moved score threshold to config.py & replaced the separator for tag.get_text in find_closest_parent_with_useful_text fn from period(.) to space( ) to keep the text more neutral.

These changes fix typos in `chunking_strategy.py` and `crawler_strategy.py` to improve code readability. Additionally, the `.test_pads/` directory is removed from the `.gitignore` file to keep the repository clean and organized.

.gitignore

@@ -165,6 +165,8 @@ Crawl4AI.egg-info/
 Crawl4AI.egg-info/*
 crawler_data.db
 .vscode/
+.tests/
+.test_pads/
 test_pad.py
 test_pad*.py
 .data/

CHANGELOG.md

@@ -1,5 +1,14 @@
 # Changelog
 
+## [v0.2.75] - 2024-07-19
+
+Minor improvements for a more maintainable codebase:
+
+- 🔄 Fixed typos in `chunking_strategy.py` and `crawler_strategy.py` to improve code readability
+- 🔄 Removed `.test_pads/` directory from `.gitignore` to keep our repository clean and organized
+
+These changes may seem small, but they contribute to a more stable and sustainable codebase. By fixing typos and updating our `.gitignore` settings, we're ensuring that our code is easier to maintain and scale in the long run.
+
 ## [v0.2.74] - 2024-07-08
 A slew of exciting updates to improve the crawler's stability and robustness! 🎉
 

README.md

@@ -1,4 +1,4 @@
-# Crawl4AI v0.2.74 🕷️🤖
+# Crawl4AI v0.2.75 🕷️🤖
 
 [![GitHub Stars](https://img.shields.io/github/stars/unclecode/crawl4ai?style=social)](https://github.com/unclecode/crawl4ai/stargazers)
 [![GitHub Forks](https://img.shields.io/github/forks/unclecode/crawl4ai?style=social)](https://github.com/unclecode/crawl4ai/network/members)

crawl4ai/chunking_strategy.py

@@ -55,7 +55,7 @@ class TopicSegmentationChunking(ChunkingStrategy):
     
     def __init__(self, num_keywords=3, **kwargs):
         import nltk as nl
-        self.tokenizer = nl.toknize.TextTilingTokenizer()
+        self.tokenizer = nl.tokenize.TextTilingTokenizer()
         self.num_keywords = num_keywords
 
     def chunk(self, text: str) -> list:

crawl4ai/config.py

@@ -27,3 +27,13 @@ WORD_TOKEN_RATE = 1.3
 
 # Threshold for the minimum number of word in a HTML tag to be considered 
 MIN_WORD_THRESHOLD = 1
+
+# Threshold for the Image extraction - Range is 1 to 6
+# Images are scored based on point based system, to filter based on usefulness. Points are assigned
+# to each image based on the following aspects.
+# If either height or width exceeds 150px
+# If image size is greater than 10Kb
+# If alt property is set
+# If image format is in jpg, png or webp
+# If image is in the first half of the total images extracted from the page
+IMAGE_SCORE_THRESHOLD = 2

crawl4ai/crawler_strategy.py

@@ -292,15 +292,22 @@ class LocalSeleniumCrawlerStrategy(CrawlerStrategy):
             # Open the screenshot with PIL
             image = Image.open(BytesIO(screenshot))
 
+            # Convert image to RGB mode
+            rgb_image = image.convert('RGB')
+
             # Convert to JPEG and compress
             buffered = BytesIO()
-            image.save(buffered, format="JPEG", quality=85)
+            rgb_image.save(buffered, format="JPEG", quality=85)
             img_base64 = base64.b64encode(buffered.getvalue()).decode('utf-8')
 
             if self.verbose:
                 print(f"[LOG] 📸 Screenshot taken and converted to base64")
 
             return img_base64
+        except Exception as e:
+            if self.verbose:
+                print(f"[ERROR] Failed to take screenshot: {str(e)}")
+            return ""
 
         except Exception as e:
             error_message = sanitize_input_encode(f"Failed to take screenshot: {str(e)}")

crawl4ai/model_loader.py

@@ -3,7 +3,7 @@ from pathlib import Path
 import subprocess, os
 import shutil
 import tarfile
-from crawl4ai.config import MODEL_REPO_BRANCH
+from .model_loader import *
 import argparse
 import urllib.request
 __location__ = os.path.realpath(os.path.join(os.getcwd(), os.path.dirname(__file__)))

crawl4ai/utils.py

@@ -11,6 +11,9 @@ from .prompts import PROMPT_EXTRACT_BLOCKS
 from .config import *
 from pathlib import Path
 from typing import Dict, Any
+from urllib.parse import urljoin
+import requests
+from requests.exceptions import InvalidSchema
 
 class InvalidCSSSelectorError(Exception):
     pass
@@ -447,6 +450,101 @@ def get_content_of_website_optimized(url: str, html: str, word_count_threshold:
     links = {'internal': [], 'external': []}
     media = {'images': [], 'videos': [], 'audios': []}
 
