers.extend([text] * colspan)
+        else:
+            first_row = table.xpath(".//tr[1]")
+            if first_row:
+                for cell in first_row[0].xpath(".//th|.//td"):
+                    text = cell.text_content().strip()
+                    colspan = int(cell.get("colspan", 1))
+                    headers.extend([text] * colspan)
+
+        # Extract rows with colspan handling
+        rows = []
+        for row in table.xpath(".//tr[not(ancestor::thead)]"):
+            row_data = []
+            for cell in row.xpath(".//td"):
+                text = cell.text_content().strip()
+                colspan = int(cell.get("colspan", 1))
+                row_data.extend([text] * colspan)
+            if row_data:
+                rows.append(row_data)
+
+        # Align rows with headers
+        max_columns = len(headers) if headers else (max(len(row) for row in rows) if rows else 0)
+        aligned_rows = []
+        for row in rows:
+            aligned = row[:max_columns] + [''] * (max_columns - len(row))
+            aligned_rows.append(aligned)
+
+        if not headers:
+            headers = [f"Column {i+1}" for i in range(max_columns)]
+
+        return {
+            "headers": headers,
+            "rows": aligned_rows,
+            "caption": caption,
+            "summary": summary,
+        }
 
     def _scrap(
         self,
@@ -728,16 +839,12 @@ class LXMLWebScrapingStrategy(ContentScrapingStrategy):
                 **kwargs,
             )
 
-            # Extract tables using the table extraction strategy if provided
             if 'table' not in excluded_tags:
-                table_extraction = kwargs.get('table_extraction')
-                if table_extraction:
-                    # Pass logger to the strategy if it doesn't have one
-                    if not table_extraction.logger:
-                        table_extraction.logger = self.logger
-                    # Extract tables using the strategy
-                    extracted_tables = table_extraction.extract_tables(body, **kwargs)
-                    media["tables"].extend(extracted_tables)
+                tables = body.xpath(".//table")
+                for table in tables:
+                    if self.is_data_table(table, **kwargs):
+                        table_data = self.extract_table_data(table)
+                        media["tables"].append(table_data)
 
             # Handle only_text option
             if kwargs.get("only_text", False):

crawl4ai/deep_crawling/bfs_strategy.py

@@ -38,7 +38,14 @@ class BFSDeepCrawlStrategy(DeepCrawlStrategy):
         self.include_external = include_external
         self.score_threshold = score_threshold
         self.max_pages = max_pages
-        self.logger = logger or logging.getLogger(__name__)
+        # Type check for logger
+        if isinstance(logger, dict):
+            logging.getLogger(__name__).warning(
+                "BFSDeepCrawlStrategy received a dict as logger; falling back to default logger."
+            )
+            self.logger = logging.getLogger(__name__)
+        else:
+            self.logger = logger or logging.getLogger(__name__)
         self.stats = TraversalStats(start_time=datetime.now())
         self._cancel_event = asyncio.Event()
         self._pages_crawled = 0

crawl4ai/docker_client.py

@@ -30,7 +30,7 @@ class Crawl4aiDockerClient:
     def __init__(
         self,
         base_url: str = "http://localhost:8000",
-        timeout: float = 30.0,
+        timeout: float = 600.0,  # Increased to 10 minutes for crawling operations
         verify_ssl: bool = True,
         verbose: bool = True,
         log_file: Optional[str] = None
@@ -113,21 +113,12 @@ class Crawl4aiDockerClient:
         self.logger.info(f"Crawling {len(urls)} URLs {'(streaming)' if is_streaming else ''}", tag="CRAWL")
         
         if is_streaming:
-            async def stream_results() -> AsyncGenerator[CrawlResult, None]:
-                async with self._http_client.stream("POST", f"{self.base_url}/crawl/stream", json=data) as response:
-                    response.raise_for_status()
-                    async for line in response.aiter_lines():
-                        if line.strip():
-                            result = json.loads(line)
-                            if "error" in result:
-                                self.logger.error_status(url=result.get("url", "unknown"), error=result["error"])
-                                continue
-                            self.logger.url_status(url=result.get("url", "unknown"), success=True, timing=result.get("timing", 0.0))
-                            if result.get("status") == "completed":
-                                continue
-                            else:
-                                yield CrawlResult(**result)
-            return stream_results()
+            # For streaming, we need to return the async generator properly
+            # The caller should be able to do: async for result in await client.crawl(...)
+            async def streaming_wrapper():
+                async for result in self._stream_crawl_results(data):
+                    yield result
+            return streaming_wrapper()
         
         response = await self._request("POST", "/crawl", json=data)
         result_data = response.json()
@@ -138,6 +129,35 @@ class Crawl4aiDockerClient:
         self.logger.success(f"Crawl completed with {len(results)} results", tag="CRAWL")
         return results[0] if len(results) == 1 else results
 
+    async def _stream_crawl_results(self, data: Dict[str, Any]) -> AsyncGenerator[CrawlResult, None]:
+        """Internal method to handle streaming crawl results."""
+        async with self._http_client.stream("POST", f"{self.base_url}/crawl/stream", json=data) as response:
+            response.raise_for_status()
+            async for line in response.aiter_lines():
+                if line.strip():
+                    try:
+                        result = json.loads(line)
+                        if "error" in result:
+                            self.logger.error_status(url=result.get("url", "unknown"), error=result["error"])
+                            continue
+                        
+                        # Check if this is a crawl result (has required fields)
+                        if "url" in result and "success" in result:
+                            self.logger.url_status(url=result.get("url", "unknown"), success=result.get("success", False), timing=result.get("timing", 0.0))
+                            
+                            # Create CrawlResult object properly
+                            crawl_result = CrawlResult(**result)
+                            yield crawl_result
+                        # Skip status-only messages
+                        elif result.get("status") == "completed":
+                            continue
+                    except json.JSONDecodeError as e:
+                        self.logger.error(f"Failed to parse streaming response: {e}", tag="STREAM")
+                        continue
+                    except Exception as e:
+                        self.logger.error(f"Error processing streaming result: {e}", tag="STREAM")
+                        continue
+
     async def get_schema(self) -> Dict[str, Any]:
         """Retrieve configuration schemas."""
         response = await self._request("GET", "/schema")

crawl4ai/memory_utils.py

@@ -0,0 +1,79 @@
+import psutil
+import platform
+import subprocess
+from typing import Tuple
+
+
+def get_true_available_memory_gb() -> float:
+    """Get truly available memory including inactive pages (cross-platform)"""
+    vm = psutil.virtual_memory()
+
+    if platform.system() == 'Darwin':  # macOS
+        # On macOS, we need to include inactive memory too
+        try:
+            # Use vm_stat to get accurate values
+            result = subprocess.run(['vm_stat'], capture_output=True, text=True)
+            lines = result.stdout.split('\n')
+
+            page_size = 16384  # macOS page size
+            pages = {}
+
+            for line in lines:
+                if 'Pages free:' in line:
+                    pages['free'] = int(line.split()[-1].rstrip('.'))
+                elif 'Pages inactive:' in line:
+                    pages['inactive'] = int(line.split()[-1].rstrip('.'))
+                elif 'Pages speculative:' in line:
+                    pages['speculative'] = int(line.split()[-1].rstrip('.'))
+                elif 'Pages purgeable:' in line:
+                    pages['purgeable'] = int(line.split()[-1].rstrip('.'))
+
+            # Calculate total available (free + inactive + speculative + purgeable)
+            total_available_pages = (
+                pages.get('free', 0) + 
+                pages.get('inactive', 0) + 
+                pages.get('speculative', 0) + 
+                pages.get('purgeable', 0)
+            )
+            available_gb = (total_available_pages * page_size) / (1024**3)
+
+            return available_gb
+        except:
+            # Fallback to psutil
+            return vm.available / (1024**3)
+    else:
+        # For Windows and Linux, psutil.available is accurate
+        return vm.available / (1024**3)
+
+
+def get_true_memory_usage_percent() -> float:
+    """
+    Get memory usage percentage that accounts for platform differences.
+    
+    Returns:
+        float: Memory usage percentage (0-100)
+    """
+    vm = psutil.virtual_memory()
+    total_gb = vm.total / (1024**3)
+    available_gb = get_true_available_memory_gb()
+    
+    # Calculate used percentage based on truly available memory
+    used_percent = 100.0 * (total_gb - available_gb) / total_gb
+    
+    # Ensure it's within valid range
+    return max(0.0, min(100.0, used_percent))
+
+
+def get_memory_stats() -> Tuple[float, float, float]:
+    """
+    Get comprehensive memory statistics.
+    
+    Returns:
+        Tuple[float, float, float]: (used_percent, available_gb, total_gb)
+    """
+    vm = psutil.virtual_memory()
+    total_gb = vm.total / (1024**3)
+    available_gb = get_true_available_memory_gb()
+    used_percent = get_true_memory_usage_percent()
+    
+    return used_percent, available_gb, total_gb

crawl4ai/models.py

@@ -1,4 +1,36 @@
-from pydantic import BaseModel, HttpUrl, PrivateAttr, Field
+
+"""
+Crawl4AI Models Module
+
+This module contains Pydantic models used throughout the Crawl4AI library.
+
+Key Features:
+- ORJSONModel: Base model with ORJSON serialization support
+- DeprecatedPropertiesMixin: Global system for handling deprecated properties
+- CrawlResult: Main result model with backward compatibility support
+
+Deprecated Properties System:
+The DeprecatedPropertiesMixin provides a global way to handle deprecated properties
+across all models. Instead of manually excluding deprecated properties in each
+model_dump() call, you can simply override the get_deprecated_properties() method:
+
+Example:
+    class MyModel(ORJSONModel):
+        name: str
+        old_field: Optional[str] = None
+        
+        def get_deprecated_properties(self) -> set[str]:
+            return {'old_field', 'another_deprecated_field'}
+        
+        @property
+        def old_field(self):
+            raise AttributeError("old_field is deprecated, use name instead")
+
+The system automatically excludes these properties from serialization, preventing
+property objects from appearing in JSON output.
+"""
+
+from pydantic import BaseModel, ConfigDict,HttpUrl, PrivateAttr, Field
 from typing import List, Dict, Optional, Callable, Awaitable, Union, Any
 from typing import AsyncGenerator
 from typing import Generic, TypeVar
@@ -8,7 +40,7 @@ from .ssl_certificate import SSLCertificate
 from datetime import datetime
 from datetime import timedelta
 
-
+import orjson
 ###############################
 # Dispatcher Models
 ###############################
@@ -91,7 +123,122 @@ class TokenUsage:
     completion_tokens_details: Optional[dict] = None
     prompt_tokens_details: Optional[dict] = None
 
-class UrlModel(BaseModel):
+
+def orjson_default(obj):
+    # Handle datetime (if not already handled by orjson)
+    if isinstance(obj, datetime):
+        return obj.isoformat()
+
+    # Handle property objects (convert to string or something else)
+    if isinstance(obj, property):
+        return str(obj)
+
+    # Last resort: convert to string
+    return str(obj)
+
+
+class DeprecatedPropertiesMixin:
+    """
+    Mixin to handle deprecated properties in Pydantic models.
+    
+    Classes that inherit from this mixin can define deprecated properties
+    that will be automatically excluded from serialization.
+    
+    Usage:
+    1. Override the get_deprecated_properties() method to return a set of deprecated property names
+    2. The model_dump method will automatically exclude these properties
+    
+    Example:
+        class MyModel(ORJSONModel):
+            def get_deprecated_properties(self) -> set[str]:
+                return {'old_field', 'legacy_property'}
+            
+            name: str
+            old_field: Optional[str] = None  # Field definition
+            
+            @property
+            def old_field(self):  # Property that overrides the field
+                raise AttributeError("old_field is deprecated, use name instead")
+    """
+    
+    def get_deprecated_properties(self) -> set[str]:
+        """
+        Get deprecated property names for this model.
+        Override this method in subclasses to define deprecated properties.
+        
+        Returns:
+            set[str]: Set of deprecated property names
+        """
+        return set()
+    
+    @classmethod
+    def get_all_deprecated_properties(cls) -> set[str]:
+        """
+        Get all deprecated properties from this class and all parent classes.
+        
+        Returns:
+            set[str]: Set of all deprecated property names
+        """
+        deprecated_props = set()
+        # Create an instance to call the instance method
+        try:
+            # Try to create a dummy instance to get deprecated properties
+            dummy_instance = cls.__new__(cls)
+            deprecated_props.update(dummy_instance.get_deprecated_properties())
+        except Exception:
+            # If we can't create an instance, check for class-level definitions
+            pass
+            
+        # Also check parent classes
+        for klass in cls.__mro__:
+            if hasattr(klass, 'get_deprecated_properties') and klass != DeprecatedPropertiesMixin:
+                try:
+                    dummy_instance = klass.__new__(klass)
+                    deprecated_props.update(dummy_instance.get_deprecated_properties())
+                except Exception:
+                    pass
+        return deprecated_props
+    
+    def model_dump(self, *args, **kwargs):
+        """
+        Override model_dump to automatically exclude deprecated properties.
+        
+        This method:
+        1. Gets the existing exclude set from kwargs
+        2. Adds all deprecated properties defined in get_deprecated_properties()
+        3. Calls the parent model_dump with the updated exclude set
+        """
+        # Get the default exclude set, or create empty set if None
+        exclude = kwargs.get('exclude', set())
+        if exclude is None:
+            exclude = set()
+        elif not isinstance(exclude, set):
+            exclude = set(exclude) if exclude else set()
+        
+        # Add deprecated properties for this instance
+        exclude.update(self.get_deprecated_properties())
+        kwargs['exclude'] = exclude
+        
+        return super().model_dump(*args, **kwargs)
+
+
+class ORJSONModel(DeprecatedPropertiesMixin, BaseModel):
+    model_config = ConfigDict(
+        ser_json_timedelta="iso8601",  # Optional: format timedelta
+        ser_json_bytes="utf8",         # Optional: bytes → UTF-8 string
+    )
+    
+    def model_dump_json(self, **kwargs) -> bytes:
+        """Custom JSON serialization using orjson"""
+        return orjson.dumps(self.model_dump(**kwargs), default=orjson_default)
+    
+    @classmethod
+    def model_validate_json(cls, json_data: Union[str, bytes], **kwargs):
+        """Custom JSON deserialization using orjson"""
+        if isinstance(json_data, str):
+            json_data = json_data.encode()
+        return cls.model_validate(orjson.loads(json_data), **kwargs)
+class UrlModel(ORJSONModel):
     url: HttpUrl
     forced: bool = False
 
@@ -108,7 +255,7 @@ class TraversalStats:
     total_depth_reached: int = 0
     current_depth: int = 0
 
