intentkit 0.5.2__py3-none-any.whl → 0.6.0__py3-none-any.whl

intentkit/skills/web_scraper/website_indexer.py

@@ -0,0 +1,456 @@
+import logging
+from typing import List, Type
+from urllib.parse import urljoin, urlparse
+
+import httpx
+import openai
+from langchain_core.runnables import RunnableConfig
+from pydantic import BaseModel, Field
+
+from intentkit.skills.web_scraper.base import WebScraperBaseTool
+from intentkit.skills.web_scraper.utils import (
+    DEFAULT_CHUNK_OVERLAP,
+    DEFAULT_CHUNK_SIZE,
+    MetadataManager,
+    ResponseFormatter,
+    VectorStoreManager,
+    scrape_and_index_urls,
+)
+
+logger = logging.getLogger(__name__)
+
+
+class WebsiteIndexerInput(BaseModel):
+    """Input for WebsiteIndexer tool."""
+
+    base_url: str = Field(
+        description="Base URL of the website to index (e.g., https://example.com). The tool will discover sitemaps and extract all URLs",
+        min_length=1,
+    )
+    max_urls: int = Field(
+        description="Maximum number of URLs to scrape from the sitemap (default: 50)",
+        default=50,
+        ge=1,
+        le=200,
+    )
+    chunk_size: int = Field(
+        description="Size of text chunks for indexing (default: 1000)",
+        default=DEFAULT_CHUNK_SIZE,
+        ge=100,
+        le=4000,
+    )
+    chunk_overlap: int = Field(
+        description="Overlap between chunks (default: 200)",
+        default=DEFAULT_CHUNK_OVERLAP,
+        ge=0,
+        le=1000,
+    )
+    include_patterns: List[str] = Field(
+        description="URL patterns to include (e.g., ['/blog/', '/docs/']). If empty, all URLs are included",
+        default=[],
+    )
+    exclude_patterns: List[str] = Field(
+        description="URL patterns to exclude (e.g., ['/admin/', '/private/'])",
+        default=[],
+    )
+
+
+class WebsiteIndexer(WebScraperBaseTool):
+    """Tool for discovering and indexing entire websites using AI-powered sitemap analysis.
+
+    This tool discovers sitemaps from robots.txt, extracts URLs from sitemap XML using GPT-4o-mini for
+    robust parsing of various sitemap formats, and then delegates to the proven scrape_and_index tool
+    for reliable content indexing.
+    """
+
+    name: str = "web_scraper_website_indexer"
+    description: str = (
+        "Index an entire website by discovering sitemaps and extracting URLs efficiently. "
+        "This tool finds sitemaps from robots.txt, parses the XML content to extract URLs, "
+        "and then uses the reliable scrape_and_index functionality for content indexing."
+    )
+    args_schema: Type[BaseModel] = WebsiteIndexerInput
+
+    def _normalize_url(self, url: str) -> str:
+        """Normalize URL by ensuring it has a proper scheme."""
+        if not url.startswith(("http://", "https://")):
+            return f"https://{url}"
+        return url
+
+    async def _get_robots_txt(self, base_url: str) -> str:
+        """Fetch robots.txt content."""
+        robots_url = urljoin(base_url, "/robots.txt")
+
+        # Import headers from utils
+        from intentkit.skills.web_scraper.utils import DEFAULT_HEADERS, FALLBACK_HEADERS
+
+        # Try with primary headers first
+        async with httpx.AsyncClient(timeout=30, headers=DEFAULT_HEADERS) as client:
+            try:
+                response = await client.get(robots_url)
+                if response.status_code == 200:
+                    return response.text
+            except Exception as e:
+                logger.warning(
+                    f"Primary headers failed for robots.txt from {robots_url}: {e}"
+                )
+
+        # Try with fallback headers
+        async with httpx.AsyncClient(timeout=30, headers=FALLBACK_HEADERS) as client:
+            try:
+                response = await client.get(robots_url)
+                if response.status_code == 200:
+                    return response.text
+            except Exception as e:
+                logger.warning(f"Could not fetch robots.txt from {robots_url}: {e}")
+        return ""
+
+    def _extract_sitemaps_from_robots(
+        self, robots_content: str, base_url: str
+    ) -> List[str]:
+        """Extract sitemap URLs from robots.txt content."""
+        sitemaps = []
+
+        for line in robots_content.split("\n"):
+            line = line.strip()
+            if line.lower().startswith("sitemap:"):
+                sitemap_url = line.split(":", 1)[1].strip()
+                # Make relative URLs absolute
+                if sitemap_url.startswith("/"):
+                    sitemap_url = urljoin(base_url, sitemap_url)
+                sitemaps.append(sitemap_url)
+
+        return sitemaps
+
+    def _get_common_sitemap_patterns(self, base_url: str) -> List[str]:
+        """Generate common sitemap URL patterns."""