+    def process_image(img, url, index, total_images):
+            #Check if an image has valid display and inside undesired html elements
+            def is_valid_image(img, parent, parent_classes):
+                style = img.get('style', '')
+                src = img.get('src', '')
+                classes_to_check = ['button', 'icon', 'logo']
+                tags_to_check = ['button', 'input']
+                return all([
+                    'display:none' not in style,
+                    src,
+                    not any(s in var for var in [src, img.get('alt', ''), *parent_classes] for s in classes_to_check),
+                    parent.name not in tags_to_check
+                ])
+
+            #Score an image for it's usefulness
+            def score_image_for_usefulness(img, base_url, index, images_count):
+                # Function to parse image height/width value and units
+                def parse_dimension(dimension):
+                    if dimension:
+                        match = re.match(r"(\d+)(\D*)", dimension)
+                        if match:
+                            number = int(match.group(1))
+                            unit = match.group(2) or 'px'  # Default unit is 'px' if not specified
+                            return number, unit
+                    return None, None
+
+                # Fetch image file metadata to extract size and extension
+                def fetch_image_file_size(img, base_url):
+                    #If src is relative path construct full URL, if not it may be CDN URL
+                    img_url = urljoin(base_url,img.get('src'))
+                    try:
+                        response = requests.head(img_url)
+                        if response.status_code == 200:
+                            return response.headers.get('Content-Length',None)
+                        else:
+                            print(f"Failed to retrieve file size for {img_url}")
+                            return None
+                    except InvalidSchema as e:
+                        return None
+                    finally:
+                        return
+
+                image_height = img.get('height')
+                height_value, height_unit = parse_dimension(image_height)
+                image_width =  img.get('width')
+                width_value, width_unit = parse_dimension(image_width)
+                image_size = int(fetch_image_file_size(img,base_url) or 0)
+                image_format = os.path.splitext(img.get('src',''))[1].lower()
+                score = 0
+                if height_value:
+                    if height_unit == 'px' and height_value > 150:
+                        score += 1
+                    if height_unit in ['%','vh','vmin','vmax'] and height_value >30:
+                        score += 1
+                if width_value:
+                    if width_unit == 'px' and width_value > 150:
+                        score += 1
+                    if width_unit in ['%','vh','vmin','vmax'] and width_value >30:
+                        score += 1
+                if image_size > 10000:
+                    score += 1
+                if img.get('alt') != '':
+                    score+=1
+                if any(image_format==format for format in ['jpg','png','webp']):
+                    score+=1
+                if index/images_count<0.5:
+                    score+=1
+                return score
+
+            # Extract meaningful text for images from closest parent
+            def find_closest_parent_with_useful_text(tag):
+                current_tag = tag
+                while current_tag:
+                    current_tag = current_tag.parent
+                    # Get the text content of the parent tag
+                    if current_tag:
+                        text_content = current_tag.get_text(separator=' ',strip=True)
+                        # Check if the text content has at least word_count_threshold
+                        if len(text_content.split()) >= word_count_threshold:
+                            return text_content
+                return None
+
+            if not is_valid_image(img, img.parent, img.parent.get('class', [])):
+                return None
+            score = score_image_for_usefulness(img, url, index, total_images)
+            if score <= IMAGE_SCORE_THRESHOLD:
+                return None
+            return {
+                'src': img.get('src', ''),
+                'alt': img.get('alt', ''),
+                'desc': find_closest_parent_with_useful_text(img),
+                'score': score,
+                'type': 'image'
+            }
+
     def process_element(element: element.PageElement) -> bool:
         try:
             if isinstance(element, NavigableString):
@@ -471,11 +569,6 @@ def get_content_of_website_optimized(url: str, html: str, word_count_threshold:
                 keep_element = True
 
             elif element.name == 'img':
-                media['images'].append({
-                    'src': element.get('src'),
-                    'alt': element.get('alt'),
-                    'type': 'image'
-                })
                 return True  # Always keep image elements
 
             elif element.name in ['video', 'audio']:
@@ -518,6 +611,14 @@ def get_content_of_website_optimized(url: str, html: str, word_count_threshold:
             print('Error processing element:', str(e))
             return False
 