-class DispatchResult(BaseModel):
+class DispatchResult(ORJSONModel):
     task_id: str
     memory_usage: float
     peak_memory: float
@@ -116,7 +263,7 @@ class DispatchResult(BaseModel):
     end_time: Union[datetime, float]
     error_message: str = ""
 
-class MarkdownGenerationResult(BaseModel):
+class MarkdownGenerationResult(ORJSONModel):
     raw_markdown: str
     markdown_with_citations: str
     references_markdown: str
@@ -126,7 +273,7 @@ class MarkdownGenerationResult(BaseModel):
     def __str__(self):
         return self.raw_markdown
     
-class CrawlResult(BaseModel):
+class CrawlResult(ORJSONModel):
     url: str
     html: str
     fit_html: Optional[str] = None
@@ -156,6 +303,10 @@ class CrawlResult(BaseModel):
     class Config:
         arbitrary_types_allowed = True
 
+    def get_deprecated_properties(self) -> set[str]:
+        """Define deprecated properties that should be excluded from serialization."""
+        return {'fit_html', 'fit_markdown', 'markdown_v2'}
+
 # NOTE: The StringCompatibleMarkdown class, custom __init__ method, property getters/setters,
 # and model_dump override all exist to support a smooth transition from markdown as a string
 # to markdown as a MarkdownGenerationResult object, while maintaining backward compatibility.
@@ -245,14 +396,16 @@ class CrawlResult(BaseModel):
         1. PrivateAttr fields are excluded from serialization by default
         2. We need to maintain backward compatibility by including the 'markdown' field
            in the serialized output
-        3. We're transitioning from 'markdown_v2' to enhancing 'markdown' to hold
-           the same type of data
+        3. Uses the DeprecatedPropertiesMixin to automatically exclude deprecated properties
         
         Future developers: This method ensures that the markdown content is properly
-        serialized despite being stored in a private attribute. If the serialization
-        requirements change, this is where you would update the logic.
+        serialized despite being stored in a private attribute. The deprecated properties
+        are automatically handled by the mixin.
         """
+        # Use the parent class method which handles deprecated properties automatically
         result = super().model_dump(*args, **kwargs)
+        
+        # Add the markdown content if it exists
         if self._markdown is not None:
             result["markdown"] = self._markdown.model_dump() 
         return result
@@ -307,7 +460,7 @@ RunManyReturn = Union[
 # 1. Replace the private attribute and property with a standard field
 # 2. Update any serialization logic that might depend on the current behavior
 
-class AsyncCrawlResponse(BaseModel):
+class AsyncCrawlResponse(ORJSONModel):
     html: str
     response_headers: Dict[str, str]
     js_execution_result: Optional[Dict[str, Any]] = None
@@ -328,7 +481,7 @@ class AsyncCrawlResponse(BaseModel):
 ###############################
 # Scraping Models
 ###############################
-class MediaItem(BaseModel):
+class MediaItem(ORJSONModel):
     src: Optional[str] = ""
     data: Optional[str] = ""
     alt: Optional[str] = ""
@@ -340,7 +493,7 @@ class MediaItem(BaseModel):
     width: Optional[int] = None
 
 
-class Link(BaseModel):
+class Link(ORJSONModel):
     href: Optional[str] = ""
     text: Optional[str] = ""
     title: Optional[str] = ""
@@ -353,7 +506,7 @@ class Link(BaseModel):
     total_score: Optional[float] = None  # Combined score from intrinsic and contextual scores
 
 
-class Media(BaseModel):
+class Media(ORJSONModel):
     images: List[MediaItem] = []
     videos: List[
         MediaItem
@@ -364,12 +517,12 @@ class Media(BaseModel):
     tables: List[Dict] = []  # Table data extracted from HTML tables
 
 
-class Links(BaseModel):
+class Links(ORJSONModel):
     internal: List[Link] = []
     external: List[Link] = []
 
 
-class ScrapingResult(BaseModel):
+class ScrapingResult(ORJSONModel):
     cleaned_html: str
     success: bool
     media: Media = Media()

crawl4ai/utils.py

@@ -16,7 +16,7 @@ from .config import MIN_WORD_THRESHOLD, IMAGE_DESCRIPTION_MIN_WORD_THRESHOLD, IM
 import httpx
 from socket import gaierror
 from pathlib import Path
-from typing import Dict, Any, List, Optional, Callable, Generator, Tuple, Iterable
+from typing import Dict, Any, List, Optional, Callable
 from urllib.parse import urljoin
 import requests
 from requests.exceptions import InvalidSchema
@@ -40,7 +40,8 @@ from typing import Sequence
 
 from itertools import chain
 from collections import deque
-import psutil
+from typing import  Generator, Iterable
+
 import numpy as np
 
 from urllib.parse import (
@@ -3413,79 +3414,3 @@ def cosine_distance(vec1: np.ndarray, vec2: np.ndarray) -> float:
     """Calculate cosine distance (1 - similarity) between two vectors"""
     return 1 - cosine_similarity(vec1, vec2)
 
-
-# Memory utilities
-
-def get_true_available_memory_gb() -> float:
-    """Get truly available memory including inactive pages (cross-platform)"""
-    vm = psutil.virtual_memory()
-
-    if platform.system() == 'Darwin':  # macOS
-        # On macOS, we need to include inactive memory too
-        try:
-            # Use vm_stat to get accurate values
-            result = subprocess.run(['vm_stat'], capture_output=True, text=True)
-            lines = result.stdout.split('\n')
-
-            page_size = 16384  # macOS page size
-            pages = {}
-
-            for line in lines:
-                if 'Pages free:' in line:
-                    pages['free'] = int(line.split()[-1].rstrip('.'))
-                elif 'Pages inactive:' in line:
-                    pages['inactive'] = int(line.split()[-1].rstrip('.'))
-                elif 'Pages speculative:' in line:
-                    pages['speculative'] = int(line.split()[-1].rstrip('.'))
-                elif 'Pages purgeable:' in line:
-                    pages['purgeable'] = int(line.split()[-1].rstrip('.'))
-
-            # Calculate total available (free + inactive + speculative + purgeable)
-            total_available_pages = (
-                pages.get('free', 0) + 
-                pages.get('inactive', 0) + 
-                pages.get('speculative', 0) + 
-                pages.get('purgeable', 0)
-            )
-            available_gb = (total_available_pages * page_size) / (1024**3)
-
-            return available_gb
-        except:
-            # Fallback to psutil
-            return vm.available / (1024**3)
-    else:
-        # For Windows and Linux, psutil.available is accurate
-        return vm.available / (1024**3)
-
-
-def get_true_memory_usage_percent() -> float:
-    """
-    Get memory usage percentage that accounts for platform differences.
-    
-    Returns:
-        float: Memory usage percentage (0-100)
-    """
-    vm = psutil.virtual_memory()
-    total_gb = vm.total / (1024**3)
-    available_gb = get_true_available_memory_gb()
-    
-    # Calculate used percentage based on truly available memory
-    used_percent = 100.0 * (total_gb - available_gb) / total_gb
-    
-    # Ensure it's within valid range
-    return max(0.0, min(100.0, used_percent))
-
-
-def get_memory_stats() -> Tuple[float, float, float]:
-    """
-    Get comprehensive memory statistics.
-    
-    Returns:
-        Tuple[float, float, float]: (used_percent, available_gb, total_gb)
-    """
-    vm = psutil.virtual_memory()
-    total_gb = vm.total / (1024**3)
-    available_gb = get_true_available_memory_gb()
-    used_percent = get_true_memory_usage_percent()
-    
-    return used_percent, available_gb, total_gb

deploy/docker/api.py

@@ -1,5 +1,6 @@
 import os
 import json
+import orjson
 import asyncio
 from typing import List, Tuple, Dict
 from functools import partial
@@ -384,27 +385,60 @@ def create_task_response(task: dict, task_id: str, base_url: str) -> dict:
 
 async def stream_results(crawler: AsyncWebCrawler, results_gen: AsyncGenerator) -> AsyncGenerator[bytes, None]:
     """Stream results with heartbeats and completion markers."""
-    import json
-    from utils import datetime_handler
+    import orjson
+    from datetime import datetime
+    import inspect
+    
+    def orjson_default(obj):
+        # Handle datetime (if not already handled by orjson)
+        if isinstance(obj, datetime):
+            return obj.isoformat()
+        # Handle property objects (convert to string or something else)
+        if isinstance(obj, property):
+            return str(obj)
+        # Last resort: convert to string
+        return str(obj)
 
     try:
-        async for result in results_gen:
-            try:
-                server_memory_mb = _get_memory_mb()
-                result_dict = result.model_dump()
-                result_dict['server_memory_mb'] = server_memory_mb
-                # If PDF exists, encode it to base64
-                if result_dict.get('pdf') is not None:
-                    result_dict['pdf'] = b64encode(result_dict['pdf']).decode('utf-8')
-                logger.info(f"Streaming result for {result_dict.get('url', 'unknown')}")
-                data = json.dumps(result_dict, default=datetime_handler) + "\n"
-                yield data.encode('utf-8')
-            except Exception as e:
-                logger.error(f"Serialization error: {e}")
-                error_response = {"error": str(e), "url": getattr(result, 'url', 'unknown')}
-                yield (json.dumps(error_response) + "\n").encode('utf-8')
+        logger.info(f"Starting streaming with results_gen type: {type(results_gen)}")
+        logger.info(f"Is results_gen async generator: {inspect.isasyncgen(results_gen)}")
+        
+        # Check if results_gen is actually an async generator vs another type
+        if inspect.isasyncgen(results_gen):
+            logger.info("Processing as async generator")
+            async for result in results_gen:
+                try:
+                    logger.info(f"Processing streaming result of type: {type(result)}")
+                    
+                    # Check if this result is actually a CrawlResult
+                    if hasattr(result, 'model_dump_json'):
+                        server_memory_mb = _get_memory_mb()
+                        result_json = result.model_dump_json()
+                        result_dict = orjson.loads(result_json)
+                        result_dict['server_memory_mb'] = server_memory_mb
+                        
+                        if result_dict.get('pdf') is not None:
+                            result_dict['pdf'] = b64encode(result_dict['pdf']).decode('utf-8')
+                        
+                        logger.info(f"Streaming result for {result_dict.get('url', 'unknown')}")
+                        data = orjson.dumps(result_dict, default=orjson_default).decode('utf-8') + "\n"
+                        yield data.encode('utf-8')
+                    else:
+                        logger.error(f"Result doesn't have model_dump_json method: {type(result)}")
+                        error_response = {"error": f"Invalid result type: {type(result)}", "url": "unknown"}
+                        yield (orjson.dumps(error_response).decode('utf-8') + "\n").encode('utf-8')
+                        
+                except Exception as e:
+                    logger.error(f"Serialization error: {e}")
+                    logger.error(f"Result type was: {type(result)}")
+                    error_response = {"error": str(e), "url": getattr(result, 'url', 'unknown')}
+                    yield (orjson.dumps(error_response).decode('utf-8') + "\n").encode('utf-8')
+        else:
+            logger.error(f"results_gen is not an async generator: {type(results_gen)}")
+            error_response = {"error": f"Invalid results_gen type: {type(results_gen)}"}
+            yield (orjson.dumps(error_response).decode('utf-8') + "\n").encode('utf-8')
 
-        yield json.dumps({"status": "completed"}).encode('utf-8')
+        yield orjson.dumps({"status": "completed"}).decode('utf-8').encode('utf-8')
         
     except asyncio.CancelledError:
         logger.warning("Client disconnected during streaming")
@@ -472,7 +506,9 @@ async def handle_crawl_request(
         # Process results to handle PDF bytes
         processed_results = []
         for result in results:
-            result_dict = result.model_dump()
+            # Use ORJSON serialization to handle property objects properly
+            result_json = result.model_dump_json()
+            result_dict = orjson.loads(result_json)
             # If PDF exists, encode it to base64
             if result_dict.get('pdf') is not None:
                 result_dict['pdf'] = b64encode(result_dict['pdf']).decode('utf-8')
@@ -522,8 +558,19 @@ async def handle_stream_crawl_request(
         browser_config.verbose = False
         crawler_config = CrawlerRunConfig.load(crawler_config)
         crawler_config.scraping_strategy = LXMLWebScrapingStrategy()
-        crawler_config.stream = True
+        # Don't force stream=True here - let the deep crawl strategy control its own streaming behavior
 
+        # Apply global base config (this was missing!)
+        base_config = config["crawler"]["base_config"]
+        for key, value in base_config.items():
+            if hasattr(crawler_config, key):
+                print(f"[DEBUG] Applying base_config: {key} = {value}")
+                setattr(crawler_config, key, value)
+
+        print(f"[DEBUG] Deep crawl strategy: {type(crawler_config.deep_crawl_strategy).__name__ if crawler_config.deep_crawl_strategy else 'None'}")
+        print(f"[DEBUG] Stream mode: {crawler_config.stream}")
+        print(f"[DEBUG] Simulate user: {getattr(crawler_config, 'simulate_user', 'Not set')}")
+        
         dispatcher = MemoryAdaptiveDispatcher(
             memory_threshold_percent=config["crawler"]["memory_threshold_percent"],
             rate_limiter=RateLimiter(
@@ -537,11 +584,58 @@ async def handle_stream_crawl_request(
         # crawler = AsyncWebCrawler(config=browser_config)
         # await crawler.start()
 
-        results_gen = await crawler.arun_many(
-            urls=urls,
-            config=crawler_config,
-            dispatcher=dispatcher
-        )
+        # Use correct method based on URL count (same as regular endpoint)
+        if len(urls) == 1:
+            # For single URL, use arun to get CrawlResult, then wrap in async generator
+            single_result_container = await crawler.arun(
+                url=urls[0],
+                config=crawler_config,
+                dispatcher=dispatcher
+            )
+            
+            async def single_result_generator():
+                # Handle CrawlResultContainer - extract the actual results
+                if hasattr(single_result_container, '_results'):
+                    # It's a CrawlResultContainer - iterate over the internal results
+                    for result in single_result_container._results:
+                        # Check if the result is an async generator (from deep crawl)
+                        if hasattr(result, '__aiter__'):
+                            async for sub_result in result:
+                                yield sub_result
+                        else:
+                            yield result
+                elif hasattr(single_result_container, '__aiter__'):
+                    # It's an async generator (from streaming deep crawl)
+                    async for result in single_result_container:
+                        yield result
+                elif hasattr(single_result_container, '__iter__') and not hasattr(single_result_container, 'url'):
+                    # It's iterable but not a CrawlResult itself
+                    for result in single_result_container:
+                        # Check if each result is an async generator
+                        if hasattr(result, '__aiter__'):
+                            async for sub_result in result:
+                                yield sub_result
+                        else:
+                            yield result
+                else:
+                    # It's a single CrawlResult
+                    yield single_result_container
+                    
+            results_gen = single_result_generator()
+        else:
+            # For multiple URLs, use arun_many
+            results_gen = await crawler.arun_many(
+                urls=urls,
+                config=crawler_config,
+                dispatcher=dispatcher
+            )
+            
+            # If results_gen is a list (e.g., from deep crawl), convert to async generator
+            if isinstance(results_gen, list):
+                async def convert_list_to_generator():
+                    for result in results_gen:
+                        yield result
+                results_gen = convert_list_to_generator()
 
         return crawler, results_gen
 

deploy/docker/server.py

@@ -7,13 +7,16 @@ Crawl4AI FastAPI entry‑point
 """
 