+        return [
+            urljoin(base_url, "/sitemap.xml"),
+            urljoin(base_url, "/sitemap_index.xml"),
+            urljoin(base_url, "/sitemaps/sitemap.xml"),
+            urljoin(base_url, "/sitemap/sitemap.xml"),
+            urljoin(base_url, "/wp-sitemap.xml"),  # WordPress
+        ]
+
+    async def _fetch_sitemap_content(self, sitemap_url: str) -> str:
+        """Fetch sitemap XML content."""
+        # Import headers from utils
+        from intentkit.skills.web_scraper.utils import DEFAULT_HEADERS, FALLBACK_HEADERS
+
+        # Try with primary headers first
+        async with httpx.AsyncClient(timeout=30, headers=DEFAULT_HEADERS) as client:
+            try:
+                response = await client.get(sitemap_url)
+                if response.status_code == 200:
+                    return response.text
+            except Exception as e:
+                logger.warning(
+                    f"Primary headers failed for sitemap from {sitemap_url}: {e}"
+                )
+
+        # Try with fallback headers
+        async with httpx.AsyncClient(timeout=30, headers=FALLBACK_HEADERS) as client:
+            try:
+                response = await client.get(sitemap_url)
+                if response.status_code == 200:
+                    return response.text
+            except Exception as e:
+                logger.warning(f"Could not fetch sitemap from {sitemap_url}: {e}")
+        return ""
+
+    async def _get_all_sitemap_content(self, base_url: str) -> tuple[str, List[str]]:
+        """Get all sitemap content for AI analysis."""
+        all_content = []
+        found_sitemaps = []
+        processed_sitemaps = set()
+
+        # First, try to get sitemaps from robots.txt
+        robots_content = await self._get_robots_txt(base_url)
+        sitemap_urls = self._extract_sitemaps_from_robots(robots_content, base_url)
+
+        # If no sitemaps found in robots.txt, try common patterns
+        if not sitemap_urls:
+            sitemap_urls = self._get_common_sitemap_patterns(base_url)
+
+        logger.info(f"Checking {len(sitemap_urls)} potential sitemap URLs...")
+
+        # Process each sitemap URL
+        sitemaps_to_process = sitemap_urls[:]
+
+        while sitemaps_to_process:
+            sitemap_url = sitemaps_to_process.pop(0)
+
+            if sitemap_url in processed_sitemaps:
+                continue
+
+            processed_sitemaps.add(sitemap_url)
+
+            xml_content = await self._fetch_sitemap_content(sitemap_url)
+            if not xml_content:
+                continue
+
+            found_sitemaps.append(sitemap_url)
+            all_content.append(f"<!-- Sitemap: {sitemap_url} -->\n{xml_content}\n")
+
+            # Check if this contains references to other sitemaps (sitemap index)
+            if "<sitemap>" in xml_content.lower() and "<loc>" in xml_content.lower():
+                # This might be a sitemap index - we'll let AI handle parsing it
+                pass
+
+        combined_xml = "\n".join(all_content) if all_content else ""
+        return combined_xml, found_sitemaps
+
+    def _create_ai_extraction_prompt(
+        self, sitemap_xml: str, include_patterns: List[str], exclude_patterns: List[str]
+    ) -> str:
+        """Create a prompt for AI to extract URLs from sitemap XML."""
+        filter_instructions = ""
+        if include_patterns:
+            filter_instructions += f"\n- INCLUDE only URLs containing these patterns: {', '.join(include_patterns)}"
+        if exclude_patterns:
+            filter_instructions += f"\n- EXCLUDE URLs containing these patterns: {', '.join(exclude_patterns)}"
+
+        return f"""Analyze this sitemap XML and extract all valid webpage URLs.
+
+SITEMAP XML CONTENT:
+{sitemap_xml}
+
+INSTRUCTIONS:
+- Extract only URLs from <loc> tags that point to actual web pages
+- Handle both standard sitemap format and sitemap index format
+- Ignore any URLs ending in .xml, .rss, .atom (these are feeds/sitemaps, not pages)
+- Skip any sitemap index entries that point to other sitemaps  
+- Handle text-based sitemaps (simple URL lists)
+- Return only unique, valid HTTP/HTTPS URLs
+- Format as a simple list, one URL per line{filter_instructions}
+
+Extract the URLs now:"""
+
+    def _parse_ai_response(self, ai_response: str) -> List[str]:
+        """Parse AI response to extract clean URLs."""
