ayrisdev/crawl4ai
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: Add hooks utility for function-based hooks with Docker client integration. ref #1377

Browse Source 
Add hooks_to_string() utility function that converts Python function objects
   to string representations for the Docker API, enabling developers to write hooks
   as regular Python functions instead of strings.

   Core Changes:
   - New hooks_to_string() utility in crawl4ai/utils.py using inspect.getsource()
   - Docker client now accepts both function objects and strings for hooks
   - Automatic detection and conversion in Crawl4aiDockerClient._prepare_request()
   - New hooks and hooks_timeout parameters in client.crawl() method

   Documentation:
   - Docker client examples with function-based hooks (docs/examples/docker_client_hooks_example.py)
   - Updated main Docker deployment guide with comprehensive hooks section
   - Added unit tests for hooks utility (tests/docker/test_hooks_utility.py)
This commit is contained in:
ntohidi
2025-10-13 12:53:33 +08:00
parent 7dadb65b80
commit 4a04b8506a
3 changed files with 269 additions and 80 deletions
Download Patch File Download Diff File Expand all files Collapse all files

82
docs/blog/release-v0.7.5.md
View File

@@ -8,7 +8,8 @@ Today I'm releasing Crawl4AI v0.7.5—focused on extensibility and security. Thi
## 🎯 What's New at a Glance
- **Docker Hooks System**: Custom Python functions at key pipeline points
- **Docker Hooks System**: Custom Python functions at key pipeline points with function-based API
- **Function-Based Hooks**: New `hooks_to_string()` utility with Docker client auto-conversion
- **Enhanced LLM Integration**: Custom providers with temperature control
- **HTTPS Preservation**: Secure internal link handling
- **Bug Fixes**: Resolved multiple community-reported issues
@@ -82,6 +83,85 @@ if result.get('success'):
- `before_retrieve_html`: Pre-extraction processing
- `before_return_html`: Final HTML processing
### Function-Based Hooks API
Writing hooks as strings works, but lacks IDE support and type checking. v0.7.5 introduces a function-based approach with automatic conversion!
**Option 1: Using the `hooks_to_string()` Utility**
```python
from crawl4ai import hooks_to_string
import requests
# Define hooks as regular Python functions (with full IDE support!)
async def on_page_context_created(page, context, **kwargs):
"""Block images to speed up crawling"""
await context.route("**/*.{png,jpg,jpeg,gif,webp}", lambda route: route.abort())
await page.set_viewport_size({"width": 1920, "height": 1080})
return page
async def before_goto(page, context, url, **kwargs):
"""Add custom headers"""
await page.set_extra_http_headers({
'X-Crawl4AI': 'v0.7.5',
'X-Custom-Header': 'my-value'
})
return page
# Convert functions to strings
hooks_code = hooks_to_string({
"on_page_context_created": on_page_context_created,
"before_goto": before_goto
})
# Use with REST API
payload = {
"urls": ["https://httpbin.org/html"],
"hooks": {"code": hooks_code, "timeout": 30}
}
response = requests.post("http://localhost:11235/crawl", json=payload)
```
**Option 2: Docker Client with Automatic Conversion (Recommended!)**
```python
from crawl4ai.docker_client import Crawl4aiDockerClient
# Define hooks as functions (same as above)
async def on_page_context_created(page, context, **kwargs):
await context.route("**/*.{png,jpg,jpeg,gif,webp}", lambda route: route.abort())
return page
async def before_retrieve_html(page, context, **kwargs):
# Scroll to load lazy content
await page.evaluate("window.scrollTo(0, document.body.scrollHeight)")
await page.wait_for_timeout(1000)
return page
# Use Docker client - conversion happens automatically!
client = Crawl4aiDockerClient(base_url="http://localhost:11235")
results = await client.crawl(
urls=["https://httpbin.org/html"],
hooks={
"on_page_context_created": on_page_context_created,
"before_retrieve_html": before_retrieve_html
},
hooks_timeout=30
)
if results and results.success:
print(f"✅ Hooks executed! HTML length: {len(results.html)}")
```
**Benefits of Function-Based Hooks:**
- ✅ Full IDE support (autocomplete, syntax highlighting)
- ✅ Type checking and linting
- ✅ Easier to test and debug
- ✅ Reusable across projects
- ✅ Automatic conversion in Docker client
- ✅ No breaking changes - string hooks still work!
## 🤖 Enhanced LLM Integration
Enhanced LLM integration with custom providers, temperature control, and base URL configuration.

