"""
Crawl4AI MCP Server - A powerful RAG-enhanced web crawling and processing tool.
This package provides MCP (Model Context Protocol) tools for web crawling,
content extraction, and intelligent document processing with vector search capabilities.
"""
__version__ = "0.1.0"
# Standard library imports
import importlib.util
import sys
from pathlib import Path
# Re-export commonly used utilities for easier imports in tests
from src.config.settings import Settings, get_settings, reset_settings
from src.core.context import (
Crawl4AIContext,
format_neo4j_error,
get_app_context,
set_app_context,
)
from src.core.decorators import track_request
from src.core.exceptions import (
Crawl4AIError,
DatabaseError,
MCPToolError,
NetworkError,
ValidationError,
)
from src.core.stdout_utils import SuppressStdout
from src.database.factory import (
create_and_initialize_database,
create_database_client,
)
from src.database.rag_queries import get_available_sources
from src.knowledge_graph.queries import query_knowledge_graph
from src.knowledge_graph.repository import parse_github_repository
from src.main import mcp
from src.services.crawling import process_urls_for_mcp
from src.services.search import search_and_process
from src.services.smart_crawl import smart_crawl_url
from src.tools import register_tools
from src.utils.code_analysis import process_code_example
from src.utils.embeddings import create_embedding, create_embeddings_batch
from src.utils.reranking import rerank_results
from src.utils.text_processing import extract_section_info, smart_chunk_markdown
from src.utils.url_helpers import (
is_sitemap,
is_txt,
parse_sitemap,
)
from src.utils.validation import (
validate_github_url,
validate_neo4j_connection,
validate_script_path,
)
# Knowledge graph operations
# Note: There's a naming conflict - validation.py file vs validation/ directory
# We need to import from the parent module with specific file reference
# Load validation.py module directly to avoid package conflict
# Use relative path resolution to work in both dev and Docker environments
_validation_path = Path(__file__).parent / "knowledge_graph" / "validation.py"
_validation_spec = importlib.util.spec_from_file_location(
"src.knowledge_graph._validation_module",
str(_validation_path),
)
if _validation_spec and _validation_spec.loader:
_validation_module = importlib.util.module_from_spec(_validation_spec)
sys.modules["src.knowledge_graph._validation_module"] = _validation_module
_validation_spec.loader.exec_module(_validation_module)
check_ai_script_hallucinations = _validation_module.check_ai_script_hallucinations
else:
check_ai_script_hallucinations = None
# Export registered MCP tool functions for tests
# These are dynamically created by @mcp.tool() decorators
# Note: Tool functions are available through the mcp instance directly
scrape_urls = None
perform_rag_query = None
search = None
search_code_examples = None
smart_crawl_url_tool = None
__all__ = [
"Crawl4AIContext",
"Crawl4AIError",
"DatabaseError",
"MCPToolError",
"NetworkError",
"Settings",
"SuppressStdout",
"ValidationError",
"__version__",
"check_ai_script_hallucinations",
"create_and_initialize_database",
"create_database_client",
"create_embedding",
"create_embeddings_batch",
"extract_section_info",
"format_neo4j_error",
"get_app_context",
"get_available_sources",
"get_settings",
"is_sitemap",
"is_txt",
"mcp",
"parse_github_repository",
"parse_sitemap",
"perform_rag_query",
"process_code_example",
"process_urls_for_mcp",
"query_knowledge_graph",
"register_tools",
"rerank_results",
"reset_settings",
"scrape_urls",
"search",
"search_and_process",
"search_code_examples",
"set_app_context",
"smart_chunk_markdown",
"smart_crawl_url",
"smart_crawl_url_tool",
"track_request",
"validate_github_url",
"validate_neo4j_connection",
"validate_script_path",
]