+        urls = []
+
+        for line in ai_response.strip().split("\n"):
+            line = line.strip()
+            # Remove any markdown formatting, bullets, numbering
+            line = line.lstrip("- •*123456789. ")
+
+            # Check if it looks like a URL
+            if line.startswith(("http://", "https://")):
+                # Basic validation
+                try:
+                    parsed = urlparse(line)
+                    if parsed.netloc and not line.endswith((".xml", ".rss", ".atom")):
+                        urls.append(line)
+                except Exception:
+                    continue
+
+        return list(set(urls))  # Remove duplicates
+
+    async def _call_ai_model(self, prompt: str, context) -> str:
+        """Call OpenAI GPT-4o-mini to extract URLs from sitemap content."""
+        try:
+            # Get OpenAI API key using the standard pattern
+            from intentkit.skills.openai.base import OpenAIBaseTool
+
+            temp_tool = OpenAIBaseTool(skill_store=self.skill_store)
+            api_key = temp_tool.get_api_key(context)
+
+            # Initialize OpenAI client
+            client = openai.AsyncOpenAI(api_key=api_key)
+
+            # Call the API
+            response = await client.chat.completions.create(
+                model="gpt-4o-mini",
+                messages=[
+                    {
+                        "role": "system",
+                        "content": "You are an expert at parsing XML sitemaps and extracting webpage URLs. Always return only clean, valid URLs, one per line.",
+                    },
+                    {"role": "user", "content": prompt},
+                ],
+                max_tokens=2000,
+                temperature=0.1,
+            )
+
+            return response.choices[0].message.content.strip()
+
+        except Exception as e:
+            logger.error(f"Error calling OpenAI API: {e}")
+            raise
+
+    async def _arun(
+        self,
+        base_url: str,
+        max_urls: int = 50,
+        chunk_size: int = DEFAULT_CHUNK_SIZE,
+        chunk_overlap: int = DEFAULT_CHUNK_OVERLAP,
+        include_patterns: List[str] = None,
+        exclude_patterns: List[str] = None,
+        config: RunnableConfig = None,
+        **kwargs,
+    ) -> str:
+        """Discover website sitemaps, extract URLs with AI, and delegate to scrape_and_index."""
+        try:
+            # Normalize inputs
+            base_url = self._normalize_url(base_url)
+            include_patterns = include_patterns or []
+            exclude_patterns = exclude_patterns or []
+
+            # Validate base URL
+            parsed_url = urlparse(base_url)
+            if not parsed_url.netloc:
+                return "Error: Invalid base URL provided. Please provide a valid URL (e.g., https://example.com)"
+
+            # Get agent context - throw error if not available
+            if not config:
+                raise ValueError("Configuration is required but not provided")
+
+            context = self.context_from_config(config)
+            if not context or not context.agent_id:
+                raise ValueError("Agent ID is required but not found in configuration")
+
+            agent_id = context.agent_id
+
+            logger.info(f"[{agent_id}] Discovering sitemaps for {base_url}...")
+
+            # Get all sitemap content
+            sitemap_xml, found_sitemaps = await self._get_all_sitemap_content(base_url)
+
+            if not sitemap_xml:
+                logger.error(
+                    f"[{agent_id}] No accessible sitemaps found for {base_url}"
+                )
+                return f"Error: No accessible sitemaps found for {base_url}. The website might not have sitemaps or they might be inaccessible."
+
+            logger.info(
+                f"[{agent_id}] Found {len(found_sitemaps)} sitemap(s). Extracting URLs with AI..."
