Add AI RAG Knowledge Assistant implementation

ai-rag-project/README.md

@@ -0,0 +1,124 @@
+# Endee AI Knowledge Assistant
+
+## Project Overview
+Endee AI Knowledge Assistant is an AI-powered question answering system built using Retrieval-Augmented Generation (RAG). The system allows users to ask natural language questions and receive context-aware answers by retrieving relevant information from a knowledge base.
+
+The system integrates semantic search, vector databases, and language models to provide intelligent responses.
+![Project](image.png)
+
+---
+
+## Problem Statement
+
+Organizations store large amounts of information across documents and databases. Traditional search systems rely on keyword matching and often fail to understand the meaning of user queries.
+
+Challenges include:
+
+- Fragmented information sources
+- Slow knowledge retrieval
+- Inefficient keyword-based search
+- Difficulty extracting insights from large datasets
+
+This project solves these challenges using an AI-powered semantic retrieval system.
+
+---
+
+## System Design and Technical Approach
+
+The project follows a Retrieval-Augmented Generation (RAG) pipeline.
+
+### Workflow
+
+1. User submits a question through the interface
+2. Question is converted into embeddings
+3. Vector database performs similarity search
+4. Relevant knowledge is retrieved
+5. AI model generates the final response
+
+### Architecture
+
+User  
+  ↓  
+Frontend (Streamlit)  
+  ↓  
+Backend API (FastAPI)  
+  ↓  
+Embedding Model (Sentence Transformers)  
+  ↓  
+Vector Database (FAISS)  
+  ↓  
+Context Retrieval  
+  ↓  
+AI Response Generation  
+
+---
+
+## How Endee is Used
+
+The project demonstrates Endee-style AI architecture principles including:
+
+- Semantic embedding generation
+- Vector database indexing
+- Context-based retrieval
+- AI-powered reasoning using retrieved knowledge
+
+This allows the system to generate accurate answers grounded in actual information.
+
+---
+
+## Technology Stack
+
+Frontend
+- Streamlit
+
+Backend
+- FastAPI
+
+AI Models
+- Sentence Transformers
+
+Vector Database
+- FAISS
+
+Programming Language
+- Python
+
+---
+
+## Setup Instructions
+
+### 1 Clone Repository
+### 2 Create Virtual Environment
+### 3 Install Dependencies
+
+
+---
+
+## Running the Project
+Start backend  -  uvicorn backend.api:app --reload
+
+Start frontend  -  streamlit run frontend/app.py
+
+Open browser -  http://localhost:8501
+---
+
+## Example Questions
+
+- What is Retrieval Augmented Generation?
+- Explain vector databases
+- What is semantic search?
+
+---
+
+## Future Improvements
+
+- Chat memory
+- Multi-document knowledge ingestion
+- Scalable vector databases
+- Improved answer citations
+
+---
+
+## Conclusion
+
+This project demonstrates how AI-powered retrieval systems can improve knowledge discovery and enable intelligent question answering using modern machine learning techniques.