+    #process images by filtering and extracting contextual text from the page
+    imgs = body.find_all('img')
+    media['images'] = [
+        result for result in
+        (process_image(img, url, i, len(imgs)) for i, img in enumerate(imgs))
+        if result is not None
+    ]
+
     process_element(body)
 
     def flatten_nested_elements(node):

docs/md/changelog.md

@@ -1,5 +1,15 @@
 # Changelog
 
+## [v0.2.75] - 2024-07-19
+
+Minor improvements for a more maintainable codebase:
+
+- 🔄 Fixed typos in `chunking_strategy.py` and `crawler_strategy.py` to improve code readability
+- 🔄 Removed `.test_pads/` directory from `.gitignore` to keep our repository clean and organized
+
+These changes may seem small, but they contribute to a more stable and sustainable codebase. By fixing typos and updating our `.gitignore` settings, we're ensuring that our code is easier to maintain and scale in the long run.
+
+
 ## v0.2.74 - 2024-07-08
 A slew of exciting updates to improve the crawler's stability and robustness! 🎉
 

docs/md/demo.md

@@ -14,6 +14,7 @@
             <div class="form-group">
                 <button class="btn btn-default" type="submit">Submit</button>
             </div>
+
         </fieldset>
     </form>
 
@@ -93,6 +94,10 @@
         </div>
     </section>
 
+    <div id="error" class="error-message" style="display: none; margin-top:1em;">
+        <div class="terminal-alert terminal-alert-error"></div>
+    </div>
+
     <script>
         function showTab(tabId) {
             const tabs = document.querySelectorAll('.tab-content');
@@ -162,7 +167,17 @@
                 },
                 body: JSON.stringify(data)
             })
-            .then(response => response.json())
+            .then(response => {
+                if (!response.ok) {
+                    if (response.status === 429) {
+                        return response.json().then(err => { 
+                            throw Object.assign(new Error('Rate limit exceeded'), { status: 429, details: err });
+                        });
+                    }
+                    throw new Error('Network response was not ok');
+                }
+                return response.json();
+            })
             .then(data => {
                 data = data.results[0]; // Only one URL is requested
                 document.getElementById('loading').style.display = 'none';
@@ -187,11 +202,29 @@ result = crawler.run(
 print(result)
                 `;
                 redo(document.getElementById('pythonCode'), pythonCode);
+                document.getElementById('error').style.display = 'none';
             })
             .catch(error => {
                 document.getElementById('loading').style.display = 'none';
-                document.getElementById('response').style.display = 'block';
-                document.getElementById('markdownContent').textContent = 'Error: ' + error;
+                document.getElementById('error').style.display = 'block';
+                let errorMessage = 'An unexpected error occurred. Please try again later.';
+                
+                if (error.status === 429) {
+                    const details = error.details;
+                    if (details.retry_after) {
+                        errorMessage = `Rate limit exceeded. Please wait ${parseFloat(details.retry_after).toFixed(1)} seconds before trying again.`;
+                    } else if (details.reset_at) {
+                        const resetTime = new Date(details.reset_at);
+                        const waitTime = Math.ceil((resetTime - new Date()) / 1000);
+                        errorMessage = `Rate limit exceeded. Please try again after ${waitTime} seconds.`;
+                    } else {
+                        errorMessage = `Rate limit exceeded. Please try again later.`;
+                    }
+                } else if (error.message) {
+                    errorMessage = error.message;
+                }
+                
+                document.querySelector('#error .terminal-alert').textContent = errorMessage;
             });
         });
     </script>

docs/md/index.md

@@ -1,4 +1,4 @@
-# Crawl4AI v0.2.74
+# Crawl4AI v0.2.75
 
 Welcome to the official documentation for Crawl4AI! 🕷️🤖 Crawl4AI is an open-source Python library designed to simplify web crawling and extract useful information from web pages. This documentation will guide you through the features, usage, and customization of Crawl4AI.
 

main.py

@@ -22,6 +22,15 @@ from typing import List, Optional
 from crawl4ai.web_crawler import WebCrawler
 from crawl4ai.database import get_total_count, clear_db
 
+import time
+from slowapi import Limiter, _rate_limit_exceeded_handler
+from slowapi.util import get_remote_address
+from slowapi.errors import RateLimitExceeded
+
+# load .env file
+from dotenv import load_dotenv
+load_dotenv()
+
 # Configuration
 __location__ = os.path.realpath(os.path.join(os.getcwd(), os.path.dirname(__file__)))
 MAX_CONCURRENT_REQUESTS = 10  # Adjust this to change the maximum concurrent requests
@@ -30,6 +39,78 @@ lock = asyncio.Lock()
 
 app = FastAPI()
 