82
docs/md_v2/blog/releases/v0.7.5.md
View File

@@ -8,7 +8,8 @@ Today I'm releasing Crawl4AI v0.7.5—focused on extensibility and security. Thi
## 🎯 What's New at a Glance
- **Docker Hooks System**: Custom Python functions at key pipeline points
- **Docker Hooks System**: Custom Python functions at key pipeline points with function-based API
- **Function-Based Hooks**: New `hooks_to_string()` utility with Docker client auto-conversion
- **Enhanced LLM Integration**: Custom providers with temperature control
- **HTTPS Preservation**: Secure internal link handling
- **Bug Fixes**: Resolved multiple community-reported issues
@@ -82,6 +83,85 @@ if result.get('success'):
- `before_retrieve_html`: Pre-extraction processing
- `before_return_html`: Final HTML processing
### Function-Based Hooks API
Writing hooks as strings works, but lacks IDE support and type checking. v0.7.5 introduces a function-based approach with automatic conversion!
**Option 1: Using the `hooks_to_string()` Utility**
```python
from crawl4ai import hooks_to_string
import requests
# Define hooks as regular Python functions (with full IDE support!)
async def on_page_context_created(page, context, **kwargs):
"""Block images to speed up crawling"""
await context.route("**/*.{png,jpg,jpeg,gif,webp}", lambda route: route.abort())
await page.set_viewport_size({"width": 1920, "height": 1080})
return page
async def before_goto(page, context, url, **kwargs):
"""Add custom headers"""
await page.set_extra_http_headers({
'X-Crawl4AI': 'v0.7.5',
'X-Custom-Header': 'my-value'
})
return page
# Convert functions to strings
hooks_code = hooks_to_string({
"on_page_context_created": on_page_context_created,
"before_goto": before_goto
})
# Use with REST API
payload = {
"urls": ["https://httpbin.org/html"],
"hooks": {"code": hooks_code, "timeout": 30}
}
response = requests.post("http://localhost:11235/crawl", json=payload)
```
**Option 2: Docker Client with Automatic Conversion (Recommended!)**
```python
from crawl4ai.docker_client import Crawl4aiDockerClient
# Define hooks as functions (same as above)
async def on_page_context_created(page, context, **kwargs):
await context.route("**/*.{png,jpg,jpeg,gif,webp}", lambda route: route.abort())
return page
async def before_retrieve_html(page, context, **kwargs):
# Scroll to load lazy content
await page.evaluate("window.scrollTo(0, document.body.scrollHeight)")
await page.wait_for_timeout(1000)
return page
# Use Docker client - conversion happens automatically!
client = Crawl4aiDockerClient(base_url="http://localhost:11235")
results = await client.crawl(
urls=["https://httpbin.org/html"],
hooks={
"on_page_context_created": on_page_context_created,
"before_retrieve_html": before_retrieve_html
},
hooks_timeout=30
)
if results and results.success:
print(f"✅ Hooks executed! HTML length: {len(results.html)}")
```
**Benefits of Function-Based Hooks:**
- ✅ Full IDE support (autocomplete, syntax highlighting)
- ✅ Type checking and linting
- ✅ Easier to test and debug
- ✅ Reusable across projects
- ✅ Automatic conversion in Docker client
- ✅ No breaking changes - string hooks still work!
## 🤖 Enhanced LLM Integration
Enhanced LLM integration with custom providers, temperature control, and base URL configuration.