ai-rag-project/backend/api.py

@@ -0,0 +1,77 @@
+import logging
+import os
+
+from fastapi import FastAPI, HTTPException
+from backend.retriever import load_knowledge, retrieve
+import openai
+
+logger = logging.getLogger(__name__)
+app = FastAPI()
+
+openai_api_key = os.getenv("OPENAI_API_KEY")
+if openai_api_key:
+    openai.api_key = openai_api_key
+
+
+@app.on_event("startup")
+def startup_event():
+    try:
+        load_knowledge()
+    except Exception:
+        logger.exception("Failed to load knowledge base")
+
+
+@app.get("/")
+def home():
+
+    return {"status": "Endee-style RAG assistant ready"}
+
+
+@app.get("/ask")
+
+def ask(question: str):
+
+    if not openai_api_key:
+        # Return a harmless response so the frontend can render an error message
+        # instead of showing "Backend not running" when no API key is configured.
+        return {
+            "answer": "OPENAI_API_KEY is not set. Set the environment variable to enable OpenAI API calls.",
+            "sources": [],
+        }
+
+    try:
+        docs = retrieve(question)
+    except Exception as e:
+        raise HTTPException(
+            status_code=500,
+            detail=f"Failed to retrieve knowledge: {e}",
+        ) from e
+
+    context = "\n".join(docs)
+
+    prompt = f"""
+    Use the following knowledge to answer the question.
+
+    Knowledge:
+    {context}
+
+    Question:
+    {question}
+
+    Provide a clear answer and reference the knowledge used.
+    """
+
+    try:
+        response = openai.ChatCompletion.create(
+            model="gpt-4o-mini",
+            messages=[{"role": "user", "content": prompt}]
+        )
+    except Exception as e:
+        raise HTTPException(status_code=502, detail=str(e)) from e
+
+    answer = response["choices"][0]["message"]["content"]
+
+    return {
+        "answer": answer,
+        "sources": docs
+    }

ai-rag-project/backend/knowledge_base.py

@@ -0,0 +1,41 @@
+from sentence_transformers import SentenceTransformer
+from pypdf import PdfReader
+import numpy as np
+from backend.vector_store import VectorStore
+
+model = SentenceTransformer("all-MiniLM-L6-v2")
+
+vector_db = None
+
+
+def load_pdf(file_path):
+
+    global vector_db
+
+    reader = PdfReader(file_path)
+
+    text = ""
+
+    for page in reader.pages:
+        text += page.extract_text()
+
+    chunks = text.split("\n")
+
+    chunks = [c.strip() for c in chunks if len(c) > 40]
+
+    embeddings = model.encode(chunks)
+
+    dimension = embeddings.shape[1]
+
+    vector_db = VectorStore(dimension)
+
+    vector_db.add(embeddings, chunks)
+
+
+def retrieve(query):
+
+    query_vec = model.encode([query])
+
+    results = vector_db.search(query_vec)
+
+    return results

ai-rag-project/backend/retriever.py

@@ -0,0 +1,38 @@
+from pathlib import Path
+
+from sentence_transformers import SentenceTransformer
+from backend.vector_store import VectorStore
+
+MODEL_NAME = "all-MiniLM-L6-v2"
+KNOWLEDGE_FILE = Path(__file__).resolve().parents[1] / "data" / "knowledge.txt"
+
+model = SentenceTransformer(MODEL_NAME)
+vector_db = None
+
+
+def load_knowledge():
+
+    global vector_db
+
+    if not KNOWLEDGE_FILE.exists():
+        raise FileNotFoundError(f"Knowledge file not found at: {KNOWLEDGE_FILE}")
+
+    with open(KNOWLEDGE_FILE, encoding="utf-8") as f:
+        docs = [line.strip() for line in f if line.strip()]
+
+    embeddings = model.encode(docs)
+
+    dim = embeddings.shape[1]
+
+    vector_db = VectorStore(dim)
+
+    vector_db.add(embeddings, docs)
+
+
+def retrieve(query):
+
+    query_vec = model.encode([query])
+
+    results = vector_db.search(query_vec)
+
+    return results

ai-rag-project/backend/vector_store.py

@@ -0,0 +1,17 @@
+import faiss
+import numpy as np
+
+
+class VectorStore:
+
+    def __init__(self, dimension):
+        self.index = faiss.IndexFlatL2(dimension)
+        self.documents = []
+
+    def add(self, embeddings, docs):
+        self.index.add(np.array(embeddings))
+        self.documents.extend(docs)
+
+    def search(self, query_embedding, k=3):
+        D, I = self.index.search(np.array(query_embedding), k)
+        return [self.documents[i] for i in I[0]]

ai-rag-project/requirements.txt

@@ -0,0 +1,8 @@
+fastapi
+uvicorn
+streamlit
+sentence-transformers
+faiss-cpu
+numpy
+requests
+openai