update faiss index

README.md

@@ -1,19 +1,43 @@
----
-title: Agriquery
-emoji: 🦀
-colorFrom: green
-colorTo: red
-sdk: docker
-app_port: 8501
-tags:
-- streamlit
-pinned: false
-short_description: LLM-powered question-answering system using RAG
----
-
-# Welcome to Streamlit!
-
-Edit `/src/streamlit_app.py` to customize this app to your heart's desire. :heart:
-
-If you have any questions, checkout our [documentation](https://docs.streamlit.io) and [community
-forums](https://discuss.streamlit.io).
+## AGRIQUERY - RAG-LLM Powered Q&A App for Agricultural Researchers.
+
+AgriQuery is an LLM-powered Q&A system built for agricultural researchers. It processes scientific publications and enables users to ask natural language questions, receiving context-aware answers backed by retrieved text. Built with LangChain, FAISS, Airflow, and Docker, it demonstrates a production-ready RAG architecture for domain-specific information retrieval.
+
+### End-to-End RAG with Airflow, FAISS, Llama, FastAPI
+
+## Usage
+1. `docker-compose up --build`
+2. Access:
+   - Airflow: http://localhost:8080
+   - FastAPI: http://localhost:8000/query?question=Your+question+here
+3. Run DAG `rag_pipeline` in Airflow UI to ingest & build FAISS.
+4. Query the running FastAPI endpoint.
+
+
+```
+agriquery-rag-llm/
+│
+├── app/
+│   ├── rag.py               # LangChain RAG setup
+│   └── main_app.py          # Streamlit or Gradio UI
+├── docker/
+│   ├── Dockerfile
+│   └── docker-compose.yml   
+│
+├── src/
+│   └── experiments.ipynb    # Testing LangChain chains and embeddings
+│   └── prepocess.py         # pdf upload, chunking and embedding setup
+├── requirements.txt
+├── .env                     # API keys
+├── README.md
+
+
+```
+pip install transformers langchain langchain_community sentence-transformers faiss-cpu
+
+## Complete Local RAG Stack
+Component	       Tool
+Chunking	LangChain Splitters
+Embedding	Sentence Transformers
+Vector DB	FAISS
+LLM	        HuggingFace Llama (Locally)
+RAG Chain	LangChain RetrievalQA

src/streamlit_app.py

@@ -50,7 +50,7 @@ prompt = PromptTemplate(
 @st.cache_resource
 def load_retriever():
     embeddings = HuggingFaceEmbeddings(model_name="all-MiniLM-L6-v2")
-    db = FAISS.load_local("./vectorstore/agriquery_faiss_index", embeddings, allow_dangerous_deserialization=True)
+    db = FAISS.load_local("./vectorstore", embeddings, allow_dangerous_deserialization=True)
     retriever = db.as_retriever()
     return retriever