Files
crawl4ai/.github/workflows/docs/ARCHITECTURE.md
Nasrin 7cac008c10 Release/v0.7.6 (#1556)
* fix(docker-api): migrate to modern datetime library API

Signed-off-by: Emmanuel Ferdman <emmanuelferdman@gmail.com>

* Fix examples in README.md

* feat(docker): add user-provided hooks support to Docker API

Implements comprehensive hooks functionality allowing users to provide custom Python
functions as strings that execute at specific points in the crawling pipeline.

Key Features:
- Support for all 8 crawl4ai hook points:
  • on_browser_created: Initialize browser settings
  • on_page_context_created: Configure page context
  • before_goto: Pre-navigation setup
  • after_goto: Post-navigation processing
  • on_user_agent_updated: User agent modification handling
  • on_execution_started: Crawl execution initialization
  • before_retrieve_html: Pre-extraction processing
  • before_return_html: Final HTML processing

Implementation Details:
- Created UserHookManager for validation, compilation, and safe execution
- Added IsolatedHookWrapper for error isolation and timeout protection
- AST-based validation ensures code structure correctness
- Sandboxed execution with restricted builtins for security
- Configurable timeout (1-120 seconds) prevents infinite loops
- Comprehensive error handling ensures hooks don't crash main process
- Execution tracking with detailed statistics and logging

API Changes:
- Added HookConfig schema with code and timeout fields
- Extended CrawlRequest with optional hooks parameter
- Added /hooks/info endpoint for hook discovery
- Updated /crawl and /crawl/stream endpoints to support hooks

Safety Features:
- Malformed hooks return clear validation errors
- Hook errors are isolated and reported without stopping crawl
- Execution statistics track success/failure/timeout rates
- All hook results are JSON-serializable

Testing:
- Comprehensive test suite covering all 8 hooks
- Error handling and timeout scenarios validated
- Authentication, performance, and content extraction examples
- 100% success rate in production testing

Documentation:
- Added extensive hooks section to docker-deployment.md
- Security warnings about user-provided code risks
- Real-world examples using httpbin.org, GitHub, BBC
- Best practices and troubleshooting guide

ref #1377

* fix(deep-crawl): BestFirst priority inversion; remove pre-scoring truncation. ref #1253

  Use negative scores in PQ to visit high-score URLs first and drop link cap prior to scoring; add test for ordering.

* docs: Update URL seeding examples to use proper async context managers
- Wrap all AsyncUrlSeeder usage with async context managers
- Update URL seeding adventure example to use "sitemap+cc" source, focus on course posts, and add stream=True parameter to fix runtime error

* fix(crawler): Removed the incorrect reference in browser_config variable #1310

* docs: update Docker instructions to use the latest release tag

* fix(docker): Fix LLM API key handling for multi-provider support

Previously, the system incorrectly used OPENAI_API_KEY for all LLM providers
due to a hardcoded api_key_env fallback in config.yml. This caused authentication
errors when using non-OpenAI providers like Gemini.

Changes:
- Remove api_key_env from config.yml to let litellm handle provider-specific env vars
- Simplify get_llm_api_key() to return None, allowing litellm to auto-detect keys
- Update validate_llm_provider() to trust litellm's built-in key detection
- Update documentation to reflect the new automatic key handling

The fix leverages litellm's existing capability to automatically find the correct
environment variable for each provider (OPENAI_API_KEY, GEMINI_API_TOKEN, etc.)
without manual configuration.

ref #1291

* docs: update adaptive crawler docs and cache defaults; remove deprecated examples (#1330)
- Replace BaseStrategy with CrawlStrategy in custom strategy examples (DomainSpecificStrategy, HybridStrategy)
- Remove “Custom Link Scoring” and “Caching Strategy” sections no longer aligned with current library
- Revise memory pruning example to use adaptive.get_relevant_content and index-based retention of top 500 docs
- Correct Quickstart note: default cache mode is CacheMode.BYPASS; instruct enabling with CacheMode.ENABLED

* fix(utils): Improve URL normalization by avoiding quote/unquote to preserve '+' signs. ref #1332

* feat: Add comprehensive website to API example with frontend

This commit adds a complete, web scraping API example that demonstrates how to get structured data from any website and use it like an API using the crawl4ai library with a minimalist frontend interface.

Core Functionality
- AI-powered web scraping with plain English queries
- Dual scraping approaches: Schema-based (faster) and LLM-based (flexible)
- Intelligent schema caching for improved performance
- Custom LLM model support with API key management
- Automatic duplicate request prevention

Modern Frontend Interface
- Minimalist black-and-white design inspired by modern web apps
- Responsive layout with smooth animations and transitions
- Three main pages: Scrape Data, Models Management, API Request History
- Real-time results display with JSON formatting
- Copy-to-clipboard functionality for extracted data
- Toast notifications for user feedback
- Auto-scroll to results when scraping starts

Model Management System
- Web-based model configuration interface
- Support for any LLM provider (OpenAI, Gemini, Anthropic, etc.)
- Simplified configuration requiring only provider and API token
- Add, list, and delete model configurations
- Secure storage of API keys in local JSON files

API Request History
- Automatic saving of all API requests and responses
- Display of request history with URL, query, and cURL commands
- Duplicate prevention (same URL + query combinations)
- Request deletion functionality
- Clean, simplified display focusing on essential information

Technical Implementation

Backend (FastAPI)
- RESTful API with comprehensive endpoints
- Pydantic models for request/response validation
- Async web scraping with crawl4ai library
- Error handling with detailed error messages
- File-based storage for models and request history

Frontend (Vanilla JS/CSS/HTML)
- No framework dependencies - pure HTML, CSS, JavaScript
- Modern CSS Grid and Flexbox layouts
- Custom dropdown styling with SVG arrows
- Responsive design for mobile and desktop
- Smooth scrolling and animations

Core Library Integration
- WebScraperAgent class for orchestration
- ModelConfig class for LLM configuration management
- Schema generation and caching system
- LLM extraction strategy support
- Browser configuration with headless mode

* fix(dependencies): add cssselect to project dependencies

Fixes bug reported in issue #1405
[Bug]: Excluded selector (excluded_selector) doesn't work

This commit reintroduces the cssselect library which was removed by PR (https://github.com/unclecode/crawl4ai/pull/1368) and merged via (437395e490).

Integration tested against 0.7.4 Docker container. Reintroducing cssselector package eliminated errors seen in logs and excluded_selector functionality was restored.

Refs: #1405

* fix(docker): resolve filter serialization and JSON encoding errors in deep crawl strategy (ref #1419)

  - Fix URLPatternFilter serialization by preventing private __slots__ from being serialized as constructor params
  - Add public attributes to URLPatternFilter to store original constructor parameters for proper serialization
  - Handle property descriptors in CrawlResult.model_dump() to prevent JSON serialization errors
  - Ensure filter chains work correctly with Docker client and REST API

  The issue occurred because:
  1. Private implementation details (_simple_suffixes, etc.) were being serialized and passed as constructor arguments during deserialization
  2. Property descriptors were being included in the serialized output, causing "Object of type property is not JSON serializable" errors

  Changes:
  - async_configs.py: Comment out __slots__ serialization logic (lines 100-109)
  - filters.py: Add patterns, use_glob, reverse to URLPatternFilter __slots__ and store as public attributes
  - models.py: Convert property descriptors to strings in model_dump() instead of including them directly

* fix(logger): ensure logger is a Logger instance in crawling strategies. ref #1437

* feat(docker): Add temperature and base_url parameters for LLM configuration. ref #1035

  Implement hierarchical configuration for LLM parameters with support for:
  - Temperature control (0.0-2.0) to adjust response creativity
  - Custom base_url for proxy servers and alternative endpoints
  - 4-tier priority: request params > provider env > global env > defaults

  Add helper functions in utils.py, update API schemas and handlers,
  support environment variables (LLM_TEMPERATURE, OPENAI_TEMPERATURE, etc.),
  and provide comprehensive documentation with examples.

* feat(docker): improve docker error handling
- Return comprehensive error messages along with status codes for api internal errors.
- Fix fit_html property serialization issue in both /crawl and /crawl/stream endpoints
- Add sanitization to ensure fit_html is always JSON-serializable (string or None)
- Add comprehensive error handling test suite.

* #1375 : refactor(proxy) Deprecate 'proxy' parameter in BrowserConfig and enhance proxy string parsing

- Updated ProxyConfig.from_string to support multiple proxy formats, including URLs with credentials.
- Deprecated the 'proxy' parameter in BrowserConfig, replacing it with 'proxy_config' for better flexibility.
- Added warnings for deprecated usage and clarified behavior when both parameters are provided.
- Updated documentation and tests to reflect changes in proxy configuration handling.

* Remove deprecated test for 'proxy' parameter in BrowserConfig and update .gitignore to include test_scripts directory.

* feat: add preserve_https_for_internal_links flag to maintain HTTPS during crawling. Ref #1410

Added a new `preserve_https_for_internal_links` configuration flag that preserves the original HTTPS scheme for same-domain links even when the server redirects to HTTP.

* feat: update documentation for preserve_https_for_internal_links. ref #1410

* fix: drop Python 3.9 support and require Python >=3.10.
The library no longer supports Python 3.9 and so it was important to drop all references to python 3.9.
Following changes have been made:
- pyproject.toml: set requires-python to ">=3.10"; remove 3.9 classifier
- setup.py: set python_requires to ">=3.10"; remove 3.9 classifier
- docs: update Python version mentions
  - deploy/docker/c4ai-doc-context.md: options -> 3.10, 3.11, 3.12, 3.13

* issue #1329 refactor(crawler): move unwanted properties to CrawlerRunConfig class

* fix(auth): fixed Docker JWT authentication. ref #1442

* remove: delete unused yoyo snapshot subproject

* fix: raise error on last attempt failure in perform_completion_with_backoff. ref #989

* Commit without API

* fix: update option labels in request builder for clarity

* fix: allow custom LLM providers for adaptive crawler embedding config. ref: #1291

  - Change embedding_llm_config from Dict to Union[LLMConfig, Dict] for type safety
  - Add backward-compatible conversion property _embedding_llm_config_dict
  - Replace all hardcoded OpenAI embedding configs with configurable options
  - Fix LLMConfig object attribute access in query expansion logic
  - Add comprehensive example demonstrating multiple provider configurations
  - Update documentation with both LLMConfig object and dictionary usage patterns

  Users can now specify any LLM provider for query expansion in embedding strategy:
  - New: embedding_llm_config=LLMConfig(provider='anthropic/claude-3', api_token='key')
  - Old: embedding_llm_config={'provider': 'openai/gpt-4', 'api_token': 'key'} (still works)

* refactor(BrowserConfig): change deprecation warning for 'proxy' parameter to UserWarning

* feat(StealthAdapter): fix stealth features for Playwright integration. ref #1481

* #1505 fix(api): update config handling to only set base config if not provided by user

* fix(docker-deployment): replace console.log with print for metadata extraction

* Release v0.7.5: The Update

- Updated version to 0.7.5
- Added comprehensive demo and release notes
- Updated documentation

* refactor(release): remove memory management section for cleaner documentation. ref #1443

* feat(docs): add brand book and page copy functionality

- Add comprehensive brand book with color system, typography, components
- Add page copy dropdown with markdown copy/view functionality
- Update mkdocs.yml with new assets and branding navigation
- Use terminal-style ASCII icons and condensed menu design

* Update gitignore add local scripts folder

* fix: remove this import as it causes python to treat "json" as a variable in the except block

