fix

.env

@@ -7,7 +7,7 @@ DID_API_KEY=b3ZpeWEuckBweWthcmEubmV0:FMWfsvU5tLYIeVzY0fyBG
 DID_SOURCE_IMAGE_URL=https://i.ibb.co/Tpq77ZJ/teacher.png
 DID_VOICE_ID=en-US-JennyNeural
 TESSERACT_CMD=C:\Program Files\Tesseract-OCR\tesseract.exe
-CHROMA_DIR=C:\path\to\your\project\chroma
+CHROMA_DIR=C:/Users/DELL/Desktop/Deploymnet/29 oct/py-learn-backend/ragg/chroma
 CHROMA_ROOT=C:/Users/DELL/Desktop/Deploymnet/29 oct/py-learn-backend/ragg/chroma
 EMBEDDING_MODEL=sentence-transformers/all-MiniLM-L6-v2
 ALLOWED_ORIGINS=http://localhost:4200,http://127.0.0.1:4200

ragg/app.py

@@ -264,9 +264,9 @@ def rag_suggest_followups():
     return jsonify(result)
 
 
+# @rag_bp.get("/_diag")
 @rag_bp.get("/_diag")
 def rag_diag():
-    # minimal imports here to avoid circulars
     try:
         from .rag_llm import CHROMA_DIR, CHROMA_ROOT, get_vectorstore, get_vectorstore_for
     except ImportError:
@@ -276,27 +276,79 @@ def rag_diag():
     from flask import jsonify
 
     def _count(vs):
-        try:
-            return vs._collection.count()
-        except Exception:
+        """Handle both LangChain and chromadb client objects."""
+        if vs is None:
+            return None
+        # 1️⃣ chromadb.Collection (your new get_vectorstore_for)
+        if hasattr(vs, "count") and callable(vs.count):
             try:
-                return vs._client.get_collection(vs._collection.name).count()
+                return vs.count()
             except Exception:
                 return None
+        # 2️⃣ LangChain vectorstore
+        if hasattr(vs, "_collection"):
+            try:
+                return vs._collection.count()  # type: ignore
+            except Exception:
+                try:
+                    return vs._client.get_collection(vs._collection.name).count()  # type: ignore
+                except Exception:
+                    return None
+        return None
+
+    # load each level safely
+    low_vs = get_vectorstore_for("low")
+    mid_vs = get_vectorstore_for("mid")
+    high_vs = get_vectorstore_for("high")
 
     info = {
-        "env_seen": {"CHROMA_DIR": CHROMA_DIR, "CHROMA_ROOT": CHROMA_ROOT},
+        "env_seen": {
+            "CHROMA_DIR": CHROMA_DIR,
+            "CHROMA_ROOT": CHROMA_ROOT
+        },
         "low_dir": {
             "path": os.path.join(CHROMA_ROOT, "low"),
             "exists": os.path.isdir(os.path.join(CHROMA_ROOT, "low")),
         },
         "counts_default": _count(get_vectorstore()),
-        "counts_low": _count(get_vectorstore_for("low")),
-        "counts_mid": _count(get_vectorstore_for("mid")),
-        "counts_high": _count(get_vectorstore_for("high")),
+        "counts_low": _count(low_vs),
+        "counts_mid": _count(mid_vs),
+        "counts_high": _count(high_vs),
     }
     return jsonify(info), 200
 