+# Initialize rate limiter
+def rate_limit_key_func(request: Request):
+    access_token = request.headers.get("access-token")
+    if access_token == os.environ.get('ACCESS_TOKEN'):
+        return None
+    return get_remote_address(request)
+
+limiter = Limiter(key_func=rate_limit_key_func)
+app.state.limiter = limiter
+
+# Dictionary to store last request times for each client
+last_request_times = {}
+last_rate_limit = {}
+
+
+def get_rate_limit():
+    limit = os.environ.get('ACCESS_PER_MIN', "5")
+    return f"{limit}/minute"
+
+# Custom rate limit exceeded handler
+async def custom_rate_limit_exceeded_handler(request: Request, exc: RateLimitExceeded) -> JSONResponse:
+    if request.client.host not in last_rate_limit or time.time() - last_rate_limit[request.client.host] > 60:
+        last_rate_limit[request.client.host] = time.time()
+    retry_after = 60 - (time.time() - last_rate_limit[request.client.host])
+    reset_at = time.time() + retry_after
+    return JSONResponse(
+        status_code=429,
+        content={
+            "detail": "Rate limit exceeded",
+            "limit": str(exc.limit.limit),
+            "retry_after": retry_after,
+            'reset_at': reset_at,
+            "message": f"You have exceeded the rate limit of {exc.limit.limit}."
+        }
+    )
+    
+app.add_exception_handler(RateLimitExceeded, custom_rate_limit_exceeded_handler)
+
+
+# Middleware for token-based bypass and per-request limit
+class RateLimitMiddleware(BaseHTTPMiddleware):
+    async def dispatch(self, request: Request, call_next):
+        SPAN = int(os.environ.get('ACCESS_TIME_SPAN', 10))
+        access_token = request.headers.get("access-token")
+        if access_token == os.environ.get('ACCESS_TOKEN'):
+            return await call_next(request)
+        
+        path = request.url.path
+        if path in ["/crawl", "/old"]:
+            client_ip = request.client.host
+            current_time = time.time()
+            
+            # Check time since last request
+            if client_ip in last_request_times:
+                time_since_last_request = current_time - last_request_times[client_ip]
+                if time_since_last_request < SPAN:
+                    return JSONResponse(
+                        status_code=429,
+                        content={
+                            "detail": "Too many requests",
+                            "message": "Rate limit exceeded. Please wait 10 seconds between requests.",
+                            "retry_after": max(0, SPAN - time_since_last_request),
+                            "reset_at": current_time + max(0, SPAN - time_since_last_request),
+                        }
+                    )
+            
+            last_request_times[client_ip] = current_time
+
+        return await call_next(request)
+
+app.add_middleware(RateLimitMiddleware)
+
 # CORS configuration
 origins = ["*"]  # Allow all origins
 app.add_middleware(
@@ -73,6 +154,7 @@ def read_root():
     return RedirectResponse(url="/mkdocs")
 
 @app.get("/old", response_class=HTMLResponse)
+@limiter.limit(get_rate_limit())
 async def read_index(request: Request):
     partials_dir = os.path.join(__location__, "pages", "partial")
     partials = {}
@@ -107,6 +189,7 @@ def import_strategy(module_name: str, class_name: str, *args, **kwargs):
         raise HTTPException(status_code=400, detail=f"Class {class_name} not found in {module_name}.")
 
 @app.post("/crawl")
+@limiter.limit(get_rate_limit())
 async def crawl_urls(crawl_request: CrawlRequest, request: Request):
     logging.debug(f"[LOG] Crawl request for URL: {crawl_request.urls}")
     global current_requests

setup.py

@@ -1,18 +1,18 @@
 from setuptools import setup, find_packages
 import os
 from pathlib import Path
-import subprocess
-from setuptools.command.install import install
+import shutil
 
 # Create the .crawl4ai folder in the user's home directory if it doesn't exist
 # If the folder already exists, remove the cache folder
-crawl4ai_folder = os.path.join(Path.home(), ".crawl4ai")
-if os.path.exists(f"{crawl4ai_folder}/cache"):
-    subprocess.run(["rm", "-rf", f"{crawl4ai_folder}/cache"])
-os.makedirs(crawl4ai_folder, exist_ok=True)
-os.makedirs(f"{crawl4ai_folder}/cache", exist_ok=True)
+crawl4ai_folder = Path.home() / ".crawl4ai"
+cache_folder = crawl4ai_folder / "cache"
 
+if cache_folder.exists():
+    shutil.rmtree(cache_folder)
 
+crawl4ai_folder.mkdir(exist_ok=True)
+cache_folder.mkdir(exist_ok=True)
 
 # Read the requirements from requirements.txt
 with open("requirements.txt") as f: