Upload 3 files

app.py

@@ -0,0 +1,24 @@
+import gradio as gr
+from indexer import index_text, answer_query
+
+# Gradio interface function to handle the RAG system
+def rag_system(query):
+    # Index the input text
+    vectorstore = index_text()
+    
+    # Answer the query based on the indexed text
+    answer = answer_query(query, vectorstore)
+    
+    return answer
+
+# Build the Gradio interface
+iface = gr.Interface(
+    fn=rag_system, 
+    inputs=["text"], 
+    outputs="text", 
+    title="AI Alignment Bot",
+    description="If the model can't answer, it will say sorry I don't know"
+)
+
+# Launch the app
+iface.launch()

indexer.py

@@ -0,0 +1,53 @@
+from langchain_community.vectorstores import FAISS
+from langchain_core.documents import Document
+from langchain_nvidia_ai_endpoints import NVIDIAEmbeddings
+import os 
+from google import genai
+from google.genai import types
+
+# Set up the Gemini API key
+import os
+
+def index_text():
+
+    os.environ["NVIDIA_API_KEY"] ="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"
+    
+    nvidia_embeddings = NVIDIAEmbeddings(
+        model="nvidia/llama-3.2-nv-embedqa-1b-v2",
+        truncate="NONE"
+    )
+    vectorstore = FAISS.load_local("nvidia_faiss_index", embeddings=nvidia_embeddings,allow_dangerous_deserialization=True)
+    return vectorstore
+
+
+def answer_query(query, vectorstore):
+    RAG_TEMPLATE = """
+#CONTEXT:
+{context}
+
+QUERY:
+{query}
+
+Use the provided context to answer the user query. Only use the provided context to answer the query.
+If you do not know the answer, or it's not contained in the provided context, respond with "I don't know".
+"""
+    os.environ["GEMINI_API_KEY"] = "AIzaSyCP3iHlrhG-aDKaZlNKzYE3yXA-7pLGCxM"
+    client = genai.Client()
+    # Get relevant documents
+    retriever = vectorstore.as_retriever()
+    search_results = retriever.invoke(query, k=2)
+    
+    # Combine context from retrieved documents
+    context = " ".join([doc.page_content for doc in search_results])
+    
+    # Build prompt
+    prompt = RAG_TEMPLATE.format(context=context, query=query)
+
+    # Generate response using Gemini
+    response = client.models.generate_content(
+        model="gemini-2.5-pro",
+        contents=prompt,
+        config=types.GenerateContentConfig(),
+    )
+
+    return response.text

requirements.txt

@@ -0,0 +1,14 @@
+gradio
+langchain
+langchain_community
+langchain-core
+transformers
+faiss-cpu
+sentence-transformers
+huggingface_hub
+pydantic
+google-genai
+langchain_nvidia_ai_endpoints
+langchain_community 
+langchain_core
+faiss-cpu