+# def rag_diag():
+#     # minimal imports here to avoid circulars
+#     try:
+#         from .rag_llm import CHROMA_DIR, CHROMA_ROOT, get_vectorstore, get_vectorstore_for
+#     except ImportError:
+#         from rag_llm import CHROMA_DIR, CHROMA_ROOT, get_vectorstore, get_vectorstore_for
+
+#     import os
+#     from flask import jsonify
+
+#     def _count(vs):
+#         try:
+#             return vs._collection.count()
+#         except Exception:
+#             try:
+#                 return vs._client.get_collection(vs._collection.name).count()
+#             except Exception:
+#                 return None
+
+#     info = {
+#         "env_seen": {"CHROMA_DIR": CHROMA_DIR, "CHROMA_ROOT": CHROMA_ROOT},
+#         "low_dir": {
+#             "path": os.path.join(CHROMA_ROOT, "low"),
+#             "exists": os.path.isdir(os.path.join(CHROMA_ROOT, "low")),
+#         },
+#         "counts_default": _count(get_vectorstore()),
+#         "counts_low": _count(get_vectorstore_for("low")),
+#         "counts_mid": _count(get_vectorstore_for("mid")),
+#         "counts_high": _count(get_vectorstore_for("high")),
+#     }
+#     return jsonify(info), 200
+
 @rag_bp.route("/search", methods=["POST", "OPTIONS"])
 def rag_search():
     if request.method == "OPTIONS":

ragg/rag_llm.py

@@ -11,6 +11,8 @@ from langchain_core.documents import Document
 from openai import OpenAI
 from dotenv import load_dotenv, find_dotenv
 load_dotenv(find_dotenv())
+import chromadb
+from chromadb.utils import embedding_functions
 
 # --- Constants ---
 # CHROMA_DIR = "./chroma"
@@ -55,28 +57,51 @@ def get_vectorstore():
         )
     return _vectorstore
 
-
 def get_vectorstore_for(db_level: Optional[str] = None):
-    """
-    Return a persistent Chroma vectorstore for the requested db_level.
-    db_level in {"low","mid","high"} → ./chroma/<db_level>
-    else → fall back to your original CHROMA_DIR (single-store).
-    """
     key = (db_level or "").strip().lower()
     if key in ("low", "mid", "high"):
         persist_dir = os.path.join(CHROMA_ROOT, key)
-        print(f"[RAG] get_vectorstore_for('{key}') -> {persist_dir}")
     else:
-        persist_dir = CHROMA_DIR  # fallback
-        print(f"[RAG] get_vectorstore_for(None) -> default ({CHROMA_DIR})")
-
-    if key not in _vectorstores:
-        print(f"🔹 Loading Chroma at: {persist_dir}")
-        _vectorstores[key] = Chroma(
-            persist_directory=persist_dir,
-            embedding_function=get_embeddings(),
-        )
-    return _vectorstores[key]
+        persist_dir = CHROMA_DIR
+
+    print(f"[RAG] Using Chroma from: {persist_dir}")
+
+    client = chromadb.PersistentClient(path=persist_dir)
+
+    # Show collections available
+    collections = client.list_collections()
+    print(f"Available collections: {[c.name for c in collections]}")
+
+    # Pick the default collection (first one)
+    if not collections:
+        print("❌ No collections found.")
+        return None
+    collection = client.get_collection(collections[0].name)
+    print(f"✅ Loaded Chroma collection: {collection.name}")
+    return collection
+
+# def get_vectorstore_for(db_level: Optional[str] = None):
+#     """
+#     Return a persistent Chroma vectorstore for the requested db_level.
+#     db_level in {"low","mid","high"} → ./chroma/<db_level>
+#     else → fall back to your original CHROMA_DIR (single-store).
+#     """
+#     key = (db_level or "").strip().lower()
+#     if key in ("low", "mid", "high"):
+#         persist_dir = os.path.join(CHROMA_ROOT, key)
+#         print(f"[RAG] get_vectorstore_for('{key}') -> {persist_dir}")
+#     else:
+#         persist_dir = CHROMA_DIR  # fallback
+#         print(f"[DEBUG] Using Chroma folder for level '{key or 'default'}' → {persist_dir}")
+#         print(f"[RAG] get_vectorstore_for(None) -> default ({CHROMA_DIR})")
+
+#     if key not in _vectorstores:
+#         print(f"🔹 Loading Chroma at: {persist_dir}")
+#         _vectorstores[key] = Chroma(
+#             persist_directory=persist_dir,
+#             embedding_function=get_embeddings(),
+#         )
+#     return _vectorstores[key]
 
 
 def get_client():