 # ── stdlib & 3rd‑party imports ───────────────────────────────
+from datetime import datetime
+
+import orjson
 from crawler_pool import get_crawler, close_all, janitor
 from crawl4ai import AsyncWebCrawler, BrowserConfig, CrawlerRunConfig
 from auth import create_access_token, get_token_dependency, TokenRequest
 from pydantic import BaseModel
 from typing import Optional, List, Dict
 from fastapi import Request, Depends
-from fastapi.responses import FileResponse
+from fastapi.responses import FileResponse, ORJSONResponse
 import base64
 import re
 from crawl4ai import AsyncWebCrawler, BrowserConfig, CrawlerRunConfig
@@ -32,6 +35,8 @@ from schemas import (
     JSEndpointRequest,
 )
 
+# Use the proper serialization functions from async_configs
+from crawl4ai.async_configs import to_serializable_dict
 from utils import (
     FilterType, load_config, setup_logging, verify_email_domain
 )
@@ -112,11 +117,26 @@ async def lifespan(_: FastAPI):
     app.state.janitor.cancel()
     await close_all()
 
+def orjson_default(obj):
+    # Handle datetime (if not already handled by orjson)
+    if isinstance(obj, datetime):
+        return obj.isoformat()
+
+    # Handle property objects (convert to string or something else)
+    if isinstance(obj, property):
+        return str(obj)
+
+    # Last resort: convert to string
+    return str(obj)
+
+def orjson_dumps(v, *, default):
+    return orjson.dumps(v, default=orjson_default).decode()
 # ───────────────────── FastAPI instance ──────────────────────
 app = FastAPI(
     title=config["app"]["title"],
     version=config["app"]["version"],
     lifespan=lifespan,
+    default_response_class=ORJSONResponse
 )
 
 # ── static playground ──────────────────────────────────────
@@ -435,15 +455,20 @@ async def crawl(
     """
     Crawl a list of URLs and return the results as JSON.
     """
-    if not crawl_request.urls:
-        raise HTTPException(400, "At least one URL required")
-    res = await handle_crawl_request(
-        urls=crawl_request.urls,
-        browser_config=crawl_request.browser_config,
-        crawler_config=crawl_request.crawler_config,
-        config=config,
-    )
-    return JSONResponse(res)
+    try:
+        if not crawl_request.urls:
+            raise HTTPException(400, "At least one URL required")
+        res = await handle_crawl_request(
+            urls=crawl_request.urls,
+            browser_config=crawl_request.browser_config,
+            crawler_config=crawl_request.crawler_config,
+            config=config,
+        )
+        # handle_crawl_request returns a dictionary, so we can pass it directly to ORJSONResponse
+        return ORJSONResponse(res)
+    except Exception as e:
+        print(f"Error occurred: {e}")
+        return ORJSONResponse({"error": str(e)}, status_code=500)
 
 
 @app.post("/crawl/stream")

docs/blog/release-v0.7.4.md

@@ -1,305 +0,0 @@
-# 🚀 Crawl4AI v0.7.4: The Intelligent Table Extraction & Performance Update
-
-*August 17, 2025 • 6 min read*
-
----
-
-Today I'm releasing Crawl4AI v0.7.4—the Intelligent Table Extraction & Performance Update. This release introduces revolutionary LLM-powered table extraction with intelligent chunking, significant performance improvements for concurrent crawling, enhanced browser management, and critical stability fixes that make Crawl4AI more robust for production workloads.
-
-## 🎯 What's New at a Glance
-
-- **🚀 LLMTableExtraction**: Revolutionary table extraction with intelligent chunking for massive tables
-- **⚡ Enhanced Concurrency**: True concurrency improvements for fast-completing tasks in batch operations
-- **🧹 Memory Management Refactor**: Streamlined memory utilities and better resource management
-- **🔧 Browser Manager Fixes**: Resolved race conditions in concurrent page creation
-- **⌨️ Cross-Platform Browser Profiler**: Improved keyboard handling and quit mechanisms
-- **🔗 Advanced URL Processing**: Better handling of raw URLs and base tag link resolution
-- **🛡️ Enhanced Proxy Support**: Flexible proxy configuration with dict and string formats
-- **🐳 Docker Improvements**: Better API handling and raw HTML support
-
-## 🚀 LLMTableExtraction: Revolutionary Table Processing
-
-**The Problem:** Complex tables with rowspan, colspan, nested structures, or massive datasets that traditional HTML parsing can't handle effectively. Large tables that exceed token limits crash extraction processes.
-
-**My Solution:** I developed LLMTableExtraction—an intelligent table extraction strategy that uses Large Language Models with automatic chunking to handle tables of any size and complexity.
-
-### Technical Implementation
-
-```python
-from crawl4ai import (
-    AsyncWebCrawler,
-    CrawlerRunConfig, 
-    LLMConfig,
-    LLMTableExtraction,
-    CacheMode
-)
-
-# Configure LLM for table extraction
-llm_config = LLMConfig(
-    provider="openai/gpt-4.1-mini",
-    api_token="env:OPENAI_API_KEY",
-    temperature=0.1,  # Low temperature for consistency
-    max_tokens=32000
-)
-
-# Create intelligent table extraction strategy
-table_strategy = LLMTableExtraction(
-    llm_config=llm_config,
-    verbose=True,
-    max_tries=2,
-    enable_chunking=True,           # Handle massive tables
-    chunk_token_threshold=5000,     # Smart chunking threshold
-    overlap_threshold=100,          # Maintain context between chunks
-    extraction_type="structured"    # Get structured data output
-)
-
-# Apply to crawler configuration
-config = CrawlerRunConfig(
-    table_extraction_strategy=table_strategy,
-    cache_mode=CacheMode.BYPASS
-)
-
-async with AsyncWebCrawler() as crawler:
-    # Extract complex tables with intelligence
-    result = await crawler.arun(
-        "https://en.wikipedia.org/wiki/List_of_countries_by_GDP", 
-        config=config
-    )
-    
-    # Access extracted tables directly
-    for i, table in enumerate(result.tables):
-        print(f"Table {i}: {len(table['data'])} rows × {len(table['headers'])} columns")
-        
-        # Convert to pandas DataFrame instantly
-        import pandas as pd
-        df = pd.DataFrame(table['data'], columns=table['headers'])
-        print(df.head())
-```
-
-**Intelligent Chunking for Massive Tables:**
-
-```python
-# Handle tables that exceed token limits
-large_table_strategy = LLMTableExtraction(
-    llm_config=llm_config,
-    enable_chunking=True,
-    chunk_token_threshold=3000,    # Conservative threshold
-    overlap_threshold=150,         # Preserve context
-    max_concurrent_chunks=3,       # Parallel processing
-    merge_strategy="intelligent"   # Smart chunk merging
-)
-
-# Process Wikipedia comparison tables, financial reports, etc.
-config = CrawlerRunConfig(
-    table_extraction_strategy=large_table_strategy,
-    # Target specific table containers
-    css_selector="div.wikitable, table.sortable",
-    delay_before_return_html=2.0
-)
-
-result = await crawler.arun(
-    "https://en.wikipedia.org/wiki/Comparison_of_operating_systems",
-    config=config
-)
-
-# Tables are automatically chunked, processed, and merged
-print(f"Extracted {len(result.tables)} complex tables")
-for table in result.tables:
-    print(f"Merged table: {len(table['data'])} total rows")
-```
-
-**Advanced Features:**
-
-- **Intelligent Chunking**: Automatically splits massive tables while preserving structure
-- **Context Preservation**: Overlapping chunks maintain column relationships
-- **Parallel Processing**: Concurrent chunk processing for speed
-- **Smart Merging**: Reconstructs complete tables from processed chunks
-- **Complex Structure Support**: Handles rowspan, colspan, nested tables
-- **Metadata Extraction**: Captures table context, captions, and relationships
-
-**Expected Real-World Impact:**
-- **Financial Analysis**: Extract complex earnings tables and financial statements
-- **Research & Academia**: Process large datasets from Wikipedia, research papers
-- **E-commerce**: Handle product comparison tables with complex layouts
-- **Government Data**: Extract census data, statistical tables from official sources
-- **Competitive Intelligence**: Process competitor pricing and feature tables
-
-## ⚡ Enhanced Concurrency: True Performance Gains
-
-**The Problem:** The `arun_many()` method wasn't achieving true concurrency for fast-completing tasks, leading to sequential processing bottlenecks in batch operations.
-
-**My Solution:** I implemented true concurrency improvements in the dispatcher that enable genuine parallel processing for fast-completing tasks.
-
-### Performance Optimization
-
-```python
-# Before v0.7.4: Sequential-like behavior for fast tasks
-# After v0.7.4: True concurrency
-
-async with AsyncWebCrawler() as crawler:
-    # These will now run with true concurrency
-    urls = [
-        "https://httpbin.org/delay/1",
-        "https://httpbin.org/delay/1", 
-        "https://httpbin.org/delay/1",
-        "https://httpbin.org/delay/1"
-    ]
-    
-    # Processes in truly parallel fashion
-    results = await crawler.arun_many(urls)
-    
-    # Performance improvement: ~4x faster for fast-completing tasks
-    print(f"Processed {len(results)} URLs with true concurrency")
-```
-
-**Expected Real-World Impact:**
-- **API Crawling**: 3-4x faster processing of REST endpoints and API documentation
-- **Batch URL Processing**: Significant speedup for large URL lists
-- **Monitoring Systems**: Faster health checks and status page monitoring
-- **Data Aggregation**: Improved performance for real-time data collection
-
-## 🧹 Memory Management Refactor: Cleaner Architecture
-
-**The Problem:** Memory utilities were scattered and difficult to maintain, with potential import conflicts and unclear organization.
-
-**My Solution:** I consolidated all memory-related utilities into the main `utils.py` module, creating a cleaner, more maintainable architecture.
-
-### Improved Memory Handling
-
-```python
-# All memory utilities now consolidated
-from crawl4ai.utils import get_true_memory_usage_percent, MemoryMonitor
-
-# Enhanced memory monitoring
-monitor = MemoryMonitor()
-monitor.start_monitoring()
-
-async with AsyncWebCrawler() as crawler:
-    # Memory-efficient batch processing
-    results = await crawler.arun_many(large_url_list)
-    
-    # Get accurate memory metrics
-    memory_usage = get_true_memory_usage_percent()
-    memory_report = monitor.get_report()
-    
-    print(f"Memory efficiency: {memory_report['efficiency']:.1f}%")
-    print(f"Peak usage: {memory_report['peak_mb']:.1f} MB")
-```
-
-**Expected Real-World Impact:**
-- **Production Stability**: More reliable memory tracking and management
-- **Code Maintainability**: Cleaner architecture for easier debugging
-- **Import Clarity**: Resolved potential conflicts and import issues
-- **Developer Experience**: Simpler API for memory monitoring
-
-## 🔧 Critical Stability Fixes
-
-### Browser Manager Race Condition Resolution
-
-**The Problem:** Concurrent page creation in persistent browser contexts caused "Target page/context closed" errors during high-concurrency operations.
-
-**My Solution:** Implemented thread-safe page creation with proper locking mechanisms.
-
-```python
-# Fixed: Safe concurrent page creation
-browser_config = BrowserConfig(
-    browser_type="chromium",
-    use_persistent_context=True,  # Now thread-safe
-    max_concurrent_sessions=10    # Safely handle concurrent requests
-)
-
-async with AsyncWebCrawler(config=browser_config) as crawler:
-    # These concurrent operations are now stable
-    tasks = [crawler.arun(url) for url in url_list]
-    results = await asyncio.gather(*tasks)  # No more race conditions
-```
-
-### Enhanced Browser Profiler
-
-**The Problem:** Inconsistent keyboard handling across platforms and unreliable quit mechanisms.
-
-**My Solution:** Cross-platform keyboard listeners with improved quit handling.
-
-### Advanced URL Processing
-
-**The Problem:** Raw URL formats (`raw://` and `raw:`) weren't properly handled, and base tag link resolution was incomplete.
-
-**My Solution:** Enhanced URL preprocessing and base tag support.
-
-```python
-# Now properly handles all URL formats
-urls = [
-    "https://example.com",
-    "raw://static-html-content", 
-    "raw:file://local-file.html"
-]
-
-# Base tag links are now correctly resolved
-config = CrawlerRunConfig(
-    include_links=True,  # Links properly resolved with base tags
-    resolve_absolute_urls=True
-)
-```
-
-## 🛡️ Enhanced Proxy Configuration
-
-**The Problem:** Proxy configuration only accepted specific formats, limiting flexibility.
-
-**My Solution:** Enhanced ProxyConfig to support both dictionary and string formats.
-
-```python
-# Multiple proxy configuration formats now supported
-from crawl4ai import BrowserConfig, ProxyConfig
-
-# String format
-proxy_config = ProxyConfig("http://proxy.example.com:8080")
-
-# Dictionary format  
-proxy_config = ProxyConfig({
-    "server": "http://proxy.example.com:8080",
-    "username": "user",
-    "password": "pass"
-})
-
-# Use with crawler
-browser_config = BrowserConfig(proxy_config=proxy_config)
-async with AsyncWebCrawler(config=browser_config) as crawler:
-    result = await crawler.arun("https://httpbin.org/ip")
-```
-
-## 🐳 Docker & Infrastructure Improvements
-
-This release includes several Docker and infrastructure improvements:
-
-- **Better API Token Handling**: Improved Docker example scripts with correct endpoints
-- **Raw HTML Support**: Enhanced Docker API to handle raw HTML content properly
-- **Documentation Updates**: Comprehensive Docker deployment examples
-- **Test Coverage**: Expanded test suite with better coverage
-
-## 📚 Documentation & Examples
-
-Enhanced documentation includes:
-
-- **LLM Table Extraction Guide**: Comprehensive examples and best practices
-- **Migration Documentation**: Updated patterns for new table extraction methods  
-- **Docker Deployment**: Complete deployment guide with examples
-- **Performance Optimization**: Guidelines for concurrent crawling
-
-## 🙏 Acknowledgments
-
-Thanks to our contributors and community for feedback, bug reports, and feature requests that made this release possible.
-
-## 📚 Resources
-
-- [Full Documentation](https://docs.crawl4ai.com)
-- [GitHub Repository](https://github.com/unclecode/crawl4ai)
-- [Discord Community](https://discord.gg/crawl4ai)
-- [LLM Table Extraction Examples](https://github.com/unclecode/crawl4ai/blob/main/docs/examples/llm_table_extraction_example.py)
-
----
-
-*Crawl4AI v0.7.4 delivers intelligent table extraction and significant performance improvements. The new LLMTableExtraction strategy handles complex tables that were previously impossible to process, while concurrency improvements make batch operations 3-4x faster. Try the intelligent table extraction—it's a game changer for data extraction workflows!*
-
-**Happy Crawling! 🕷️**
-
-*- The Crawl4AI Team*