+            )
+
+            try:
+                # Use AI to extract URLs from sitemap
+                prompt = self._create_ai_extraction_prompt(
+                    sitemap_xml, include_patterns, exclude_patterns
+                )
+                ai_response = await self._call_ai_model(prompt, context)
+                all_urls = self._parse_ai_response(ai_response)
+
+                logger.info(
+                    f"[{agent_id}] AI extracted {len(all_urls)} URLs from sitemap"
+                )
+
+            except Exception as e:
+                logger.error(
+                    f"[{agent_id}] AI extraction failed: {e}, falling back to regex"
+                )
+                # Fallback to simple regex if AI fails
+                import re
+
+                url_pattern = r"<loc>(https?://[^<]+)</loc>"
+                all_urls = re.findall(url_pattern, sitemap_xml)
+
+                # Basic filtering for fallback
+                filtered_urls = []
+                for url in all_urls:
+                    # Skip XML files (sitemaps)
+                    if url.endswith((".xml", ".rss", ".atom")):
+                        continue
+
+                    # Apply exclude patterns
+                    if exclude_patterns and any(
+                        pattern in url for pattern in exclude_patterns
+                    ):
+                        continue
+
+                    # Apply include patterns
+                    if include_patterns:
+                        if any(pattern in url for pattern in include_patterns):
+                            filtered_urls.append(url)
+                    else:
+                        filtered_urls.append(url)
+
+                all_urls = filtered_urls
+                logger.info(
+                    f"[{agent_id}] Regex fallback extracted {len(all_urls)} URLs from sitemap"
+                )
+
+            # Remove duplicates and limit
+            unique_urls = list(set(all_urls))[:max_urls]
+
+            if not unique_urls:
+                logger.error(
+                    f"[{agent_id}] No valid URLs found in sitemaps after filtering"
+                )
+                return f"Error: No valid URLs found in sitemaps after filtering. Found sitemaps: {', '.join(found_sitemaps)}"
+
+            logger.info(
+                f"[{agent_id}] Extracted {len(unique_urls)} URLs from sitemaps. Scraping and indexing..."
+            )
+
+            # Use the utility function to scrape and index URLs directly
+            total_chunks, was_merged, valid_urls = await scrape_and_index_urls(
+                unique_urls, agent_id, self.skill_store, chunk_size, chunk_overlap
+            )
+
+            if total_chunks == 0:
+                logger.error(
+                    f"[{agent_id}] No content could be extracted from discovered URLs"
+                )
+                return f"Error: No content could be extracted from the discovered URLs. Found sitemaps: {', '.join(found_sitemaps)}"
+
+            # Get current storage size for response
+            vs_manager = VectorStoreManager(self.skill_store)
+            current_size = await vs_manager.get_content_size(agent_id)
+            size_limit_reached = len(valid_urls) < len(unique_urls)
+
+            # Update metadata
+            metadata_manager = MetadataManager(self.skill_store)
+            new_metadata = metadata_manager.create_url_metadata(
+                valid_urls, [], "website_indexer"
+            )
+            await metadata_manager.update_metadata(agent_id, new_metadata)
+
+            logger.info(f"[{agent_id}] Website indexing completed successfully")
+
+            # Format the indexing result
+            result = ResponseFormatter.format_indexing_response(
+                "scraped and indexed",
+                valid_urls,
+                total_chunks,
+                chunk_size,
+                chunk_overlap,
+                was_merged,
+                current_size_bytes=current_size,
+                size_limit_reached=size_limit_reached,
+                total_requested_urls=len(unique_urls),
+            )
+
+            # Enhance the response with sitemap discovery info
+            enhanced_result = (
+                f"WEBSITE INDEXING COMPLETE\n"
+                f"Base URL: {base_url}\n"
+                f"Sitemaps discovered: {len(found_sitemaps)}\n"
+                f"URLs extracted: {len(unique_urls)}\n"
+                f"URLs successfully indexed: {len(valid_urls)}\n"
+                f"Include patterns: {', '.join(include_patterns) if include_patterns else 'None (all URLs)'}\n"
+                f"Exclude patterns: {', '.join(exclude_patterns) if exclude_patterns else 'None'}\n\n"
+                f"DISCOVERED SITEMAPS:\n"
+                f"{chr(10).join(['- ' + sitemap for sitemap in found_sitemaps])}\n\n"
+                f"INDEXING RESULTS:\n{result}"
+            )
+
+            return enhanced_result
+
+        except Exception as e:
+            # Extract agent_id for error logging if possible
+            agent_id = "UNKNOWN"
+            try:
+                if config:
+                    context = self.context_from_config(config)
+                    if context and context.agent_id:
+                        agent_id = context.agent_id
+            except Exception:
+                pass
+
+            logger.error(f"[{agent_id}] Error in WebsiteIndexer: {e}", exc_info=True)
+            raise type(e)(f"[agent:{agent_id}]: {e}") from e

intentkit/utils/logging.py

@@ -43,7 +43,7 @@ def setup_logging(env: str, debug: bool = False):
         debug: Debug mode flag
     """
 
-    if env == "local" or debug:
+    if debug:
         # Set up logging configuration for local/debug
         logging.basicConfig(
             level=logging.DEBUG,

{intentkit-0.5.2.dist-info → intentkit-0.6.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: intentkit
-Version: 0.5.2
+Version: 0.6.0
 Summary: Intent-based AI Agent Platform - Core Package
 Project-URL: Homepage, https://github.com/crestal-network/intentkit
 Project-URL: Repository, https://github.com/crestal-network/intentkit