* fix: always return a list, even if we catch an exception

* feat(marketplace): Add Crawl4AI marketplace with secure configuration

- Implement marketplace frontend and admin dashboard
- Add FastAPI backend with environment-based configuration
- Use .env file for secrets management
- Include data generation scripts
- Add proper CORS configuration
- Remove hardcoded password from admin login
- Update gitignore for security

* fix(marketplace): Update URLs to use /marketplace path and relative API endpoints

- Change API_BASE to relative '/api' for production
- Move marketplace to /marketplace instead of /marketplace/frontend
- Update MkDocs navigation
- Fix logo path in marketplace index

* fix(docs): hide copy menu on non-markdown pages

* feat(marketplace): add sponsor logo uploads

Co-authored-by: factory-droid[bot] <138933559+factory-droid[bot]@users.noreply.github.com>

* feat(docs): add chatgpt quick link to page actions

* fix(marketplace): align admin api with backend endpoints

* fix(marketplace): isolate api under marketplace prefix

* fix(marketplace): resolve app detail page routing and styling issues

- Fixed JavaScript errors from missing HTML elements (install-code, usage-code, integration-code)
- Added missing CSS classes for tabs, overview layout, sidebar, and integration content
- Fixed tab navigation to display horizontally in single line
- Added proper padding to tab content sections (removed from container, added to content)
- Fixed tab selector from .nav-tab to .tab-btn to match HTML structure
- Added sidebar styling with stats grid and metadata display
- Improved responsive design with mobile-friendly tab scrolling
- Fixed code block positioning for copy buttons
- Removed margin from first headings to prevent extra spacing
- Added null checks for DOM elements in JavaScript to prevent errors

These changes resolve the routing issue where clicking on apps caused page redirects,
and fix the broken layout where CSS was not properly applied to the app detail page.

* fix(marketplace): prevent hero image overflow and secondary card stretching

- Fixed hero image to 200px height with min/max constraints
- Added object-fit: cover to hero-image img elements
- Changed secondary-featured align-items from stretch to flex-start
- Fixed secondary-card height to 118px (no flex: 1 stretching)
- Updated responsive grid layouts for wider screens
- Added flex: 1 to hero-content for better content distribution

These changes ensure a rigid, predictable layout that prevents:
1. Large images from pushing text content down
2. Single secondary cards from stretching to fill entire height

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

   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)

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

   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)

* fix(docs): clarify Docker Hooks System with function-based API in README

* docs: Add demonstration files for v0.7.5 release, showcasing the new Docker Hooks System and all other features.

* docs: Update 0.7.5 video walkthrough

* docs: add complete SDK reference documentation

Add comprehensive single-page SDK reference combining:
- Installation & Setup
- Quick Start
- Core API (AsyncWebCrawler, arun, arun_many, CrawlResult)
- Configuration (BrowserConfig, CrawlerConfig, Parameters)
- Crawling Patterns
- Content Processing (Markdown, Fit Markdown, Selection, Interaction, Link & Media)
- Extraction Strategies (LLM and No-LLM)
- Advanced Features (Session Management, Hooks & Auth)

Generated using scripts/generate_sdk_docs.py in ultra-dense mode
optimized for AI assistant consumption.

Stats: 23K words, 185 code blocks, 220KB

* feat: add AI assistant skill package for Crawl4AI

- Create comprehensive skill package for AI coding assistants
- Include complete SDK reference (23K words, v0.7.4)
- Add three extraction scripts (basic, batch, pipeline)
- Implement version tracking in skill and scripts
- Add prominent download section on homepage
- Place skill in docs/assets for web distribution

The skill enables AI assistants like Claude, Cursor, and Windsurf
to effectively use Crawl4AI with optimized workflows for markdown
generation and data extraction.

* fix: remove non-existent wiki link and clarify skill usage instructions

* fix: update Crawl4AI skill with corrected parameters and examples

- Fixed CrawlerConfig → CrawlerRunConfig throughout
- Fixed parameter names (timeout → page_timeout, store_html removed)
- Fixed schema format (selector → baseSelector)
- Corrected proxy configuration (in BrowserConfig, not CrawlerRunConfig)
- Fixed fit_markdown usage with content filters
- Added comprehensive references to docs/examples/ directory
- Created safe packaging script to avoid root directory pollution
- All scripts tested and verified working

* fix: thoroughly verify and fix all Crawl4AI skill examples

- Cross-checked every section against actual docs
- Fixed BM25ContentFilter parameters (user_query, bm25_threshold)
- Removed incorrect wait_for selector from basic example
- Added comprehensive test suite (4 test files)
- All examples now tested and verified working
- Tests validate: basic crawling, markdown generation, data extraction, advanced patterns
- Package size: 76.6 KB (includes tests for future validation)

* feat(ci): split release pipeline and add Docker caching

- Split release.yml into PyPI/GitHub release and Docker workflows
- Add GitHub Actions cache for Docker builds (10-15x faster rebuilds)
- Implement dual-trigger for docker-release.yml (auto + manual)
- Add comprehensive workflow documentation in .github/workflows/docs/
- Backup original workflow as release.yml.backup

* feat: add webhook notifications for crawl job completion

Implements webhook support for the crawl job API to eliminate polling requirements.

Changes:
- Added WebhookConfig and WebhookPayload schemas to schemas.py
- Created webhook.py with WebhookDeliveryService class
- Integrated webhook notifications in api.py handle_crawl_job
- Updated job.py CrawlJobPayload to accept webhook_config
- Added webhook configuration section to config.yml
- Included comprehensive usage examples in WEBHOOK_EXAMPLES.md

Features:
- Webhook notifications on job completion (success/failure)
- Configurable data inclusion in webhook payload
- Custom webhook headers support
- Global default webhook URL configuration
- Exponential backoff retry logic (5 attempts: 1s, 2s, 4s, 8s, 16s)
- 30-second timeout per webhook call

Usage:
POST /crawl/job with optional webhook_config:
- webhook_url: URL to receive notifications
- webhook_data_in_payload: include full results (default: false)
- webhook_headers: custom headers for authentication

Generated with Claude Code https://claude.com/claude-code

Co-Authored-By: Claude <noreply@anthropic.com>

* docs: add webhook documentation to Docker README

Added comprehensive webhook section to README.md including:
- Overview of asynchronous job queue with webhooks
- Benefits and use cases
- Quick start examples
- Webhook authentication
- Global webhook configuration
- Job status polling alternative

Updated table of contents and summary to include webhook feature.
Maintains consistent tone and style with rest of README.

