update_document
Modify existing documents in a knowledge base by replacing content and re-indexing for accurate retrieval in local RAG systems.
Instructions
Update an existing document in the knowledge base.
Removes old chunks and re-indexes with new content.
Args:
filepath: Full path to the document file
content: New content for the document
Returns:
JSON string with update resultsInput Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| filepath | Yes | ||
| content | Yes |
Output Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| result | Yes |
Implementation Reference
- mcp_server/server.py:982-1039 (handler)The 'update_document_content' method in 'KnowledgeOrchestrator' handles the removal of old document chunks and re-indexing of updated content.
def update_document_content(self, filepath: str, content: str) -> Dict[str, Any]: """Update an existing document. Removes old chunks and re-indexes.""" filepath = Path(filepath) if not filepath.exists(): return {"error": f"File not found: {filepath}"} # Resolve to absolute for consistent comparison with stored metadata filepath_resolved = str(filepath.resolve()) doc_id = None for did, info in self._indexed_docs.items(): stored = str(Path(info.get("source", "")).resolve()) if stored == filepath_resolved: doc_id = did break old_chunks_removed = 0 if doc_id: old_chunks_removed = self._remove_document_chunks(doc_id) del self._indexed_docs[doc_id] filepath.write_text(content, encoding="utf-8") doc = self.parser.parse_file(filepath) if not doc: self._save_metadata() return {"error": "Failed to parse updated content", "old_chunks_removed": old_chunks_removed} new_chunks_added, dedup_skipped = self._index_document(doc) try: file_stat = filepath.stat() file_mtime = datetime.fromtimestamp(file_stat.st_mtime).isoformat() file_size = file_stat.st_size except OSError: file_mtime = datetime.now().isoformat() file_size = 0 self._indexed_docs[doc.id] = { "source": str(filepath), "category": doc.category, "format": doc.format, "chunks": new_chunks_added, "keywords": doc.keywords, "indexed_at": datetime.now().isoformat(), "file_mtime": file_mtime, "file_size": file_size, } self._save_metadata() self.query_cache.invalidate() self.bm25_index.build_index() return { "old_chunks_removed": old_chunks_removed, "new_chunks_added": new_chunks_added, "dedup_skipped": dedup_skipped, "filepath": str(filepath), } - mcp_server/server.py:1452-1476 (registration)The MCP tool 'update_document' registration and entry point.
def update_document(filepath: str, content: str) -> str: """ Update an existing document in the knowledge base. Removes old chunks and re-indexes with new content. Args: filepath: Full path to the document file content: New content for the document Returns: JSON string with update results """ if not filepath: return json.dumps({"status": "error", "message": "Filepath required"}) if not content or not content.strip(): return json.dumps({"status": "error", "message": "Content cannot be empty"}) orchestrator = get_orchestrator() result = orchestrator.update_document_content(filepath, content.strip()) if "error" in result: return json.dumps({"status": "error", "message": result["error"]}) return json.dumps({"status": "success", **result}, indent=2)