docs/examples/llm_table_extraction_example.py

@@ -1,356 +0,0 @@
-#!/usr/bin/env python3
-"""
-Example demonstrating LLM-based table extraction in Crawl4AI.
-
-This example shows how to use the LLMTableExtraction strategy to extract
-complex tables from web pages, including handling rowspan, colspan, and nested tables.
-"""
-
-import os
-import sys
-
-# Get the grandparent directory
-grandparent_dir = os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
-sys.path.append(grandparent_dir)
-__location__ = os.path.realpath(os.path.join(os.getcwd(), os.path.dirname(__file__)))
-
-
-
-import asyncio
-from crawl4ai import (
-    AsyncWebCrawler,
-    CrawlerRunConfig,
-    LLMConfig,
-    LLMTableExtraction,
-    CacheMode
-)
-import pandas as pd
-
-
-# Example 1: Basic LLM Table Extraction
-async def basic_llm_extraction():
-    """Extract tables using LLM with default settings."""
-    print("\n=== Example 1: Basic LLM Table Extraction ===")
-    
-    # Configure LLM (using OpenAI GPT-4o-mini for cost efficiency)
-    llm_config = LLMConfig(
-        provider="openai/gpt-4.1-mini",
-        api_token="env:OPENAI_API_KEY",  # Uses environment variable
-        temperature=0.1,  # Low temperature for consistency
-        max_tokens=32000
-    )
-    
-    # Create LLM table extraction strategy
-    table_strategy = LLMTableExtraction(
-        llm_config=llm_config,
-        verbose=True,
-        # css_selector="div.mw-content-ltr",
-        max_tries=2,
-        enable_chunking=True,
-        chunk_token_threshold=5000,  # Lower threshold to force chunking
-        min_rows_per_chunk=10,
-        max_parallel_chunks=3
-    )
-    
-    # Configure crawler with the strategy
-    config = CrawlerRunConfig(
-        cache_mode=CacheMode.BYPASS,
-        table_extraction=table_strategy
-    )
-    
-    async with AsyncWebCrawler() as crawler:
-        # Extract tables from a Wikipedia page
-        result = await crawler.arun(
-            url="https://en.wikipedia.org/wiki/List_of_chemical_elements",
-            config=config
-        )
-        
-        if result.success:
-            print(f"✓ Found {len(result.tables)} tables")
-            
-            # Display first table
-            if result.tables:
-                first_table = result.tables[0]
-                print(f"\nFirst table:")
-                print(f"  Headers: {first_table['headers'][:5]}...")
-                print(f"  Rows: {len(first_table['rows'])}")
-                
-                # Convert to pandas DataFrame
-                df = pd.DataFrame(
-                    first_table['rows'],
-                    columns=first_table['headers']
-                )
-                print(f"\nDataFrame shape: {df.shape}")
-                print(df.head())
-        else:
-            print(f"✗ Extraction failed: {result.error}")
-
-
-# Example 2: Focused Extraction with CSS Selector
-async def focused_extraction():
-    """Extract tables from specific page sections using CSS selectors."""
-    print("\n=== Example 2: Focused Extraction with CSS Selector ===")
-    
-    # HTML with multiple tables
-    test_html = """
-    <html>
-    <body>
-        <div class="sidebar">
-            <table role="presentation">
-                <tr><td>Navigation</td></tr>
-            </table>
-        </div>
-        
-        <div class="main-content">
-            <table id="data-table">
-                <caption>Quarterly Sales Report</caption>
-                <thead>
-                    <tr>
-                        <th rowspan="2">Product</th>
-                        <th colspan="3">Q1 2024</th>
-                    </tr>
-                    <tr>
-                        <th>Jan</th>
-                        <th>Feb</th>
-                        <th>Mar</th>
-                    </tr>
-                </thead>
-                <tbody>
-                    <tr>
-                        <td>Widget A</td>
-                        <td>100</td>
-                        <td>120</td>
-                        <td>140</td>
-                    </tr>
-                    <tr>
-                        <td>Widget B</td>
-                        <td>200</td>
-                        <td>180</td>
-                        <td>220</td>
-                    </tr>
-                </tbody>
-            </table>
-        </div>
-    </body>
-    </html>
-    """
-    
-    llm_config = LLMConfig(
-        provider="openai/gpt-4.1-mini",
-        api_token="env:OPENAI_API_KEY"
-    )
-    
-    # Focus only on main content area
-    table_strategy = LLMTableExtraction(
-        llm_config=llm_config,
-        css_selector=".main-content",  # Only extract from main content
-        verbose=True
-    )
-    
-    config = CrawlerRunConfig(
-        cache_mode=CacheMode.BYPASS,
-        table_extraction=table_strategy
-    )
-    
-    async with AsyncWebCrawler() as crawler:
-        result = await crawler.arun(
-            url=f"raw:{test_html}",
-            config=config
-        )
-        
-        if result.success and result.tables:
-            table = result.tables[0]
-            print(f"✓ Extracted table: {table.get('caption', 'No caption')}")
-            print(f"  Headers: {table['headers']}")
-            print(f"  Metadata: {table['metadata']}")
-            
-            # The LLM should have handled the rowspan/colspan correctly
-            print("\nProcessed data (rowspan/colspan handled):")
-            for i, row in enumerate(table['rows']):
-                print(f"  Row {i+1}: {row}")
-
-
-# Example 3: Comparing with Default Extraction
-async def compare_strategies():
-    """Compare LLM extraction with default extraction on complex tables."""
-    print("\n=== Example 3: Comparing LLM vs Default Extraction ===")
-    
-    # Complex table with nested structure
-    complex_html = """
-    <html>
-    <body>
-        <table>
-            <tr>
-                <th rowspan="3">Category</th>
-                <th colspan="2">2023</th>
-                <th colspan="2">2024</th>
-            </tr>
-            <tr>
-                <th>H1</th>
-                <th>H2</th>
-                <th>H1</th>
-                <th>H2</th>
-            </tr>
-            <tr>
-                <td colspan="4">All values in millions</td>
-            </tr>
-            <tr>
-                <td>Revenue</td>
-                <td>100</td>
-                <td>120</td>
-                <td>130</td>
-                <td>145</td>
-            </tr>
-            <tr>
-                <td>Profit</td>
-                <td>20</td>
-                <td>25</td>
-                <td>28</td>
-                <td>32</td>
-            </tr>
-        </table>
-    </body>
-    </html>
-    """
-    
-    async with AsyncWebCrawler() as crawler:
-        # Test with default extraction
-        from crawl4ai import DefaultTableExtraction
-        
-        default_strategy = DefaultTableExtraction(
-            table_score_threshold=3,
-            verbose=True
-        )
-        
-        config_default = CrawlerRunConfig(
-            cache_mode=CacheMode.BYPASS,
-            table_extraction=default_strategy
-        )
-        
-        result_default = await crawler.arun(
-            url=f"raw:{complex_html}",
-            config=config_default
-        )
-        
-        # Test with LLM extraction
-        llm_strategy = LLMTableExtraction(
-            llm_config=LLMConfig(
-                provider="openai/gpt-4.1-mini",
-                api_token="env:OPENAI_API_KEY"
-            ),
-            verbose=True
-        )
-        
-        config_llm = CrawlerRunConfig(
-            cache_mode=CacheMode.BYPASS,
-            table_extraction=llm_strategy
-        )
-        
-        result_llm = await crawler.arun(
-            url=f"raw:{complex_html}",
-            config=config_llm
-        )
-        
-        # Compare results
-        print("\nDefault Extraction:")
-        if result_default.tables:
-            table = result_default.tables[0]
-            print(f"  Headers: {table.get('headers', [])}")
-            print(f"  Rows: {len(table.get('rows', []))}")
-            for i, row in enumerate(table.get('rows', [])[:3]):
-                print(f"    Row {i+1}: {row}")
-        
-        print("\nLLM Extraction (handles complex structure better):")
-        if result_llm.tables:
-            table = result_llm.tables[0]
-            print(f"  Headers: {table.get('headers', [])}")
-            print(f"  Rows: {len(table.get('rows', []))}")
-            for i, row in enumerate(table.get('rows', [])):
-                print(f"    Row {i+1}: {row}")
-            print(f"  Metadata: {table.get('metadata', {})}")
-
-# Example 4: Batch Processing Multiple Pages
-async def batch_extraction():
-    """Extract tables from multiple pages efficiently."""
-    print("\n=== Example 4: Batch Table Extraction ===")
-    
-    urls = [
-        "https://www.worldometers.info/geography/alphabetical-list-of-countries/",
-        # "https://en.wikipedia.org/wiki/List_of_chemical_elements",
-    ]
-    
-    llm_config = LLMConfig(
-        provider="openai/gpt-4.1-mini",
-        api_token="env:OPENAI_API_KEY",
-        temperature=0.1,
-        max_tokens=1500
-    )
-    
-    table_strategy = LLMTableExtraction(
-        llm_config=llm_config,
-        css_selector="div.datatable-container",  # Wikipedia data tables
-        verbose=False,
-        enable_chunking=True,
-        chunk_token_threshold=5000,  # Lower threshold to force chunking
-        min_rows_per_chunk=10,
-        max_parallel_chunks=3
-    )
-    
-    config = CrawlerRunConfig(
-        table_extraction=table_strategy,
-        cache_mode=CacheMode.BYPASS
-    )
-    
-    all_tables = []
-    
-    async with AsyncWebCrawler() as crawler:
-        for url in urls:
-            print(f"\nProcessing: {url.split('/')[-1][:50]}...")
-            result = await crawler.arun(url=url, config=config)
-            
-            if result.success and result.tables:
-                print(f"  ✓ Found {len(result.tables)} tables")
-                # Store first table from each page
-                if result.tables:
-                    all_tables.append({
-                        'url': url,
-                        'table': result.tables[0]
-                    })
-    
-    # Summary
-    print(f"\n=== Summary ===")
-    print(f"Extracted {len(all_tables)} tables from {len(urls)} pages")
-    for item in all_tables:
-        table = item['table']
-        print(f"\nFrom {item['url'].split('/')[-1][:30]}:")
-        print(f"  Columns: {len(table['headers'])}")
-        print(f"  Rows: {len(table['rows'])}")
-
-
-async def main():
-    """Run all examples."""
-    print("=" * 60)
-    print("LLM TABLE EXTRACTION EXAMPLES")
-    print("=" * 60)
-    
-    # Run examples (comment out ones you don't want to run)
-    
-    # Basic extraction
-    await basic_llm_extraction()
-    
-    # # Focused extraction with CSS
-    # await focused_extraction()
-    
-    # # Compare strategies
-    # await compare_strategies()
-    
-    # # Batch processing
-    # await batch_extraction()
-    
-    print("\n" + "=" * 60)
-    print("ALL EXAMPLES COMPLETED")
-    print("=" * 60)
-
-
-if __name__ == "__main__":
-    asyncio.run(main())