Generated with Claude Code https://claude.com/claude-code

Co-Authored-By: Claude <noreply@anthropic.com>

* docs: add webhook example for Docker deployment

Added docker_webhook_example.py demonstrating:
- Submitting crawl jobs with webhook configuration
- Flask-based webhook receiver implementation
- Three usage patterns:
  1. Webhook notification only (fetch data separately)
  2. Webhook with full data in payload
  3. Traditional polling approach for comparison

Includes comprehensive comments explaining:
- Webhook payload structure
- Authentication headers setup
- Error handling
- Production deployment tips

Example is fully functional and ready to run with Flask installed.

Generated with Claude Code https://claude.com/claude-code

Co-Authored-By: Claude <noreply@anthropic.com>

* test: add webhook implementation validation tests

Added comprehensive test suite to validate webhook implementation:
- Module import verification
- WebhookDeliveryService initialization
- Pydantic model validation (WebhookConfig)
- Payload construction logic
- Exponential backoff calculation
- API integration checks

All tests pass (6/6), confirming implementation is correct.

Generated with Claude Code https://claude.com/claude-code

Co-Authored-By: Claude <noreply@anthropic.com>

* test: add comprehensive webhook feature test script

Added end-to-end test script that automates webhook feature testing:

Script Features (test_webhook_feature.sh):
- Automatic branch switching and dependency installation
- Redis and server startup/shutdown management
- Webhook receiver implementation
- Integration test for webhook notifications
- Comprehensive cleanup and error handling
- Returns to original branch after completion

Test Flow:
1. Fetch and checkout webhook feature branch
2. Activate venv and install dependencies
3. Start Redis and Crawl4AI server
4. Submit crawl job with webhook config
5. Verify webhook delivery and payload
6. Clean up all processes and return to original branch

Documentation:
- WEBHOOK_TEST_README.md with usage instructions
- Troubleshooting guide
- Exit codes and safety features

Usage: ./tests/test_webhook_feature.sh

Generated with Claude Code https://claude.com/claude-code

Co-Authored-By: Claude <noreply@anthropic.com>

* fix: properly serialize Pydantic HttpUrl in webhook config

Use model_dump(mode='json') instead of deprecated dict() method to ensure
Pydantic special types (HttpUrl, UUID, etc.) are properly serialized to
JSON-compatible native Python types.

This fixes webhook delivery failures caused by HttpUrl objects remaining
as Pydantic types in the webhook_config dict, which caused JSON
serialization errors and httpx request failures.

Also update mcp requirement to >=1.18.0 for compatibility.

* feat: add webhook support for /llm/job endpoint

Add comprehensive webhook notification support for the /llm/job endpoint,
following the same pattern as the existing /crawl/job implementation.

Changes:
- Add webhook_config field to LlmJobPayload model (job.py)
- Implement webhook notifications in process_llm_extraction() with 4
  notification points: success, provider validation failure, extraction
  failure, and general exceptions (api.py)
- Store webhook_config in Redis task data for job tracking
- Initialize WebhookDeliveryService with exponential backoff retry logic
Documentation:
- Add Example 6 to WEBHOOK_EXAMPLES.md showing LLM extraction with webhooks
- Update Flask webhook handler to support both crawl and llm_extraction tasks
- Add TypeScript client examples for LLM jobs
- Add comprehensive examples to docker_webhook_example.py with schema support
- Clarify data structure differences between webhook and API responses

Testing:
- Add test_llm_webhook_feature.py with 7 validation tests (all passing)
- Verify pattern consistency with /crawl/job implementation
- Add implementation guide (WEBHOOK_LLM_JOB_IMPLEMENTATION.md)

* fix: remove duplicate comma in webhook_config parameter

* fix: update Crawl4AI Docker container port from 11234 to 11235

* Release v0.7.6: The 0.7.6 Update

- Updated version to 0.7.6
- Added comprehensive demo and release notes
- Updated all documentation
- Update the veriosn in Dockerfile to 0.7.6

---------

Signed-off-by: Emmanuel Ferdman <emmanuelferdman@gmail.com>
Co-authored-by: Emmanuel Ferdman <emmanuelferdman@gmail.com>
Co-authored-by: Nezar Ali <abu5sohaib@gmail.com>
Co-authored-by: Soham Kukreti <kukretisoham@gmail.com>
Co-authored-by: James T. Wood <jamesthomaswood@gmail.com>
Co-authored-by: AHMET YILMAZ <tawfik@kidocode.com>
Co-authored-by: nafeqq-1306 <nafiquee@yahoo.com>
Co-authored-by: unclecode <unclecode@kidocode.com>
Co-authored-by: Martin Sjöborg <martin.sjoborg@quartr.se>
Co-authored-by: Martin Sjöborg <martin@sjoborg.org>
Co-authored-by: factory-droid[bot] <138933559+factory-droid[bot]@users.noreply.github.com>
Co-authored-by: Claude <noreply@anthropic.com>
2025-10-22 20:41:06 +08:00