185
docs/releases_review/demo_v0.7.5.py
View File

@@ -4,7 +4,7 @@
This demo showcases key features introduced in v0.7.5 with real, executable examples.
Featured Demos:
1. ✅ Docker Hooks System - Real API calls with custom hooks
1. ✅ Docker Hooks System - Real API calls with custom hooks (string & function-based)
2. ✅ Enhanced LLM Integration - Working LLM configurations
3. ✅ HTTPS Preservation - Live crawling with HTTPS maintenance
@@ -19,8 +19,10 @@ import requests
import time
import sys
from crawl4ai import (AsyncWebCrawler, CrawlerRunConfig, BrowserConfig,
CacheMode, FilterChain, URLPatternFilter, BFSDeepCrawlStrategy)
from crawl4ai import (AsyncWebCrawler, CrawlerRunConfig, BrowserConfig,
CacheMode, FilterChain, URLPatternFilter, BFSDeepCrawlStrategy,
hooks_to_string)
from crawl4ai.docker_client import Crawl4aiDockerClient
def print_section(title: str, description: str = ""):
@@ -36,13 +38,13 @@ async def demo_1_docker_hooks_system():
"""Demo 1: Docker Hooks System - Real API calls with custom hooks"""
print_section(
"Demo 1: Docker Hooks System",
"Testing real Docker hooks with live API calls"
"Testing both string-based and function-based hooks (NEW in v0.7.5!)"
)
# Check Docker service availability
def check_docker_service():
try:
response = requests.get("http://localhost:11234/", timeout=3)
response = requests.get("http://localhost:11235/", timeout=3)
return response.status_code == 200
except:
return False
@@ -60,108 +62,132 @@ async def demo_1_docker_hooks_system():
print("✓ Docker service detected!")
# Define real working hooks
hooks_config = {
# ============================================================================
# PART 1: Traditional String-Based Hooks (Works with REST API)
# ============================================================================
print("\n" + "─" * 60)
print("Part 1: String-Based Hooks (REST API)")
print("─" * 60)
hooks_config_string = {
"on_page_context_created": """
async def hook(page, context, **kwargs):
print("Hook: Setting up page context")
# Block images to speed up crawling
print("[String Hook] Setting up page context")
await context.route("**/*.{png,jpg,jpeg,gif,webp}", lambda route: route.abort())
print("Hook: Images blocked")
return page
""",
"before_retrieve_html": """
async def hook(page, context, **kwargs):
print("Hook: Before retrieving HTML")
# Scroll to bottom to load lazy content
print("[String Hook] Before retrieving HTML")
await page.evaluate("window.scrollTo(0, document.body.scrollHeight)")
await page.wait_for_timeout(1000)
print("Hook: Scrolled to bottom")
return page
""",
"before_goto": """
async def hook(page, context, url, **kwargs):
print(f"Hook: About to navigate to {url}")
# Add custom headers
await page.set_extra_http_headers({
'X-Test-Header': 'crawl4ai-hooks-test'
})
return page
"""
}
# Test with a reliable URL
test_url = "https://httpbin.org/html"
payload = {
"urls": ["https://httpbin.org/html"],
"hooks": {
"code": hooks_config,
"code": hooks_config_string,
"timeout": 30
}
}
print(f"🎯 Testing URL: {test_url}")
print("🔧 Configured 3 hooks: on_page_context_created, before_retrieve_html, before_goto\n")
# Make the request
print("🔄 Executing hooks...")
print("🔧 Using string-based hooks for REST API...")
try:
start_time = time.time()
response = requests.post(
"http://localhost:11234/crawl",
json=payload,
timeout=60
)
response = requests.post("http://localhost:11235/crawl", json=payload, timeout=60)
execution_time = time.time() - start_time
if response.status_code == 200:
result = response.json()
print(f"🎉 Success! Execution time: {execution_time:.2f}s\n")
# Display results
success = result.get('success', False)
print(f"✅ Crawl Status: {'Success' if success else 'Failed'}")
if success:
markdown_content = result.get('markdown', '')
print(f"📄 Content Length: {len(markdown_content)} characters")
# Show content preview
if markdown_content:
preview = markdown_content[:300] + "..." if len(markdown_content) > 300 else markdown_content
print("\n--- Content Preview ---")
print(preview)
print("--- End Preview ---\n")
# Check if our hook marker is present
raw_html = result.get('html', '')
if "Crawl4AI v0.7.5 Docker Hook" in raw_html:
print("✓ Hook marker found in HTML - hooks executed successfully!")
# Display hook execution info if available
print("\nHook Execution Summary:")