docs/examples/table_extraction_example.py

@@ -1,276 +0,0 @@
-"""
-Example: Using Table Extraction Strategies in Crawl4AI
-
-This example demonstrates how to use different table extraction strategies
-to extract tables from web pages.
-"""
-
-import asyncio
-import pandas as pd
-from crawl4ai import (
-    AsyncWebCrawler,
-    CrawlerRunConfig,
-    CacheMode,
-    DefaultTableExtraction,
-    NoTableExtraction,
-    TableExtractionStrategy
-)
-from typing import Dict, List, Any
-
-
-async def example_default_extraction():
-    """Example 1: Using default table extraction (automatic)."""
-    print("\n" + "="*50)
-    print("Example 1: Default Table Extraction")
-    print("="*50)
-    
-    async with AsyncWebCrawler() as crawler:
-        # No need to specify table_extraction - uses DefaultTableExtraction automatically
-        config = CrawlerRunConfig(
-            cache_mode=CacheMode.BYPASS,
-            table_score_threshold=7  # Adjust sensitivity (default: 7)
-        )
-        
-        result = await crawler.arun(
-            "https://en.wikipedia.org/wiki/List_of_countries_by_GDP_(nominal)",
-            config=config
-        )
-        
-        if result.success and result.tables:
-            print(f"Found {len(result.tables)} tables")
-            
-            # Convert first table to pandas DataFrame
-            if result.tables:
-                first_table = result.tables[0]
-                df = pd.DataFrame(
-                    first_table['rows'],
-                    columns=first_table['headers'] if first_table['headers'] else None
-                )
-                print(f"\nFirst table preview:")
-                print(df.head())
-                print(f"Shape: {df.shape}")
-
-
-async def example_custom_configuration():
-    """Example 2: Custom table extraction configuration."""
-    print("\n" + "="*50)
-    print("Example 2: Custom Table Configuration")
-    print("="*50)
-    
-    async with AsyncWebCrawler() as crawler:
-        # Create custom extraction strategy with specific settings
-        table_strategy = DefaultTableExtraction(
-            table_score_threshold=5,  # Lower threshold for more permissive detection
-            min_rows=3,  # Only extract tables with at least 3 rows
-            min_cols=2,  # Only extract tables with at least 2 columns
-            verbose=True
-        )
-        
-        config = CrawlerRunConfig(
-            cache_mode=CacheMode.BYPASS,
-            table_extraction=table_strategy,
-            # Target specific tables using CSS selector
-            css_selector="div.main-content"
-        )
-        
-        result = await crawler.arun(
-            "https://example.com/data",
-            config=config
-        )
-        
-        if result.success:
-            print(f"Found {len(result.tables)} tables matching criteria")
-            
-            for i, table in enumerate(result.tables):
-                print(f"\nTable {i+1}:")
-                print(f"  Caption: {table.get('caption', 'No caption')}")
-                print(f"  Size: {table['metadata']['row_count']} rows × {table['metadata']['column_count']} columns")
-                print(f"  Has headers: {table['metadata']['has_headers']}")
-
-
-async def example_disable_extraction():
-    """Example 3: Disable table extraction when not needed."""
-    print("\n" + "="*50)
-    print("Example 3: Disable Table Extraction")
-    print("="*50)
-    
-    async with AsyncWebCrawler() as crawler:
-        # Use NoTableExtraction to skip table processing entirely
-        config = CrawlerRunConfig(
-            cache_mode=CacheMode.BYPASS,
-            table_extraction=NoTableExtraction()  # No tables will be extracted
-        )
-        
-        result = await crawler.arun(
-            "https://example.com",
-            config=config
-        )
-        
-        if result.success:
-            print(f"Tables extracted: {len(result.tables)} (should be 0)")
-            print("Table extraction disabled - better performance for non-table content")
-
-
-class FinancialTableExtraction(TableExtractionStrategy):
-    """
-    Custom strategy for extracting financial tables with specific requirements.
-    """
-    
-    def __init__(self, currency_symbols=None, **kwargs):
-        super().__init__(**kwargs)
-        self.currency_symbols = currency_symbols or ['$', '€', '£', '¥']
-    
-    def extract_tables(self, element, **kwargs):
-        """Extract only tables that appear to contain financial data."""
-        tables_data = []
-        
-        for table in element.xpath(".//table"):
-            # Check if table contains currency symbols
-            table_text = ''.join(table.itertext())
-            has_currency = any(symbol in table_text for symbol in self.currency_symbols)
-            
-            if not has_currency:
-                continue
-            
-            # Extract using base logic (could reuse DefaultTableExtraction logic)
-            headers = []
-            rows = []
-            
-            # Extract headers
-            for th in table.xpath(".//thead//th | .//tr[1]//th"):
-                headers.append(th.text_content().strip())
-            
-            # Extract rows
-            for tr in table.xpath(".//tbody//tr | .//tr[position()>1]"):
-                row = []
-                for td in tr.xpath(".//td"):
-                    cell_text = td.text_content().strip()
-                    # Clean currency values
-                    for symbol in self.currency_symbols:
-                        cell_text = cell_text.replace(symbol, '')
-                    row.append(cell_text)
-                if row:
-                    rows.append(row)
-            
-            if headers or rows:
-                tables_data.append({
-                    "headers": headers,
-                    "rows": rows,
-                    "caption": table.xpath(".//caption/text()")[0] if table.xpath(".//caption") else "",
-                    "summary": table.get("summary", ""),
-                    "metadata": {
-                        "type": "financial",
-                        "has_currency": True,
-                        "row_count": len(rows),
-                        "column_count": len(headers) if headers else len(rows[0]) if rows else 0
-                    }
-                })
-        
-        return tables_data
-
-
-async def example_custom_strategy():
-    """Example 4: Custom table extraction strategy."""
-    print("\n" + "="*50)
-    print("Example 4: Custom Financial Table Strategy")
-    print("="*50)
-    
-    async with AsyncWebCrawler() as crawler:
-        # Use custom strategy for financial tables
-        config = CrawlerRunConfig(
-            cache_mode=CacheMode.BYPASS,
-            table_extraction=FinancialTableExtraction(
-                currency_symbols=['$', '€'],
-                verbose=True
-            )
-        )
-        
-        result = await crawler.arun(
-            "https://finance.yahoo.com/",
-            config=config
-        )
-        
-        if result.success:
-            print(f"Found {len(result.tables)} financial tables")
-            
-            for table in result.tables:
-                if table['metadata'].get('type') == 'financial':
-                    print(f"  ✓ Financial table with {table['metadata']['row_count']} rows")
-
-
-async def example_combined_extraction():
-    """Example 5: Combine table extraction with other strategies."""
-    print("\n" + "="*50)
-    print("Example 5: Combined Extraction Strategies")
-    print("="*50)
-    
-    from crawl4ai import LLMExtractionStrategy, LLMConfig
-    
-    async with AsyncWebCrawler() as crawler:
-        # Define schema for structured extraction
-        schema = {
-            "type": "object",
-            "properties": {
-                "page_title": {"type": "string"},
-                "main_topic": {"type": "string"},
-                "key_figures": {
-                    "type": "array",
-                    "items": {"type": "string"}
-                }
-            }
-        }
-        
-        config = CrawlerRunConfig(
-            cache_mode=CacheMode.BYPASS,
-            # Table extraction
-            table_extraction=DefaultTableExtraction(
-                table_score_threshold=6,
-                min_rows=2
-            ),
-            # LLM extraction for structured data
-            extraction_strategy=LLMExtractionStrategy(
-                llm_config=LLMConfig(provider="openai"),
-                schema=schema
-            )
-        )
-        
-        result = await crawler.arun(
-            "https://en.wikipedia.org/wiki/Economy_of_the_United_States",
-            config=config
-        )
-        
-        if result.success:
-            print(f"Tables found: {len(result.tables)}")
-            
-            # Tables are in result.tables
-            if result.tables:
-                print(f"First table has {len(result.tables[0]['rows'])} rows")
-            
-            # Structured data is in result.extracted_content
-            if result.extracted_content:
-                import json
-                structured_data = json.loads(result.extracted_content)
-                print(f"Page title: {structured_data.get('page_title', 'N/A')}")
-                print(f"Main topic: {structured_data.get('main_topic', 'N/A')}")
-
-
-async def main():
-    """Run all examples."""
-    print("\n" + "="*60)
-    print("CRAWL4AI TABLE EXTRACTION EXAMPLES")
-    print("="*60)
-    
-    # Run examples
-    await example_default_extraction()
-    await example_custom_configuration()
-    await example_disable_extraction()
-    await example_custom_strategy()
-    # await example_combined_extraction()  # Requires OpenAI API key
-    
-    print("\n" + "="*60)
-    print("EXAMPLES COMPLETED")
-    print("="*60)
-
-
-if __name__ == "__main__":
-    asyncio.run(main())

docs/md_v2/blog/index.md

@@ -20,22 +20,136 @@ Ever wondered why your AI coding assistant struggles with your library despite c
 
 ## Latest Release
 
-### [Crawl4AI v0.7.4 – The Intelligent Table Extraction & Performance Update](../blog/release-v0.7.4.md)
-*August 17, 2025*
+### [Crawl4AI v0.7.3 – The Multi-Config Intelligence Update](releases/0.7.3.md)
+*August 6, 2025*
 
-Crawl4AI v0.7.4 introduces revolutionary LLM-powered table extraction with intelligent chunking, performance improvements for concurrent crawling, enhanced browser management, and critical stability fixes that make Crawl4AI more robust for production workloads.
+Crawl4AI v0.7.3 brings smarter URL-specific configurations, flexible Docker deployments, and critical stability improvements. Configure different crawling strategies for different URL patterns in a single batch—perfect for mixed content sites with docs, blogs, and APIs.
 
 Key highlights:
-- **🚀 LLMTableExtraction**: Revolutionary table extraction with intelligent chunking for massive tables
-- **⚡ Dispatcher Bug Fix**: Fixed sequential processing issue in arun_many for fast-completing tasks
-- **🧹 Memory Management Refactor**: Streamlined memory utilities and better resource management
-- **🔧 Browser Manager Fixes**: Resolved race conditions in concurrent page creation
-- **🔗 Advanced URL Processing**: Better handling of raw URLs and base tag link resolution
+- **Multi-URL Configurations**: Different strategies for different URL patterns in one crawl
+- **Flexible Docker LLM Providers**: Configure providers via environment variables  
+- **Bug Fixes**: Critical stability improvements for production deployments
+- **Documentation Updates**: Clearer examples and improved API documentation
 
-[Read full release notes →](../blog/release-v0.7.4.md)
+[Read full release notes →](releases/0.7.3.md)
 
 ---
 
+## Previous Releases
+
+### [Crawl4AI v0.7.0 – The Adaptive Intelligence Update](releases/0.7.0.md)
+*January 28, 2025*
+
+Introduced groundbreaking intelligence features including Adaptive Crawling, Virtual Scroll support, intelligent Link Preview, and the Async URL Seeder for massive URL discovery.
+
+[Read release notes →](releases/0.7.0.md)
+
+### [Crawl4AI v0.6.0 – World-Aware Crawling, Pre-Warmed Browsers, and the MCP API](releases/0.6.0.md)
+*December 23, 2024*
+
+Crawl4AI v0.6.0 brought major architectural upgrades including world-aware crawling (set geolocation, locale, and timezone), real-time traffic capture, and a memory-efficient crawler pool with pre-warmed pages.  
+
+The Docker server now exposes a full-featured MCP socket + SSE interface, supports streaming, and comes with a new Playground UI. Plus, table extraction is now native, and the new stress-test framework supports crawling 1,000+ URLs.  
+
+Other key changes:  
+
+*   Native support for `result.media["tables"]` to export DataFrames  
+* Full network + console logs and MHTML snapshot per crawl  
+* Browser pooling and pre-warming for faster cold starts  
+* New streaming endpoints via MCP API and Playground  
+* Robots.txt support, proxy rotation, and improved session handling  
+* Deprecated old markdown names, legacy modules cleaned up  
+* Massive repo cleanup: ~36K insertions, ~5K deletions across 121 files
+
+[Read full release notes →](releases/0.6.0.md)
+
+---
+
+### [Crawl4AI v0.5.0: Deep Crawling, Scalability, and a New CLI!](releases/0.5.0.md)
+
+My dear friends and crawlers, there you go, this is the release of Crawl4AI v0.5.0! This release brings a wealth of new features, performance improvements, and a more streamlined developer experience.  Here's a breakdown of what's new:
+
+**Major New Features:**
+
+*   **Deep Crawling:** Explore entire websites with configurable strategies (BFS, DFS, Best-First).  Define custom filters and URL scoring for targeted crawls.
+*   **Memory-Adaptive Dispatcher:**  Handle large-scale crawls with ease!  Our new dispatcher dynamically adjusts concurrency based on available memory and includes built-in rate limiting.
+*   **Multiple Crawler Strategies:** Choose between the full-featured Playwright browser-based crawler or a new, *much* faster HTTP-only crawler for simpler tasks.
+*   **Docker Deployment:**  Deploy Crawl4AI as a scalable, self-contained service with built-in API endpoints and optional JWT authentication.
+*   **Command-Line Interface (CLI):**  Interact with Crawl4AI directly from your terminal.  Crawl, configure, and extract data with simple commands.
+*   **LLM Configuration (`LLMConfig`):** A new, unified way to configure LLM providers (OpenAI, Anthropic, Ollama, etc.) for extraction, filtering, and schema generation.  Simplifies API key management and switching between models.
+
+**Minor Updates & Improvements:**
+
+*   **LXML Scraping Mode:** Faster HTML parsing with `LXMLWebScrapingStrategy`.
+*   **Proxy Rotation:** Added `ProxyRotationStrategy` with a `RoundRobinProxyStrategy` implementation.
+*   **PDF Processing:** Extract text, images, and metadata from PDF files.
+*   **URL Redirection Tracking:**  Automatically follows and records redirects.
+*   **Robots.txt Compliance:**  Optionally respect website crawling rules.
+*   **LLM-Powered Schema Generation:**  Automatically create extraction schemas using an LLM.
+*   **`LLMContentFilter`:** Generate high-quality, focused markdown using an LLM.
+*   **Improved Error Handling & Stability:** Numerous bug fixes and performance enhancements.
+*   **Enhanced Documentation:**  Updated guides and examples.
+
+**Breaking Changes & Migration:**
+
+This release includes several breaking changes to improve the library's structure and consistency.  Here's what you need to know:
+
+*   **`arun_many()` Behavior:** Now uses the `MemoryAdaptiveDispatcher` by default.  The return type depends on the `stream` parameter in `CrawlerRunConfig`.  Adjust code that relied on unbounded concurrency.
+*   **`max_depth` Location:** Moved to `CrawlerRunConfig` and now controls *crawl depth*.
+*   **Deep Crawling Imports:**  Import `DeepCrawlStrategy` and related classes from `crawl4ai.deep_crawling`.
+*   **`BrowserContext` API:**  Updated; the old `get_context` method is deprecated.
+*   **Optional Model Fields:** Many data model fields are now optional.  Handle potential `None` values.
+*   **`ScrapingMode` Enum:** Replaced with strategy pattern (`WebScrapingStrategy`, `LXMLWebScrapingStrategy`).
+*   **`content_filter` Parameter:** Removed from `CrawlerRunConfig`. Use extraction strategies or markdown generators with filters.
+*   **Removed Functionality:** The synchronous `WebCrawler`, the old CLI, and docs management tools have been removed.
+*   **Docker:**  Significant changes to deployment.  See the [Docker documentation](../deploy/docker/README.md).
+*   **`ssl_certificate.json`:** This file has been removed.
+* **Config**: FastFilterChain has been replaced with FilterChain
+* **Deep-Crawl**: DeepCrawlStrategy.arun now returns Union[CrawlResultT, List[CrawlResultT], AsyncGenerator[CrawlResultT, None]]
+* **Proxy**: Removed synchronous WebCrawler support and related rate limiting configurations
+*   **LLM Parameters:** Use the new `LLMConfig` object instead of passing `provider`, `api_token`, `base_url`, and `api_base` directly to `LLMExtractionStrategy` and `LLMContentFilter`.
+
+**In short:** Update imports, adjust `arun_many()` usage, check for optional fields, and review the Docker deployment guide.
+
+## License Change
+
+Crawl4AI v0.5.0 updates the license to Apache 2.0 *with a required attribution clause*.  This means you are free to use, modify, and distribute Crawl4AI (even commercially), but you *must* clearly attribute the project in any public use or distribution.  See the updated `LICENSE` file for the full legal text and specific requirements.
+
+**Get Started:**
+
+*   **Installation:** `pip install "crawl4ai[all]"` (or use the Docker image)
+*   **Documentation:** [https://docs.crawl4ai.com](https://docs.crawl4ai.com)
+*   **GitHub:** [https://github.com/unclecode/crawl4ai](https://github.com/unclecode/crawl4ai)
+
+I'm very excited to see what you build with Crawl4AI v0.5.0!
+
+---
+
+### [0.4.2 - Configurable Crawlers, Session Management, and Smarter Screenshots](releases/0.4.2.md)
+*December 12, 2024*
+
+The 0.4.2 update brings massive improvements to configuration, making crawlers and browsers easier to manage with dedicated objects. You can now import/export local storage for seamless session management. Plus, long-page screenshots are faster and cleaner, and full-page PDF exports are now possible. Check out all the new features to make your crawling experience even smoother.
+
+[Read full release notes →](releases/0.4.2.md)
+
+---
+
+### [0.4.1 - Smarter Crawling with Lazy-Load Handling, Text-Only Mode, and More](releases/0.4.1.md)
+*December 8, 2024*
+
+This release brings major improvements to handling lazy-loaded images, a blazing-fast Text-Only Mode, full-page scanning for infinite scrolls, dynamic viewport adjustments, and session reuse for efficient crawling. If you're looking to improve speed, reliability, or handle dynamic content with ease, this update has you covered.
+
+[Read full release notes →](releases/0.4.1.md)
+
+---
+
+### [0.4.0 - Major Content Filtering Update](releases/0.4.0.md)
+*December 1, 2024*
+
+Introduced significant improvements to content filtering, multi-threaded environment handling, and user-agent generation. This release features the new PruningContentFilter, enhanced thread safety, and improved test coverage.
+
+[Read full release notes →](releases/0.4.0.md)
+
 ## Project History
 