918 lines
30 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
# Workflow Architecture Documentation
## Overview
This document describes the technical architecture of the split release pipeline for Crawl4AI.
---
## Architecture Diagram
```
┌─────────────────────────────────────────────────────────────────┐
│ Developer │
│ │ │
│ ▼ │
│ git tag v1.2.3 │
│ git push --tags │
└──────────────────────────────┬──────────────────────────────────┘
┌─────────────────────────────────────────────────────────────────┐
│ GitHub Repository │
│ │
│ ┌────────────────────────────────────────────────────────┐ │
│ │ Tag Event: v1.2.3 │ │
│ └────────────────────────────────────────────────────────┘ │
│ │ │
│ ▼ │
│ ┌────────────────────────────────────────────────────────┐ │
│ │ release.yml (Release Pipeline) │ │
│ │ ┌──────────────────────────────────────────────┐ │ │
│ │ │ 1. Extract Version │ │ │
│ │ │ v1.2.3 → 1.2.3 │ │ │
│ │ └──────────────────────────────────────────────┘ │ │
│ │ ┌──────────────────────────────────────────────┐ │ │
│ │ │ 2. Validate Version │ │ │
│ │ │ Tag == __version__.py │ │ │
│ │ └──────────────────────────────────────────────┘ │ │
│ │ ┌──────────────────────────────────────────────┐ │ │
│ │ │ 3. Build Python Package │ │ │
│ │ │ - Source dist (.tar.gz) │ │ │
│ │ │ - Wheel (.whl) │ │ │
│ │ └──────────────────────────────────────────────┘ │ │
│ │ ┌──────────────────────────────────────────────┐ │ │
│ │ │ 4. Upload to PyPI │ │ │
│ │ │ - Authenticate with token │ │ │
│ │ │ - Upload dist/* │ │ │
│ │ └──────────────────────────────────────────────┘ │ │
│ │ ┌──────────────────────────────────────────────┐ │ │
│ │ │ 5. Create GitHub Release │ │ │
│ │ │ - Tag: v1.2.3 │ │ │
│ │ │ - Body: Install instructions │ │ │
│ │ │ - Status: Published │ │ │
│ │ └──────────────────────────────────────────────┘ │ │
│ └────────────────────────────────────────────────────────┘ │
│ │ │
│ ▼ │
│ ┌────────────────────────────────────────────────────────┐ │
│ │ Release Event: published (v1.2.3) │ │
│ └────────────────────────────────────────────────────────┘ │
│ │ │
│ ▼ │
│ ┌────────────────────────────────────────────────────────┐ │
│ │ docker-release.yml (Docker Pipeline) │ │
│ │ ┌──────────────────────────────────────────────┐ │ │
│ │ │ 1. Extract Version from Release │ │ │
│ │ │ github.event.release.tag_name → 1.2.3 │ │ │
│ │ └──────────────────────────────────────────────┘ │ │
│ │ ┌──────────────────────────────────────────────┐ │ │
│ │ │ 2. Parse Semantic Versions │ │ │
│ │ │ 1.2.3 → Major: 1, Minor: 1.2 │ │ │
│ │ └──────────────────────────────────────────────┘ │ │
│ │ ┌──────────────────────────────────────────────┐ │ │
│ │ │ 3. Setup Multi-Arch Build │ │ │
│ │ │ - Docker Buildx │ │ │
│ │ │ - QEMU emulation │ │ │
│ │ └──────────────────────────────────────────────┘ │ │
│ │ ┌──────────────────────────────────────────────┐ │ │
│ │ │ 4. Authenticate Docker Hub │ │ │
│ │ │ - Username: DOCKER_USERNAME │ │ │
│ │ │ - Token: DOCKER_TOKEN │ │ │
│ │ └──────────────────────────────────────────────┘ │ │
│ │ ┌──────────────────────────────────────────────┐ │ │
│ │ │ 5. Build Multi-Arch Images │ │ │
│ │ │ ┌────────────────┬────────────────┐ │ │ │
│ │ │ │ linux/amd64 │ linux/arm64 │ │ │ │
│ │ │ └────────────────┴────────────────┘ │ │ │
│ │ │ Cache: GitHub Actions (type=gha) │ │ │
│ │ └──────────────────────────────────────────────┘ │ │
│ │ ┌──────────────────────────────────────────────┐ │ │
│ │ │ 6. Push to Docker Hub │ │ │
│ │ │ Tags: │ │ │
│ │ │ - unclecode/crawl4ai:1.2.3 │ │ │
│ │ │ - unclecode/crawl4ai:1.2 │ │ │
│ │ │ - unclecode/crawl4ai:1 │ │ │
│ │ │ - unclecode/crawl4ai:latest │ │ │
│ │ └──────────────────────────────────────────────┘ │ │
│ └────────────────────────────────────────────────────────┘ │
└─────────────────────────────────────────────────────────────────┘
┌─────────────────────────────────────────────────────────────────┐
│ External Services │
│ │
│ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐ │
│ │ PyPI │ │ Docker Hub │ │ GitHub │ │
│ │ │ │ │ │ │ │
│ │ crawl4ai │ │ unclecode/ │ │ Releases │ │
│ │ 1.2.3 │ │ crawl4ai │ │ v1.2.3 │ │
│ └──────────────┘ └──────────────┘ └──────────────┘ │
└─────────────────────────────────────────────────────────────────┘
```
---
## Component Details
### 1. Release Pipeline (release.yml)
#### Purpose
Fast publication of Python package and GitHub release.
#### Input
- **Trigger**: Git tag matching `v*` (excluding `test-v*`)
- **Example**: `v1.2.3`
#### Processing Stages
##### Stage 1: Version Extraction
```bash
Input: refs/tags/v1.2.3
Output: VERSION=1.2.3
```
**Implementation**:
```bash
TAG_VERSION=${GITHUB_REF#refs/tags/v} # Remove 'refs/tags/v' prefix
echo "VERSION=$TAG_VERSION" >> $GITHUB_OUTPUT
```
##### Stage 2: Version Validation
```bash