print("🔗 before_goto: URL modified with tracking parameter")
print("✅ after_goto: Page navigation completed")
print("📝 before_return_html: Content processed and marked")
print(f"✅ String-based hooks executed in {execution_time:.2f}s")
if result.get('results') and result['results'][0].get('success'):
html_length = len(result['results'][0].get('html', ''))
print(f" 📄 HTML length: {html_length} characters")
else:
print(f"❌ Request failed: {response.status_code}")
try:
error_data = response.json()
print(f"Error: {error_data}")
except:
print(f"Raw response: {response.text[:500]}")
except requests.exceptions.Timeout:
print("⏰ Request timed out after 60 seconds")
except Exception as e:
print(f"❌ Error: {str(e)}")
# ============================================================================
# PART 2: NEW Function-Based Hooks with Docker Client (v0.7.5)
# ============================================================================
print("\n" + "─" * 60)
print("Part 2: Function-Based Hooks with Docker Client (✨ NEW!)")
print("─" * 60)
# Define hooks as regular Python functions
async def on_page_context_created_func(page, context, **kwargs):
"""Block images to speed up crawling"""
print("[Function Hook] Setting up page context")
await context.route("**/*.{png,jpg,jpeg,gif,webp}", lambda route: route.abort())
await page.set_viewport_size({"width": 1920, "height": 1080})
return page
async def before_goto_func(page, context, url, **kwargs):
"""Add custom headers before navigation"""
print(f"[Function Hook] About to navigate to {url}")
await page.set_extra_http_headers({
'X-Crawl4AI': 'v0.7.5-function-hooks',
'X-Test-Header': 'demo'
})
return page
async def before_retrieve_html_func(page, context, **kwargs):
"""Scroll to load lazy content"""
print("[Function Hook] Scrolling page for lazy-loaded content")
await page.evaluate("window.scrollTo(0, document.body.scrollHeight)")
await page.wait_for_timeout(500)
await page.evaluate("window.scrollTo(0, 0)")
return page
# Use the hooks_to_string utility (can be used standalone)
print("\n📦 Converting functions to strings with hooks_to_string()...")
hooks_as_strings = hooks_to_string({
"on_page_context_created": on_page_context_created_func,
"before_goto": before_goto_func,
"before_retrieve_html": before_retrieve_html_func
})
print(f" ✓ Converted {len(hooks_as_strings)} hooks to string format")
# OR use Docker Client which does conversion automatically!
print("\n🐳 Using Docker Client with automatic conversion...")
try:
client = Crawl4aiDockerClient(base_url="http://localhost:11235")
# Pass function objects directly - conversion happens automatically!
results = await client.crawl(
urls=["https://httpbin.org/html"],
hooks={
"on_page_context_created": on_page_context_created_func,
"before_goto": before_goto_func,
"before_retrieve_html": before_retrieve_html_func
},
hooks_timeout=30
)
if results and results.success:
print(f"✅ Function-based hooks executed successfully!")
print(f" 📄 HTML length: {len(results.html)} characters")
print(f" 🎯 URL: {results.url}")
else:
print("⚠️ Crawl completed but may have warnings")
except Exception as e:
print(f"❌ Docker client error: {str(e)}")
# Show the benefits
print("\n" + "=" * 60)
print("✨ Benefits of Function-Based Hooks:")
print("=" * 60)
print("✓ Full IDE support (autocomplete, syntax highlighting)")
print("✓ Type checking and linting")
print("✓ Easier to test and debug")
print("✓ Reusable across projects")
print("✓ Automatic conversion in Docker client")
print("=" * 60)
async def demo_2_enhanced_llm_integration():
"""Demo 2: Enhanced LLM Integration - Working LLM configurations"""
@@ -182,7 +208,7 @@ async def demo_2_enhanced_llm_integration():
}
try:
response = requests.post(
"http://localhost:11234/md",
"http://localhost:11235/md",
json=payload,
timeout=60
)
@@ -285,7 +311,10 @@ async def main():
print("You've experienced the power of Crawl4AI v0.7.5!")
print("")
print("Key Features Demonstrated:")
print("🔧 Docker Hooks - Custom pipeline modifications")
print("🔧 Docker Hooks - String-based & function-based (NEW!)")
print(" • hooks_to_string() utility for function conversion")
print(" • Docker client with automatic conversion")
print(" • Full IDE support and type checking")
print("🤖 Enhanced LLM - Better AI integration")
print("🔒 HTTPS Preservation - Secure link handling")
print("")