 Curious about how Crawl4AI has evolved? Check out our [complete changelog](https://github.com/unclecode/crawl4ai/blob/main/CHANGELOG.md) for a detailed history of all versions and updates.

docs/md_v2/core/table_extraction.md

@@ -1,807 +0,0 @@
-# Table Extraction Strategies
-
-## Overview
-
-**New in v0.7.3+**: Table extraction now follows the **Strategy Design Pattern**, providing unprecedented flexibility and power for handling different table structures. Don't worry - **your existing code still works!** We maintain full backward compatibility while offering new capabilities.
-
-### What's Changed?
-- **Architecture**: Table extraction now uses pluggable strategies
-- **Backward Compatible**: Your existing code with `table_score_threshold` continues to work
-- **More Power**: Choose from multiple strategies or create your own
-- **Same Default Behavior**: By default, uses `DefaultTableExtraction` (same as before)
-
-### Key Points
-✅ **Old code still works** - No breaking changes  
-✅ **Same default behavior** - Uses the proven extraction algorithm  
-✅ **New capabilities** - Add LLM extraction or custom strategies when needed  
-✅ **Strategy pattern** - Clean, extensible architecture
-
-## Quick Start
-
-### The Simplest Way (Works Like Before)
-
-If you're already using Crawl4AI, nothing changes:
-
-```python
-import asyncio
-from crawl4ai import AsyncWebCrawler, CrawlerRunConfig
-
-async def extract_tables():
-    async with AsyncWebCrawler() as crawler:
-        # This works exactly like before - uses DefaultTableExtraction internally
-        result = await crawler.arun("https://example.com/data")
-        
-        # Tables are automatically extracted and available in result.tables
-        for table in result.tables:
-            print(f"Table with {len(table['rows'])} rows and {len(table['headers'])} columns")
-            print(f"Headers: {table['headers']}")
-            print(f"First row: {table['rows'][0] if table['rows'] else 'No data'}")
-
-asyncio.run(extract_tables())
-```
-
-### Using the Old Configuration (Still Supported)
-
-Your existing code with `table_score_threshold` continues to work:
-
-```python
-# This old approach STILL WORKS - we maintain backward compatibility
-config = CrawlerRunConfig(
-    table_score_threshold=7  # Internally creates DefaultTableExtraction(table_score_threshold=7)
-)
-result = await crawler.arun(url, config)
-```
-
-## Table Extraction Strategies
-
-### Understanding the Strategy Pattern
-
-The strategy pattern allows you to choose different table extraction algorithms at runtime. Think of it as having different tools in a toolbox - you pick the right one for the job:
-
-- **No explicit strategy?** → Uses `DefaultTableExtraction` automatically (same as v0.7.2 and earlier)
-- **Need complex table handling?** → Choose `LLMTableExtraction` (costs money, use sparingly)
-- **Want to disable tables?** → Use `NoTableExtraction`
-- **Have special requirements?** → Create a custom strategy
-
-### Available Strategies
-
-| Strategy | Description | Use Case | Cost | When to Use |
-|----------|-------------|----------|------|-------------|
-| `DefaultTableExtraction` | **RECOMMENDED**: Same algorithm as before v0.7.3 | General purpose (default) | Free | **Use this first - handles 95% of cases** |
-| `LLMTableExtraction` | AI-powered extraction for complex tables | Tables with complex rowspan/colspan | **$$$ Per API call** | Only when DefaultTableExtraction fails |
-| `NoTableExtraction` | Disables table extraction | When tables aren't needed | Free | For text-only extraction |
-| Custom strategies | User-defined extraction logic | Specialized requirements | Free | Domain-specific needs |
-
-> **⚠️ CRITICAL COST WARNING for LLMTableExtraction**: 
-> 
-> **DO NOT USE `LLMTableExtraction` UNLESS ABSOLUTELY NECESSARY!**
-> 
-> - **Always try `DefaultTableExtraction` first** - It's free and handles most tables perfectly
-> - LLM extraction **costs money** with every API call
-> - For large tables (100+ rows), LLM extraction can be **very slow**
-> - **For large tables**: If you must use LLM, choose fast providers:
->   - ✅ **Groq** (fastest inference)
->   - ✅ **Cerebras** (optimized for speed)
->   - ⚠️ Avoid: OpenAI, Anthropic for large tables (slower)
-> 
-> **🚧 WORK IN PROGRESS**: 
-> We are actively developing an **advanced non-LLM algorithm** that will handle complex table structures (rowspan, colspan, nested tables) for **FREE**. This will replace the need for costly LLM extraction in most cases. Coming soon!
-
-### DefaultTableExtraction
-
-The default strategy uses a sophisticated scoring system to identify data tables:
-
-```python
-from crawl4ai import DefaultTableExtraction, CrawlerRunConfig
-
-# Customize the default extraction
-table_strategy = DefaultTableExtraction(
-    table_score_threshold=7,  # Scoring threshold (default: 7)
-    min_rows=2,               # Minimum rows required
-    min_cols=2,               # Minimum columns required
-    verbose=True              # Enable detailed logging
-)
-
-config = CrawlerRunConfig(
-    table_extraction=table_strategy
-)
-```
-
-#### Scoring System
-
-The scoring system evaluates multiple factors:
-
-| Factor | Score Impact | Description |
-|--------|--------------|-------------|
-| Has `<thead>` | +2 | Semantic table structure |
-| Has `<tbody>` | +1 | Organized table body |
-| Has `<th>` elements | +2 | Header cells present |
-| Headers in correct position | +1 | Proper semantic structure |
-| Consistent column count | +2 | Regular data structure |
-| Has caption | +2 | Descriptive caption |
-| Has summary | +1 | Summary attribute |
-| High text density | +2 to +3 | Content-rich cells |
-| Data attributes | +0.5 each | Data-* attributes |
-| Nested tables | -3 | Often indicates layout |
-| Role="presentation" | -3 | Explicitly non-data |
-| Too few rows | -2 | Insufficient data |
-
-### LLMTableExtraction (Use Sparingly!)
-
-**⚠️ WARNING**: Only use this when `DefaultTableExtraction` fails with complex tables!
-
-LLMTableExtraction uses AI to understand complex table structures that traditional parsers struggle with. It automatically handles large tables through intelligent chunking and parallel processing:
-
-```python
-from crawl4ai import LLMTableExtraction, LLMConfig, CrawlerRunConfig
-
-# Configure LLM (costs money per call!)
-llm_config = LLMConfig(
-    provider="groq/llama-3.3-70b-versatile",  # Fast provider for large tables
-    api_token="your_api_key",
-    temperature=0.1
-)
-
-# Create LLM extraction strategy with smart chunking
-table_strategy = LLMTableExtraction(
-    llm_config=llm_config,
-    max_tries=3,                      # Retry up to 3 times if extraction fails
-    css_selector="table",             # Optional: focus on specific tables
-    enable_chunking=True,             # Automatically chunk large tables (default: True)
-    chunk_token_threshold=3000,       # Split tables larger than this (default: 3000 tokens)
-    min_rows_per_chunk=10,            # Minimum rows per chunk (default: 10)
-    max_parallel_chunks=5,            # Process up to 5 chunks in parallel (default: 5)
-    verbose=True
-)
-
-config = CrawlerRunConfig(
-    table_extraction=table_strategy
-)
-
-result = await crawler.arun(url, config)
-```
-
-#### When to Use LLMTableExtraction
-
-✅ **Use ONLY when**:
-- Tables have complex merged cells (rowspan/colspan) that break DefaultTableExtraction
-- Nested tables that need semantic understanding
-- Tables with irregular structures
-- You've tried DefaultTableExtraction and it failed
-
-❌ **Never use when**:
-- DefaultTableExtraction works (99% of cases)
-- Tables are simple or well-structured
-- You're processing many pages (costs add up!)
-- Tables have 100+ rows (very slow)
-
-#### How Smart Chunking Works
-
-LLMTableExtraction automatically handles large tables through intelligent chunking:
-
-1. **Automatic Detection**: Tables exceeding the token threshold are automatically split
-2. **Smart Splitting**: Chunks are created at row boundaries, preserving table structure
-3. **Header Preservation**: Each chunk includes the original headers for context
-4. **Parallel Processing**: Multiple chunks are processed simultaneously for speed
-5. **Intelligent Merging**: Results are merged back into a single, complete table
-
-**Chunking Parameters**:
-- `enable_chunking` (default: `True`): Automatically handle large tables
-- `chunk_token_threshold` (default: `3000`): When to split tables
-- `min_rows_per_chunk` (default: `10`): Ensures meaningful chunk sizes
-- `max_parallel_chunks` (default: `5`): Concurrent processing for speed
-
-The chunking is completely transparent - you get the same output format whether the table was processed in one piece or multiple chunks.
-
-#### Performance Optimization for LLMTableExtraction
-
-**Provider Recommendations by Table Size**:
-
-| Table Size | Recommended Providers | Why |
-|------------|----------------------|-----|
-| Small (<50 rows) | Any provider | Fast enough |
-| Medium (50-200 rows) | Groq, Cerebras | Optimized inference |
-| Large (200+ rows) | **Groq** (best), Cerebras | Fastest inference + automatic chunking |
-| Very Large (500+ rows) | Groq with chunking | Parallel processing keeps it fast |
-
-### NoTableExtraction
-
-Disable table extraction for better performance when tables aren't needed:
-
-```python
-from crawl4ai import NoTableExtraction, CrawlerRunConfig
-
-config = CrawlerRunConfig(
-    table_extraction=NoTableExtraction()
-)
-
-# Tables won't be extracted, improving performance
-result = await crawler.arun(url, config)
-assert len(result.tables) == 0
-```
-
-## Extracted Table Structure
-
-Each extracted table contains:
-
-```python
-{
-    "headers": ["Column 1", "Column 2", ...],  # Column headers
-    "rows": [                                   # Data rows
-        ["Row 1 Col 1", "Row 1 Col 2", ...],
-        ["Row 2 Col 1", "Row 2 Col 2", ...],
-    ],
-    "caption": "Table Caption",                # If present
-    "summary": "Table Summary",                # If present
-    "metadata": {
-        "row_count": 10,                       # Number of rows
-        "column_count": 3,                      # Number of columns
-        "has_headers": True,                    # Headers detected
-        "has_caption": True,                    # Caption exists
-        "has_summary": False,                   # Summary exists
-        "id": "data-table-1",                   # Table ID if present
-        "class": "financial-data"               # Table class if present
-    }
-}
-```
-
-## Configuration Options
-
-### Basic Configuration
-
-```python
-config = CrawlerRunConfig(
-    # Table extraction settings
-    table_score_threshold=7,      # Default threshold (backward compatible)
-    table_extraction=strategy,     # Optional: custom strategy
-    
-    # Filter what to process
-    css_selector="main",          # Focus on specific area
-    excluded_tags=["nav", "aside"] # Exclude page sections
-)
-```
-
-### Advanced Configuration
-
-```python
-from crawl4ai import DefaultTableExtraction, CrawlerRunConfig
-
-# Fine-tuned extraction
-strategy = DefaultTableExtraction(
-    table_score_threshold=5,      # Lower = more permissive
-    min_rows=3,                   # Require at least 3 rows
-    min_cols=2,                   # Require at least 2 columns
-    verbose=True                  # Detailed logging
-)
-
-config = CrawlerRunConfig(
-    table_extraction=strategy,
-    css_selector="article.content", # Target specific content
-    exclude_domains=["ads.com"],   # Exclude ad domains
-    cache_mode=CacheMode.BYPASS    # Fresh extraction
-)
-```
-
-## Working with Extracted Tables
-
-### Convert to Pandas DataFrame
-
-```python
-import pandas as pd
-
-async def tables_to_dataframes(url):
-    async with AsyncWebCrawler() as crawler:
-        result = await crawler.arun(url)
-        
-        dataframes = []
-        for table_data in result.tables:
-            # Create DataFrame
-            if table_data['headers']:
-                df = pd.DataFrame(
-                    table_data['rows'],
-                    columns=table_data['headers']
-                )
-            else:
-                df = pd.DataFrame(table_data['rows'])
-            
-            # Add metadata as DataFrame attributes
-            df.attrs['caption'] = table_data.get('caption', '')
-            df.attrs['metadata'] = table_data.get('metadata', {})
-            
-            dataframes.append(df)
-        
-        return dataframes
-```
-
-### Filter Tables by Criteria
-
-```python
-async def extract_large_tables(url):
-    async with AsyncWebCrawler() as crawler:
-        # Configure minimum size requirements
-        strategy = DefaultTableExtraction(
-            min_rows=10,
-            min_cols=3,
-            table_score_threshold=6
-        )
-        
-        config = CrawlerRunConfig(
-            table_extraction=strategy
-        )
-        
-        result = await crawler.arun(url, config)
-        
-        # Further filter results
-        large_tables = [
-            table for table in result.tables
-            if table['metadata']['row_count'] > 10
-            and table['metadata']['column_count'] > 3
-        ]
-        
-        return large_tables
-```
-
-### Export Tables to Different Formats
-
-```python
-import json
-import csv
-
-async def export_tables(url):
-    async with AsyncWebCrawler() as crawler:
-        result = await crawler.arun(url)
-        
-        for i, table in enumerate(result.tables):
-            # Export as JSON
-            with open(f'table_{i}.json', 'w') as f:
-                json.dump(table, f, indent=2)
-            
-            # Export as CSV
-            with open(f'table_{i}.csv', 'w', newline='') as f:
-                writer = csv.writer(f)
-                if table['headers']:
-                    writer.writerow(table['headers'])