Input: TAG_VERSION=1.2.3
Check: crawl4ai/__version__.py contains __version__ = "1.2.3"
Output: Pass/Fail
```
**Implementation**:
```bash
PACKAGE_VERSION=$(python -c "from crawl4ai.__version__ import __version__; print(__version__)")
if [ "$TAG_VERSION" != "$PACKAGE_VERSION" ]; then
exit 1
fi
```
##### Stage 3: Package Build
```bash
Input: Source code + pyproject.toml
Output: dist/crawl4ai-1.2.3.tar.gz
dist/crawl4ai-1.2.3-py3-none-any.whl
```
**Implementation**:
```bash
python -m build
# Uses build backend defined in pyproject.toml
```
##### Stage 4: PyPI Upload
```bash
Input: dist/*.{tar.gz,whl}
Auth: PYPI_TOKEN
Output: Package published to PyPI
```
**Implementation**:
```bash
twine upload dist/*
# Environment:
# TWINE_USERNAME: __token__
# TWINE_PASSWORD: ${{ secrets.PYPI_TOKEN }}
```
##### Stage 5: GitHub Release Creation
```bash
Input: Tag: v1.2.3
Body: Markdown content
Output: Published GitHub release
```
**Implementation**:
```yaml
uses: softprops/action-gh-release@v2
with:
tag_name: v1.2.3
name: Release v1.2.3
body: |
Installation instructions and changelog
draft: false
prerelease: false
```
#### Output
- **PyPI Package**: https://pypi.org/project/crawl4ai/1.2.3/
- **GitHub Release**: Published release on repository
- **Event**: `release.published` (triggers Docker workflow)
#### Timeline
```
0:00 - Tag pushed
0:01 - Checkout + Python setup
0:02 - Version validation
0:03 - Package build
0:04 - PyPI upload starts
0:06 - PyPI upload complete
0:07 - GitHub release created
0:08 - Workflow complete
```
---
### 2. Docker Release Pipeline (docker-release.yml)
#### Purpose
Build and publish multi-architecture Docker images.
#### Inputs
##### Input 1: Release Event (Automatic)
```yaml
Event: release.published
Data: github.event.release.tag_name = "v1.2.3"
```
##### Input 2: Docker Rebuild Tag (Manual)
```yaml
Tag: docker-rebuild-v1.2.3
```
#### Processing Stages
##### Stage 1: Version Detection
```bash
# From release event:
VERSION = github.event.release.tag_name.strip("v")
# Result: "1.2.3"
# From rebuild tag:
VERSION = GITHUB_REF.replace("refs/tags/docker-rebuild-v", "")
# Result: "1.2.3"
```
##### Stage 2: Semantic Version Parsing
```bash
Input: VERSION=1.2.3
Output: MAJOR=1
MINOR=1.2
PATCH=3 (implicit)
```
**Implementation**:
```bash
MAJOR=$(echo $VERSION | cut -d. -f1) # Extract first component
MINOR=$(echo $VERSION | cut -d. -f1-2) # Extract first two components
```
##### Stage 3: Multi-Architecture Setup
```yaml
Setup:
- Docker Buildx (multi-platform builder)
- QEMU (ARM emulation on x86)
Platforms:
- linux/amd64 (x86_64)
- linux/arm64 (aarch64)
```
**Architecture**:
```
GitHub Runner (linux/amd64)
├─ Buildx Builder
│ ├─ Native: Build linux/amd64 image
│ └─ QEMU: Emulate ARM to build linux/arm64 image
└─ Generate manifest list (points to both images)
```
##### Stage 4: Docker Hub Authentication
```bash
Input: DOCKER_USERNAME
DOCKER_TOKEN
Output: Authenticated Docker client
```
##### Stage 5: Build with Cache
```yaml
Cache Configuration:
cache-from: type=gha # Read from GitHub Actions cache
cache-to: type=gha,mode=max # Write all layers
Cache Key Components:
- Workflow file path
- Branch name
- Architecture (amd64/arm64)
```
**Cache Hierarchy**:
```
Cache Entry: main/docker-release.yml/linux-amd64
├─ Layer: sha256:abc123... (FROM python:3.12)
├─ Layer: sha256:def456... (RUN apt-get update)
├─ Layer: sha256:ghi789... (COPY requirements.txt)
├─ Layer: sha256:jkl012... (RUN pip install)
└─ Layer: sha256:mno345... (COPY . /app)
Cache Hit/Miss Logic:
- If layer input unchanged → cache hit → skip build
- If layer input changed → cache miss → rebuild + all subsequent layers
```
##### Stage 6: Tag Generation
```bash
Input: VERSION=1.2.3, MAJOR=1, MINOR=1.2
Output Tags:
- unclecode/crawl4ai:1.2.3 (exact version)
- unclecode/crawl4ai:1.2 (minor version)
- unclecode/crawl4ai:1 (major version)
- unclecode/crawl4ai:latest (latest stable)
```
**Tag Strategy**:
- All tags point to same image SHA
- Users can pin to desired stability level
- Pushing new version updates `1`, `1.2`, and `latest` automatically
##### Stage 7: Push to Registry
```bash
For each tag:
For each platform (amd64, arm64):
Push image to Docker Hub
Create manifest list:
Manifest: unclecode/crawl4ai:1.2.3
├─ linux/amd64: sha256:abc...
└─ linux/arm64: sha256:def...
Docker CLI automatically selects correct platform on pull
```
#### Output
- **Docker Images**: 4 tags × 2 platforms = 8 image variants + 4 manifests
- **Docker Hub**: https://hub.docker.com/r/unclecode/crawl4ai/tags
#### Timeline
**Cold Cache (First Build)**:
```
0:00 - Release event received
0:01 - Checkout + Buildx setup
0:02 - Docker Hub auth
0:03 - Start build (amd64)
0:08 - Complete amd64 build
0:09 - Start build (arm64)
0:14 - Complete arm64 build
