ch6

app.py

@@ -3,6 +3,7 @@ from sentence_transformers import SentenceTransformer
 import uvicorn
 import os
 from pathlib import Path
+import chromadb
 from chromadb.config import Settings
 from chromadb.server.fastapi import FastAPI as ChromaFastAPI
 import torch
@@ -72,10 +73,33 @@ def write_bucket_probe(path: str) -> None:
     )
 
 
+def seed_chroma_data(path: str) -> None:
+    client = chromadb.PersistentClient(path=path)
+    collection = client.get_or_create_collection(name="knowledge_base")
+    if collection.count() > 0:
+        logger.info("[*] Chroma already has data; skipping seed.")
+        return
+
+    documents = [
+        "Chroma is a lightweight, open-source vector database built for AI.",
+        "Python is a high-level programming language used extensively in data science.",
+        "The celestial body closest to Earth is the Moon.",
+    ]
+    metadatas = [
+        {"category": "tech", "source": "docs"},
+        {"category": "tech", "source": "wiki"},
+        {"category": "science", "source": "space-facts"},
+    ]
+    ids = ["doc1", "doc2", "doc3"]
+    collection.add(documents=documents, metadatas=metadatas, ids=ids)
+    logger.info("[+] Seeded Chroma with sample documents.")
+
+
 @app.on_event("startup")
 def load_model():
     global model
     write_bucket_probe(chroma_persist_directory)
+    seed_chroma_data(chroma_persist_directory)
     logger.info(f"[*] Loading Harrier OSS 0.6B model: {MODEL_NAME}...")
     try:
         model = SentenceTransformer(MODEL_NAME, trust_remote_code=True, device=device)

vector.py

@@ -0,0 +1,72 @@
+import argparse
+from pathlib import Path
+from typing import List
+
+import chromadb
+
+
+def _parse_args() -> argparse.Namespace:
+    parser = argparse.ArgumentParser(description="Local ChromaDB persistence demo")
+    parser.add_argument(
+        "--path",
+        default="chroma_data",
+        help="Local persistence directory",
+    )
+    parser.add_argument(
+        "--collection",
+        default="knowledge_base",
+        help="Collection name",
+    )
+    parser.add_argument(
+        "--query",
+        default="Tell me about vector stores",
+        help="Query text",
+    )
+    return parser.parse_args()
+
+
+def _seed_collection(collection: chromadb.Collection) -> None:
+    documents = [
+        "Chroma is a lightweight, open-source vector database built for AI.",
+        "Python is a high-level programming language used extensively in data science.",
+        "The celestial body closest to Earth is the Moon.",
+    ]
+    metadatas = [
+        {"category": "tech", "source": "docs"},
+        {"category": "tech", "source": "wiki"},
+        {"category": "science", "source": "space-facts"},
+    ]
+    ids = ["doc1", "doc2", "doc3"]
+    collection.add(documents=documents, metadatas=metadatas, ids=ids)
+
+
+def main() -> None:
+    args = _parse_args()
+    persist_path = Path(args.path).resolve()
+    persist_path.mkdir(parents=True, exist_ok=True)
+
+    print(f"Using local Chroma persistence at: {persist_path}")
+    client = chromadb.PersistentClient(path=str(persist_path))
+
+    collection = client.get_or_create_collection(name=args.collection)
+    if collection.count() == 0:
+        print("Seeding collection with sample documents...")
+        _seed_collection(collection)
+    print(f"Collection '{args.collection}' has {collection.count()} documents.")
+
+    results = collection.query(query_texts=[args.query], n_results=2)
+    print("\n--- Search Results ---")
+    for doc, meta, distance in zip(
+        results["documents"][0],
+        results["metadatas"][0],
+        results["distances"][0],
+    ):
+        print(f"Matched Document: {doc}")
+        print(f"Metadata: {meta}")
+        print(f"Distance Score (Lower is better): {distance:.4f}")
+        print()
+    print("----------------------")
+
+
+if __name__ == "__main__":
+    main()