-                writer.writerows(table['rows'])
-            
-            # Export as Markdown
-            with open(f'table_{i}.md', 'w') as f:
-                # Write headers
-                if table['headers']:
-                    f.write('| ' + ' | '.join(table['headers']) + ' |\n')
-                    f.write('|' + '---|' * len(table['headers']) + '\n')
-                
-                # Write rows
-                for row in table['rows']:
-                    f.write('| ' + ' | '.join(str(cell) for cell in row) + ' |\n')
-```
-
-## Creating Custom Strategies
-
-Extend `TableExtractionStrategy` to create custom extraction logic:
-
-### Example: Financial Table Extractor
-
-```python
-from crawl4ai import TableExtractionStrategy
-from typing import List, Dict, Any
-import re
-
-class FinancialTableExtractor(TableExtractionStrategy):
-    """Extract tables containing financial data."""
-    
-    def __init__(self, currency_symbols=None, require_numbers=True, **kwargs):
-        super().__init__(**kwargs)
-        self.currency_symbols = currency_symbols or ['$', '€', '£', '¥']
-        self.require_numbers = require_numbers
-        self.number_pattern = re.compile(r'\d+[,.]?\d*')
-    
-    def extract_tables(self, element, **kwargs):
-        tables_data = []
-        
-        for table in element.xpath(".//table"):
-            # Check if table contains financial indicators
-            table_text = ''.join(table.itertext())
-            
-            # Must contain currency symbols
-            has_currency = any(sym in table_text for sym in self.currency_symbols)
-            if not has_currency:
-                continue
-            
-            # Must contain numbers if required
-            if self.require_numbers:
-                numbers = self.number_pattern.findall(table_text)
-                if len(numbers) < 3:  # Arbitrary minimum
-                    continue
-            
-            # Extract the table data
-            table_data = self._extract_financial_data(table)
-            if table_data:
-                tables_data.append(table_data)
-        
-        return tables_data
-    
-    def _extract_financial_data(self, table):
-        """Extract and clean financial data from table."""
-        headers = []
-        rows = []
-        
-        # Extract headers
-        for th in table.xpath(".//thead//th | .//tr[1]//th"):
-            headers.append(th.text_content().strip())
-        
-        # Extract and clean rows
-        for tr in table.xpath(".//tbody//tr | .//tr[position()>1]"):
-            row = []
-            for td in tr.xpath(".//td"):
-                text = td.text_content().strip()
-                # Clean currency formatting
-                text = re.sub(r'[$€£¥,]', '', text)
-                row.append(text)
-            if row:
-                rows.append(row)
-        
-        return {
-            "headers": headers,
-            "rows": rows,
-            "caption": self._get_caption(table),
-            "summary": table.get("summary", ""),
-            "metadata": {
-                "type": "financial",
-                "row_count": len(rows),
-                "column_count": len(headers) or len(rows[0]) if rows else 0
-            }
-        }
-    
-    def _get_caption(self, table):
-        caption = table.xpath(".//caption/text()")
-        return caption[0].strip() if caption else ""
-
-# Usage
-strategy = FinancialTableExtractor(
-    currency_symbols=['$', 'EUR'],
-    require_numbers=True
-)
-
-config = CrawlerRunConfig(
-    table_extraction=strategy
-)
-```
-
-### Example: Specific Table Extractor
-
-```python
-class SpecificTableExtractor(TableExtractionStrategy):
-    """Extract only tables matching specific criteria."""
-    
-    def __init__(self, 
-                 required_headers=None, 
-                 id_pattern=None,
-                 class_pattern=None,
-                 **kwargs):
-        super().__init__(**kwargs)
-        self.required_headers = required_headers or []
-        self.id_pattern = id_pattern
-        self.class_pattern = class_pattern
-    
-    def extract_tables(self, element, **kwargs):
-        tables_data = []
-        
-        for table in element.xpath(".//table"):
-            # Check ID pattern
-            if self.id_pattern:
-                table_id = table.get('id', '')
-                if not re.match(self.id_pattern, table_id):
-                    continue
-            
-            # Check class pattern
-            if self.class_pattern:
-                table_class = table.get('class', '')
-                if not re.match(self.class_pattern, table_class):
-                    continue
-            
-            # Extract headers to check requirements
-            headers = self._extract_headers(table)
-            
-            # Check if required headers are present
-            if self.required_headers:
-                if not all(req in headers for req in self.required_headers):
-                    continue
-            
-            # Extract full table data
-            table_data = self._extract_table_data(table, headers)
-            tables_data.append(table_data)
-        
-        return tables_data
-```
-
-## Combining with Other Strategies
-
-Table extraction works seamlessly with other Crawl4AI strategies:
-
-```python
-from crawl4ai import (
-    AsyncWebCrawler,
-    CrawlerRunConfig,
-    DefaultTableExtraction,
-    LLMExtractionStrategy,
-    JsonCssExtractionStrategy
-)
-
-async def combined_extraction(url):
-    async with AsyncWebCrawler() as crawler:
-        config = CrawlerRunConfig(
-            # Table extraction
-            table_extraction=DefaultTableExtraction(
-                table_score_threshold=6,
-                min_rows=2
-            ),
-            
-            # CSS-based extraction for specific elements
-            extraction_strategy=JsonCssExtractionStrategy({
-                "title": "h1",
-                "summary": "p.summary",
-                "date": "time"
-            }),
-            
-            # Focus on main content
-            css_selector="main.content"
-        )
-        
-        result = await crawler.arun(url, config)
-        
-        # Access different extraction results
-        tables = result.tables  # Table data
-        structured = json.loads(result.extracted_content)  # CSS extraction
-        
-        return {
-            "tables": tables,
-            "structured_data": structured,
-            "markdown": result.markdown
-        }
-```
-
-## Performance Considerations
-
-### Optimization Tips
-
-1. **Disable when not needed**: Use `NoTableExtraction` if tables aren't required
-2. **Target specific areas**: Use `css_selector` to limit processing scope
-3. **Set minimum thresholds**: Filter out small/irrelevant tables early
-4. **Cache results**: Use appropriate cache modes for repeated extractions
-
-```python
-# Optimized configuration for large pages
-config = CrawlerRunConfig(
-    # Only process main content area
-    css_selector="article.main-content",
-    
-    # Exclude navigation and sidebars
-    excluded_tags=["nav", "aside", "footer"],
-    
-    # Higher threshold for stricter filtering
-    table_extraction=DefaultTableExtraction(
-        table_score_threshold=8,
-        min_rows=5,
-        min_cols=3
-    ),
-    
-    # Enable caching for repeated access
-    cache_mode=CacheMode.ENABLED
-)
-```
-
-## Migration Guide
-
-### Important: Your Code Still Works!
-
-**No changes required!** The transition to the strategy pattern is **fully backward compatible**.
-
-### How It Works Internally
-
-#### v0.7.2 and Earlier
-```python
-# Old way - directly passing table_score_threshold
-config = CrawlerRunConfig(
-    table_score_threshold=7
-)
-# Internally: No strategy pattern, direct implementation
-```
-
-#### v0.7.3+ (Current)
-```python
-# Old way STILL WORKS - we handle it internally
-config = CrawlerRunConfig(
-    table_score_threshold=7
-)
-# Internally: Automatically creates DefaultTableExtraction(table_score_threshold=7)
-```
-
-### Taking Advantage of New Features
-
-While your old code works, you can now use the strategy pattern for more control:
-
-```python
-# Option 1: Keep using the old way (perfectly fine!)
-config = CrawlerRunConfig(
-    table_score_threshold=7  # Still supported
-)
-
-# Option 2: Use the new strategy pattern (more flexibility)
-from crawl4ai import DefaultTableExtraction
-
-strategy = DefaultTableExtraction(
-    table_score_threshold=7,
-    min_rows=2,  # New capability!
-    min_cols=2   # New capability!
-)
-
-config = CrawlerRunConfig(
-    table_extraction=strategy
-)
-
-# Option 3: Use advanced strategies when needed
-from crawl4ai import LLMTableExtraction, LLMConfig
-
-# Only for complex tables that DefaultTableExtraction can't handle
-# Automatically handles large tables with smart chunking
-llm_strategy = LLMTableExtraction(
-    llm_config=LLMConfig(
-        provider="groq/llama-3.3-70b-versatile",
-        api_token="your_key"
-    ),
-    max_tries=3,
-    enable_chunking=True,  # Automatically chunk large tables
-    chunk_token_threshold=3000,  # Chunk when exceeding 3000 tokens
-    max_parallel_chunks=5  # Process up to 5 chunks in parallel
-)
-
-config = CrawlerRunConfig(
-    table_extraction=llm_strategy  # Advanced extraction with automatic chunking
-)
-```
-
-### Summary
-
-- ✅ **No breaking changes** - Old code works as-is
-- ✅ **Same defaults** - DefaultTableExtraction is automatically used
-- ✅ **Gradual adoption** - Use new features when you need them
-- ✅ **Full compatibility** - result.tables structure unchanged
-
-## Best Practices
-
-### 1. Choose the Right Strategy (Cost-Conscious Approach)
-
-**Decision Flow**:
-```
-1. Do you need tables? 
-   → No: Use NoTableExtraction
-   → Yes: Continue to #2
-
-2. Try DefaultTableExtraction first (FREE)
-   → Works? Done! ✅
-   → Fails? Continue to #3
-
-3. Is the table critical and complex?
-   → No: Accept DefaultTableExtraction results
-   → Yes: Continue to #4
-
-4. Use LLMTableExtraction (COSTS MONEY)
-   → Small table (<50 rows): Any LLM provider
-   → Large table (50+ rows): Use Groq or Cerebras
-   → Very large (500+ rows): Reconsider - maybe chunk the page
-```
-
-**Strategy Selection Guide**:
-- **DefaultTableExtraction**: Use for 99% of cases - it's free and effective
-- **LLMTableExtraction**: Only for complex tables with merged cells that break DefaultTableExtraction
-- **NoTableExtraction**: When you only need text/markdown content
-- **Custom Strategy**: For specialized requirements (financial, scientific, etc.)
-
-### 2. Validate Extracted Data
-
-```python
-def validate_table(table):
-    """Validate table data quality."""
-    # Check structure
-    if not table.get('rows'):
-        return False
-    
-    # Check consistency
-    if table.get('headers'):
-        expected_cols = len(table['headers'])
-        for row in table['rows']:
-            if len(row) != expected_cols:
-                return False
-    
-    # Check minimum content
-    total_cells = sum(len(row) for row in table['rows'])
-    non_empty = sum(1 for row in table['rows'] 
-                    for cell in row if cell.strip())
-    
-    if non_empty / total_cells < 0.5:  # Less than 50% non-empty
-        return False
-    
-    return True
-
-# Filter valid tables
-valid_tables = [t for t in result.tables if validate_table(t)]
-```
-
-### 3. Handle Edge Cases
-
-```python
-async def robust_table_extraction(url):
-    """Extract tables with error handling."""
-    async with AsyncWebCrawler() as crawler:
-        try:
-            config = CrawlerRunConfig(
-                table_extraction=DefaultTableExtraction(
-                    table_score_threshold=6,
-                    verbose=True
-                )
-            )
-            
-            result = await crawler.arun(url, config)
-            
-            if not result.success:
-                print(f"Crawl failed: {result.error}")
-                return []
-            
-            # Process tables safely
-            processed_tables = []
-            for table in result.tables:
-                try:
-                    # Validate and process
-                    if validate_table(table):
-                        processed_tables.append(table)
-                except Exception as e:
-                    print(f"Error processing table: {e}")
-                    continue
-            
-            return processed_tables
-            
-        except Exception as e:
-            print(f"Extraction error: {e}")
-            return []
-```
-
-## Troubleshooting
-
-### Common Issues and Solutions
-
-| Issue | Cause | Solution |
-|-------|-------|----------|
-| No tables extracted | Score too high | Lower `table_score_threshold` |
-| Layout tables included | Score too low | Increase `table_score_threshold` |
-| Missing tables | CSS selector too specific | Broaden or remove `css_selector` |
-| Incomplete data | Complex table structure | Create custom strategy |
-| Performance issues | Processing entire page | Use `css_selector` to limit scope |
-
-### Debug Logging
-
-Enable verbose logging to understand extraction decisions:
-
-```python
-import logging
-
-# Configure logging
-logging.basicConfig(level=logging.DEBUG)
-
-# Enable verbose mode in strategy
-strategy = DefaultTableExtraction(
-    table_score_threshold=7,
-    verbose=True  # Detailed extraction logs
-)
-
-config = CrawlerRunConfig(
-    table_extraction=strategy,
-    verbose=True  # General crawler logs
-)
-```
-
-## See Also
-
-- [Extraction Strategies](extraction-strategies.md) - Overview of all extraction strategies
-- [Content Selection](content-selection.md) - Using CSS selectors and filters
-- [Performance Optimization](../optimization/performance-tuning.md) - Speed up extraction
-- [Examples](../examples/table_extraction_example.py) - Complete working examples