0:15 - Generate manifests
0:16 - Push all tags
0:17 - Workflow complete
```
**Warm Cache (Code Change Only)**:
```
0:00 - Release event received
0:01 - Checkout + Buildx setup
0:02 - Docker Hub auth
0:03 - Start build (amd64) - cache hit for layers 1-4
0:04 - Complete amd64 build (only layer 5 rebuilt)
0:05 - Start build (arm64) - cache hit for layers 1-4
0:06 - Complete arm64 build (only layer 5 rebuilt)
0:07 - Generate manifests
0:08 - Push all tags
0:09 - Workflow complete
```
---
## Data Flow
### Version Information Flow
```
Developer
crawl4ai/__version__.py
__version__ = "1.2.3"
├─► Git Tag
│ v1.2.3
│ │
│ ▼
│ release.yml
│ │
│ ├─► Validation
│ │ ✓ Match
│ │
│ ├─► PyPI Package
│ │ crawl4ai==1.2.3
│ │
│ └─► GitHub Release
│ v1.2.3
│ │
│ ▼
│ docker-release.yml
│ │
│ └─► Docker Tags
│ 1.2.3, 1.2, 1, latest
└─► Package Metadata
pyproject.toml
version = "1.2.3"
```
### Secrets Flow
```
GitHub Secrets (Encrypted at Rest)
├─► PYPI_TOKEN
│ │
│ ▼
│ release.yml
│ │
│ ▼
│ TWINE_PASSWORD env var (masked in logs)
│ │
│ ▼
│ PyPI API (HTTPS)
├─► DOCKER_USERNAME
│ │
│ ▼
│ docker-release.yml
│ │
│ ▼
│ docker/login-action (masked in logs)
│ │
│ ▼
│ Docker Hub API (HTTPS)
└─► DOCKER_TOKEN
docker-release.yml
docker/login-action (masked in logs)
Docker Hub API (HTTPS)
```
### Artifact Flow
```
Source Code
├─► release.yml
│ │
│ ▼
│ python -m build
│ │
│ ├─► crawl4ai-1.2.3.tar.gz
│ │ │
│ │ ▼
│ │ PyPI Storage
│ │ │
│ │ ▼
│ │ pip install crawl4ai
│ │
│ └─► crawl4ai-1.2.3-py3-none-any.whl
│ │
│ ▼
│ PyPI Storage
│ │
│ ▼
│ pip install crawl4ai
└─► docker-release.yml
docker build
├─► Image: linux/amd64
│ │
│ └─► Docker Hub
│ unclecode/crawl4ai:1.2.3-amd64
└─► Image: linux/arm64
└─► Docker Hub
unclecode/crawl4ai:1.2.3-arm64
```
---
## State Machines
### Release Pipeline State Machine
```
┌─────────┐
│ START │
└────┬────┘
┌──────────────┐
│ Extract │
│ Version │
└──────┬───────┘
┌──────────────┐ ┌─────────┐
│ Validate │─────►│ FAILED │
│ Version │ No │ (Exit 1)│
└──────┬───────┘ └─────────┘
│ Yes
┌──────────────┐
│ Build │
│ Package │
└──────┬───────┘
┌──────────────┐ ┌─────────┐
│ Upload │─────►│ FAILED │
│ to PyPI │ Error│ (Exit 1)│
└──────┬───────┘ └─────────┘
│ Success
┌──────────────┐
│ Create │
│ GH Release │
└──────┬───────┘
┌──────────────┐
│ SUCCESS │
│ (Emit Event) │
└──────────────┘
```
### Docker Pipeline State Machine
```
┌─────────┐
│ START │
│ (Event) │
└────┬────┘
┌──────────────┐
│ Detect │
│ Version │
│ Source │
└──────┬───────┘
┌──────────────┐
│ Parse │
│ Semantic │
│ Versions │
└──────┬───────┘
┌──────────────┐ ┌─────────┐
│ Authenticate │─────►│ FAILED │
│ Docker Hub │ Error│ (Exit 1)│
└──────┬───────┘ └─────────┘
│ Success
┌──────────────┐
│ Build │
│ amd64 │
└──────┬───────┘
┌──────────────┐ ┌─────────┐
│ Build │─────►│ FAILED │
│ arm64 │ Error│ (Exit 1)│
└──────┬───────┘ └─────────┘
│ Success
┌──────────────┐
│ Push All │
│ Tags │
└──────┬───────┘
┌──────────────┐
│ SUCCESS │
└──────────────┘
```
---
## Security Architecture
### Threat Model
#### Threats Mitigated
1. **Secret Exposure**
- Mitigation: GitHub Actions secret masking
- Evidence: Secrets never appear in logs
2. **Unauthorized Package Upload**
- Mitigation: Scoped PyPI tokens
- Evidence: Token limited to `crawl4ai` project
3. **Man-in-the-Middle**
- Mitigation: HTTPS for all API calls
- Evidence: PyPI, Docker Hub, GitHub all use TLS
4. **Supply Chain Tampering**
- Mitigation: Immutable artifacts, content checksums
- Evidence: PyPI stores SHA256, Docker uses content-addressable storage
#### Trust Boundaries
```
┌─────────────────────────────────────────┐
│ Trusted Zone │
│ ┌────────────────────────────────┐ │
│ │ GitHub Actions Runner │ │
│ │ - Ephemeral VM │ │
│ │ - Isolated environment │ │
│ │ - Access to secrets │ │
│ └────────────────────────────────┘ │
│ │ │
│ │ HTTPS (TLS 1.2+) │
│ ▼ │
└─────────────────────────────────────────┘
┌────────────┼────────────┐
│ │ │
▼ ▼ ▼
┌────────┐ ┌─────────┐ ┌──────────┐
│ PyPI │ │ Docker │ │ GitHub │
│ API │ │ Hub │ │ API │
└────────┘ └─────────┘ └──────────┘
External External External
Service Service Service
```
### Secret Management
#### Secret Lifecycle
```
Creation (Developer)
├─► PyPI: Create API token (scoped to project)
├─► Docker Hub: Create access token (read/write)
Storage (GitHub)
├─► Encrypted at rest (AES-256)
├─► Access controlled (repo-scoped)
Usage (Workflow)
├─► Injected as env vars
├─► Masked in logs (GitHub redacts on output)
├─► Never persisted to disk (in-memory only)
Transmission (API Call)
├─► HTTPS only
├─► TLS 1.2+ with strong ciphers
Rotation (Manual)
└─► Regenerate on PyPI/Docker Hub
Update GitHub secret
```
---
## Performance Characteristics
### Release Pipeline Performance
| Metric | Value | Notes |
|--------|-------|-------|
