add init_documents.py

init_documents.py

@@ -1,25 +1,20 @@
 """
-Script to load all .txt files from a directory into the knowledge base.
+Script to load all .txt files from a directory into the vector store.
 """
-
+import os
-import argparse
 from pathlib import Path
-from agent import agent
+from chunker import load_and_split
-from langchain_core.messages import HumanMessage
+from vector_store import vector_store
 
-async def init_docs(dir_path: Path):
+if __name__ == "__main__":
-    for txt_file in dir_path.rglob("*.txt"):
+    docs_dir = Path("docs")
-        content = txt_file.read_text(encoding="utf-8")
+    if not docs_dir.exists():
-        title = txt_file.stem
+        print("Docs directory not found.")
-        await agent.ainvoke(
+        exit(1)
-            {"messages": [HumanMessage(content=f"Add document: {title}")]},
+    chunks = load_and_split(docs_dir)
-            {"configurable": {"thread_id": f"init-{txt_file.name}"}},
+    # Each chunk is (text, metadata)
-        )
+    documents = []
-    print("Initialization complete.")
+    for idx, (chunk, meta) in enumerate(chunks):
-
+        documents.append({"content": chunk, "metadata": meta})
-if __name__ == "__main__":  # pragma: no cover
+    vector_store.add_documents(documents)
-    parser = argparse.ArgumentParser(description="Initialize knowledge base from directory")
+    print(f"Loaded {len(documents)} chunks into the vector store.")
-    parser.add_argument("dir", type=Path, help="Directory with text files")
-    args = parser.parse_args()
-    import asyncio
-    asyncio.run(init_docs(args.dir))