docs/md_v2/migration/table_extraction_v073.md

@@ -1,376 +0,0 @@
-# Migration Guide: Table Extraction v0.7.3
-
-## Overview
-
-Version 0.7.3 introduces the **Table Extraction Strategy Pattern**, providing a more flexible and extensible approach to table extraction while maintaining full backward compatibility.
-
-## What's New
-
-### Strategy Pattern Implementation
-
-Table extraction now follows the same strategy pattern used throughout Crawl4AI:
-
-- **Consistent Architecture**: Aligns with extraction, chunking, and markdown strategies
-- **Extensibility**: Easy to create custom table extraction strategies
-- **Better Separation**: Table logic moved from content scraping to dedicated module
-- **Full Control**: Fine-grained control over table detection and extraction
-
-### New Classes
-
-```python
-from crawl4ai import (
-    TableExtractionStrategy,    # Abstract base class
-    DefaultTableExtraction,      # Current implementation (default)
-    NoTableExtraction           # Explicitly disable extraction
-)
-```
-
-## Backward Compatibility
-
-**✅ All existing code continues to work without changes.**
-
-### No Changes Required
-
-If your code looks like this, it will continue to work:
-
-```python
-# This still works exactly the same
-config = CrawlerRunConfig(
-    table_score_threshold=7
-)
-result = await crawler.arun(url, config)
-tables = result.tables  # Same structure, same data
-```
-
-### What Happens Behind the Scenes
-
-When you don't specify a `table_extraction` strategy:
-
-1. `CrawlerRunConfig` automatically creates `DefaultTableExtraction`
-2. It uses your `table_score_threshold` parameter
-3. Tables are extracted exactly as before
-4. Results appear in `result.tables` with the same structure
-
-## New Capabilities
-
-### 1. Explicit Strategy Configuration
-
-You can now explicitly configure table extraction:
-
-```python
-# New: Explicit control
-strategy = DefaultTableExtraction(
-    table_score_threshold=7,
-    min_rows=2,              # New: minimum row filter
-    min_cols=2,              # New: minimum column filter
-    verbose=True             # New: detailed logging
-)
-
-config = CrawlerRunConfig(
-    table_extraction=strategy
-)
-```
-
-### 2. Disable Table Extraction
-
-Improve performance when tables aren't needed:
-
-```python
-# New: Skip table extraction entirely
-config = CrawlerRunConfig(
-    table_extraction=NoTableExtraction()
-)
-# No CPU cycles spent on table detection/extraction
-```
-
-### 3. Custom Extraction Strategies
-
-Create specialized extractors:
-
-```python
-class MyTableExtractor(TableExtractionStrategy):
-    def extract_tables(self, element, **kwargs):
-        # Custom extraction logic
-        return custom_tables
-
-config = CrawlerRunConfig(
-    table_extraction=MyTableExtractor()
-)
-```
-
-## Migration Scenarios
-
-### Scenario 1: Basic Usage (No Changes Needed)
-
-**Before (v0.7.2):**
-```python
-config = CrawlerRunConfig()
-result = await crawler.arun(url, config)
-for table in result.tables:
-    print(table['headers'])
-```
-
-**After (v0.7.3):**
-```python
-# Exactly the same - no changes required
-config = CrawlerRunConfig()
-result = await crawler.arun(url, config)
-for table in result.tables:
-    print(table['headers'])
-```
-
-### Scenario 2: Custom Threshold (No Changes Needed)
-
-**Before (v0.7.2):**
-```python
-config = CrawlerRunConfig(
-    table_score_threshold=5
-)
-```
-
-**After (v0.7.3):**
-```python
-# Still works the same
-config = CrawlerRunConfig(
-    table_score_threshold=5
-)
-
-# Or use new explicit approach for more control
-strategy = DefaultTableExtraction(
-    table_score_threshold=5,
-    min_rows=2  # Additional filtering
-)
-config = CrawlerRunConfig(
-    table_extraction=strategy
-)
-```
-
-### Scenario 3: Advanced Filtering (New Feature)
-
-**Before (v0.7.2):**
-```python
-# Had to filter after extraction
-config = CrawlerRunConfig(
-    table_score_threshold=5
-)
-result = await crawler.arun(url, config)
-
-# Manual filtering
-large_tables = [
-    t for t in result.tables 
-    if len(t['rows']) >= 5 and len(t['headers']) >= 3
-]
-```
-
-**After (v0.7.3):**
-```python
-# Filter during extraction (more efficient)
-strategy = DefaultTableExtraction(
-    table_score_threshold=5,
-    min_rows=5,
-    min_cols=3
-)
-config = CrawlerRunConfig(
-    table_extraction=strategy
-)
-result = await crawler.arun(url, config)
-# result.tables already filtered
-```
-
-## Code Organization Changes
-
-### Module Structure
-
-**Before (v0.7.2):**
-```
-crawl4ai/
-  content_scraping_strategy.py
-    - LXMLWebScrapingStrategy
-      - is_data_table()      # Table detection
-      - extract_table_data() # Table extraction
-```
-
-**After (v0.7.3):**
-```
-crawl4ai/
-  content_scraping_strategy.py
-    - LXMLWebScrapingStrategy
-      # Table methods removed, uses strategy
-  
-  table_extraction.py (NEW)
-    - TableExtractionStrategy    # Base class
-    - DefaultTableExtraction      # Moved logic here
-    - NoTableExtraction          # New option
-```
-
-### Import Changes
-
-**New imports available (optional):**
-```python
-# These are now available but not required for existing code
-from crawl4ai import (
-    TableExtractionStrategy,
-    DefaultTableExtraction,
-    NoTableExtraction
-)
-```
-
-## Performance Implications
-
-### No Performance Impact
-
-For existing code, performance remains identical:
-- Same extraction logic
-- Same scoring algorithm
-- Same processing time
-
-### Performance Improvements Available
-
-New options for better performance:
-
-```python
-# Skip tables entirely (faster)
-config = CrawlerRunConfig(
-    table_extraction=NoTableExtraction()
-)
-
-# Process only specific areas (faster)
-config = CrawlerRunConfig(
-    css_selector="main.content",
-    table_extraction=DefaultTableExtraction(
-        min_rows=5,  # Skip small tables
-        min_cols=3
-    )
-)
-```
-
-## Testing Your Migration
-
-### Verification Script
-
-Run this to verify your extraction still works:
-
-```python
-import asyncio
-from crawl4ai import AsyncWebCrawler, CrawlerRunConfig
-
-async def verify_extraction():
-    url = "your_url_here"
-    
-    async with AsyncWebCrawler() as crawler:
-        # Test 1: Old approach
-        config_old = CrawlerRunConfig(
-            table_score_threshold=7
-        )
-        result_old = await crawler.arun(url, config_old)
-        
-        # Test 2: New explicit approach
-        from crawl4ai import DefaultTableExtraction
-        config_new = CrawlerRunConfig(
-            table_extraction=DefaultTableExtraction(
-                table_score_threshold=7
-            )
-        )
-        result_new = await crawler.arun(url, config_new)
-        
-        # Compare results
-        assert len(result_old.tables) == len(result_new.tables)
-        print(f"✓ Both approaches extracted {len(result_old.tables)} tables")
-        
-        # Verify structure
-        for old, new in zip(result_old.tables, result_new.tables):
-            assert old['headers'] == new['headers']
-            assert old['rows'] == new['rows']
-        
-        print("✓ Table content identical")
-
-asyncio.run(verify_extraction())
-```
-
-## Deprecation Notes
-
-### No Deprecations
-
-- All existing parameters continue to work
-- `table_score_threshold` in `CrawlerRunConfig` is still supported
-- No breaking changes
-
-### Internal Changes (Transparent to Users)
-
-- `LXMLWebScrapingStrategy.is_data_table()` - Moved to `DefaultTableExtraction`
-- `LXMLWebScrapingStrategy.extract_table_data()` - Moved to `DefaultTableExtraction`
-
-These methods were internal and not part of the public API.
-
-## Benefits of Upgrading
-
-While not required, using the new pattern provides:
-
-1. **Better Control**: Filter tables during extraction, not after
-2. **Performance Options**: Skip extraction when not needed
-3. **Extensibility**: Create custom extractors for specific needs
-4. **Consistency**: Same pattern as other Crawl4AI strategies
-5. **Future-Proof**: Ready for upcoming advanced strategies
-
-## Troubleshooting
-
-### Issue: Different Number of Tables
-
-**Cause**: Threshold or filtering differences
-
-**Solution**: 
-```python
-# Ensure same threshold
-strategy = DefaultTableExtraction(
-    table_score_threshold=7,  # Match your old setting
-    min_rows=0,               # No filtering (default)
-    min_cols=0                # No filtering (default)
-)
-```
-
-### Issue: Import Errors
-
-**Cause**: Using new classes without importing
-
-**Solution**:
-```python
-# Add imports if using new features
-from crawl4ai import (
-    DefaultTableExtraction,
-    NoTableExtraction,
-    TableExtractionStrategy
-)
-```
-
-### Issue: Custom Strategy Not Working
-
-**Cause**: Incorrect method signature
-
-**Solution**:
-```python
-class CustomExtractor(TableExtractionStrategy):
-    def extract_tables(self, element, **kwargs):  # Correct signature
-        # Not: extract_tables(self, html)
-        # Not: extract(self, element)
-        return tables_list
-```
-
-## Getting Help
-
-If you encounter issues:
-
-1. Check your `table_score_threshold` matches previous settings
-2. Verify imports if using new classes
-3. Enable verbose logging: `DefaultTableExtraction(verbose=True)`
-4. Review the [Table Extraction Documentation](../core/table_extraction.md)
-5. Check [examples](../examples/table_extraction_example.py)
-
-## Summary
-
-- ✅ **Full backward compatibility** - No code changes required
-- ✅ **Same results** - Identical extraction behavior by default
-- ✅ **New options** - Additional control when needed
-- ✅ **Better architecture** - Consistent with Crawl4AI patterns
-- ✅ **Ready for future** - Foundation for advanced strategies
-
-The migration to v0.7.3 is seamless with no required changes while providing new capabilities for those who need them.

tests/test_llm_simple_url.py

@@ -1,170 +0,0 @@
-#!/usr/bin/env python3
-"""
-Test LLMTableExtraction with controlled HTML
-"""
-
-import os
-import sys
-sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
-
-import asyncio
-from crawl4ai import (
-    AsyncWebCrawler,
-    CrawlerRunConfig,
-    LLMConfig,
-    LLMTableExtraction,
-    DefaultTableExtraction,
-    CacheMode
-)
-
-async def test_controlled_html():
-    """Test with controlled HTML content."""
-    print("\n" + "=" * 60)
-    print("LLM TABLE EXTRACTION TEST")
-    print("=" * 60)
-    
-    url = "https://en.wikipedia.org/wiki/List_of_chemical_elements"
-    # url = "https://en.wikipedia.org/wiki/List_of_prime_ministers_of_India"
-    
-    # Configure LLM
-    llm_config = LLMConfig(
-        # provider="openai/gpt-4.1-mini",
-        # api_token=os.getenv("OPENAI_API_KEY"),
-        provider="groq/llama-3.3-70b-versatile",
-        api_token="GROQ_API_TOKEN",
-        temperature=0.1,
-        max_tokens=32000
-    )
-    
-    print("\n1. Testing LLMTableExtraction:")
-    
-    # Create LLM extraction strategy
-    llm_strategy = LLMTableExtraction(
-        llm_config=llm_config,
-        verbose=True,
-        # css_selector="div.w3-example"
-        css_selector="div.mw-content-ltr",
-        # css_selector="table.wikitable",
-        max_tries=2,
-        
-        enable_chunking=True,
-        chunk_token_threshold=5000,  # Lower threshold to force chunking
-        min_rows_per_chunk=10,
-        max_parallel_chunks=3
-    )
-    
-    config_llm = CrawlerRunConfig(
-        cache_mode=CacheMode.BYPASS,
-        table_extraction=llm_strategy
-    )
-    
-    async with AsyncWebCrawler() as crawler:
-        # Test with LLM extraction
-        result_llm = await crawler.arun(
-            # url=f"raw:{test_html}",
-            url=url,
-            config=config_llm
-        )
-        
-        if result_llm.success:
-            print(f"\n   ✓ LLM Extraction: Found {len(result_llm.tables)} table(s)")
-            
-            for i, table in enumerate(result_llm.tables, 1):
-                print(f"\n   Table {i}:")
-                print(f"   - Caption: {table.get('caption', 'No caption')}")
-                print(f"   - Headers: {table['headers']}")
-                print(f"   - Rows: {len(table['rows'])}")
-                
-                # Show how colspan/rowspan were handled
-                print(f"   - Sample rows:")
-                for j, row in enumerate(table['rows'][:2], 1):
-                    print(f"     Row {j}: {row}")
-                
-                metadata = table.get('metadata', {})
-                print(f"   - Metadata:")
-                print(f"     • Has merged cells: {metadata.get('has_merged_cells', False)}")
-                print(f"     • Table type: {metadata.get('table_type', 'unknown')}")
-        
-        # # Compare with default extraction
-        # print("\n2. Comparing with DefaultTableExtraction:")
-        
-        # default_strategy = DefaultTableExtraction(
-        #     table_score_threshold=3,
-        #     verbose=False
-        # )
-        
-        # config_default = CrawlerRunConfig(
-        #     cache_mode=CacheMode.BYPASS,
-        #     table_extraction=default_strategy
-        # )
-        
-        # result_default = await crawler.arun(
-        #     # url=f"raw:{test_html}",
-        #     url=url,
-        #     config=config_default
-        # )
-        
-        # if result_default.success:
-        #     print(f"   ✓ Default Extraction: Found {len(result_default.tables)} table(s)")
-            
-        #     # Compare handling of complex structures
-        #     print("\n3. Comparison Summary:")
-        #     print(f"   LLM found: {len(result_llm.tables)} tables")
-        #     print(f"   Default found: {len(result_default.tables)} tables")
-            
-        #     if result_llm.tables and result_default.tables:
-        #         llm_first = result_llm.tables[0]
-        #         default_first = result_default.tables[0]
-                
-        #         print(f"\n   First table comparison:")
-        #         print(f"   LLM headers: {len(llm_first['headers'])} columns")
-        #         print(f"   Default headers: {len(default_first['headers'])} columns")
-                
-        #         # Check if LLM better handled the complex structure
-        #         if llm_first.get('metadata', {}).get('has_merged_cells'):
-        #             print("   ✓ LLM correctly identified merged cells")
-                
-        #         # Test pandas compatibility
-        #         try:
-        #             import pandas as pd
-                    
-        #             print("\n4. Testing Pandas compatibility:")
-                    
-        #             # Create DataFrame from LLM extraction
-        #             df_llm = pd.DataFrame(
-        #                 llm_first['rows'],
-        #                 columns=llm_first['headers']
-        #             )
-        #             print(f"   ✓ LLM table -> DataFrame: Shape {df_llm.shape}")
-                    
-        #             # Create DataFrame from default extraction
-        #             df_default = pd.DataFrame(
-        #                 default_first['rows'],
-        #                 columns=default_first['headers']
-        #             )
-        #             print(f"   ✓ Default table -> DataFrame: Shape {df_default.shape}")
-                    
-        #             print("\n   LLM DataFrame preview:")
-        #             print(df_llm.head(2).to_string())
-                    
-        #         except ImportError:
-        #             print("\n4. Pandas not installed, skipping DataFrame test")
-        
-        print("\n✅ Test completed successfully!")
-
-async def main():
-    """Run the test."""
-    
-    # Check for API key
-    if not os.getenv("OPENAI_API_KEY"):
-        print("⚠️  OPENAI_API_KEY not set. Please set it to test LLM extraction.")
-        print("   You can set it with: export OPENAI_API_KEY='your-key-here'")
-        return
-    
-    await test_controlled_html()
-
-if __name__ == "__main__":
-    asyncio.run(main())
-    
-    
-    

tests/test_memory_macos.py

@@ -4,7 +4,7 @@
 import psutil
 import platform
 import time
-from crawl4ai.utils import get_true_memory_usage_percent, get_memory_stats, get_true_available_memory_gb
+from crawl4ai.memory_utils import get_true_memory_usage_percent, get_memory_stats, get_true_available_memory_gb
 
 
 def test_memory_calculation():