| Cold start | ~2-3 min | First run on new runner |
| Warm start | ~2-3 min | Minimal caching benefit |
| PyPI upload | ~30-60 sec | Network-bound |
| Package build | ~30 sec | CPU-bound |
| Parallelization | None | Sequential by design |
### Docker Pipeline Performance
| Metric | Cold Cache | Warm Cache (code) | Warm Cache (deps) |
|--------|-----------|-------------------|-------------------|
| Total time | 10-15 min | 1-2 min | 3-5 min |
| amd64 build | 5-7 min | 30-60 sec | 1-2 min |
| arm64 build | 5-7 min | 30-60 sec | 1-2 min |
| Push time | 1-2 min | 30 sec | 30 sec |
| Cache hit rate | 0% | 85% | 60% |
### Cache Performance Model
```python
def estimate_build_time(changes):
base_time = 60 # seconds (setup + push)
if "Dockerfile" in changes:
return base_time + (10 * 60) # Full rebuild: ~11 min
elif "requirements.txt" in changes:
return base_time + (3 * 60) # Deps rebuild: ~4 min
elif any(f.endswith(".py") for f in changes):
return base_time + 60 # Code only: ~2 min
else:
return base_time # No changes: ~1 min
```
---
## Scalability Considerations
### Current Limits
| Resource | Limit | Impact |
|----------|-------|--------|
| Workflow concurrency | 20 (default) | Max 20 releases in parallel |
| Artifact storage | 500 MB/artifact | PyPI packages small (<10 MB) |
| Cache storage | 10 GB/repo | Docker layers fit comfortably |
| Workflow run time | 6 hours | Plenty of headroom |
### Scaling Strategies
#### Horizontal Scaling (Multiple Repos)
```
crawl4ai (main)
├─ release.yml
└─ docker-release.yml
crawl4ai-plugins (separate)
├─ release.yml
└─ docker-release.yml
Each repo has independent:
- Secrets
- Cache (10 GB each)
- Concurrency limits (20 each)
```
#### Vertical Scaling (Larger Runners)
```yaml
jobs:
docker:
runs-on: ubuntu-latest-8-cores # GitHub-hosted larger runner
# 4x faster builds for CPU-bound layers
```
---
## Disaster Recovery
### Failure Scenarios
#### Scenario 1: Release Pipeline Fails
**Failure Point**: PyPI upload fails (network error)
**State**:
- ✓ Version validated
- ✓ Package built
- ✗ PyPI upload
- ✗ GitHub release
**Recovery**:
```bash
# Manual upload
twine upload dist/*
# Retry workflow (re-run from GitHub Actions UI)
```
**Prevention**: Add retry logic to PyPI upload
#### Scenario 2: Docker Pipeline Fails
**Failure Point**: ARM build fails (dependency issue)
**State**:
- ✓ PyPI published
- ✓ GitHub release created
- ✓ amd64 image built
- ✗ arm64 image build
**Recovery**:
```bash
# Fix Dockerfile
git commit -am "fix: ARM build dependency"
# Trigger rebuild
git tag docker-rebuild-v1.2.3
git push origin docker-rebuild-v1.2.3
```
**Impact**: PyPI package available, only Docker ARM users affected
#### Scenario 3: Partial Release
**Failure Point**: GitHub release creation fails
**State**:
- ✓ PyPI published
- ✗ GitHub release
- ✗ Docker images
**Recovery**:
```bash
# Create release manually
gh release create v1.2.3 \
--title "Release v1.2.3" \
--notes "..."
# This triggers docker-release.yml automatically
```
---
## Monitoring and Observability
### Metrics to Track
#### Release Pipeline
- Success rate (target: >99%)
- Duration (target: <3 min)
- PyPI upload time (target: <60 sec)
#### Docker Pipeline
- Success rate (target: >95%)
- Duration (target: <15 min cold, <2 min warm)
- Cache hit rate (target: >80% for code changes)
### Alerting
**Critical Alerts**:
- Release pipeline failure (blocks release)
- PyPI authentication failure (expired token)
**Warning Alerts**:
- Docker build >15 min (performance degradation)
- Cache hit rate <50% (cache issue)
### Logging
**GitHub Actions Logs**:
- Retention: 90 days
- Downloadable: Yes
- Searchable: Limited
**Recommended External Logging**:
```yaml
- name: Send logs to external service
if: failure()
run: |
curl -X POST https://logs.example.com/api/v1/logs \
-H "Content-Type: application/json" \
-d "{\"workflow\": \"${{ github.workflow }}\", \"status\": \"failed\"}"
```
---
## Future Enhancements
### Planned Improvements
1. **Automated Changelog Generation**
- Use conventional commits
- Generate CHANGELOG.md automatically
2. **Pre-release Testing**
- Test builds on `test-v*` tags
- Upload to TestPyPI
3. **Notification System**
- Slack/Discord notifications on release
- Email on failure
4. **Performance Optimization**
- Parallel Docker builds (amd64 + arm64 simultaneously)
- Persistent runners for better caching
5. **Enhanced Validation**
- Smoke tests after PyPI upload
- Container security scanning
---
## References
- [GitHub Actions Architecture](https://docs.github.com/en/actions/learn-github-actions/understanding-github-actions)
- [Docker Build Cache](https://docs.docker.com/build/cache/)
- [PyPI API Documentation](https://warehouse.pypa.io/api-reference/)
---
**Last Updated**: 2025-01-21
**Version**: 2.0