fix

.env

@@ -8,7 +8,7 @@ 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_ROOT="C:/Users/DELL/Desktop/Deploymnet/29 oct/py-learn-backend/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
 RAG_INGEST_URL=http://localhost:5000/rag/ingest

ragg/rag_backend.py

@@ -19,7 +19,10 @@ except Exception:
 
 from langchain_community.embeddings import HuggingFaceEmbeddings
 from langchain_community.vectorstores import Chroma
-from langchain.schema import Document
+try:
+    from langchain_core.documents import Document   # LC >= 0.2
+except Exception:
+    from langchain.schema import Document    
 from pdf2image import convert_from_path
 from PIL import Image  # noqa: F401  (used implicitly via pdf2image)
 import pytesseract

requirements.txt

@@ -15,10 +15,12 @@ boto3==1.35.13
 ffmpeg-python==0.2.0
 imageio-ffmpeg
 pydantic>=2
-chromadb
+chromadb==0.5.4
 langchain>=0.3
-langchain-community>=0.3
-langchain-text-splitters>=0.3
+langchain-core==0.2.38
+langchain-community==0.2.11
+langchain-text-splitters==0.2.2
+sentence-transformers==2.2.2
 pypdf>=4
 tiktoken