Initial code, download index from Hugging Face Hub

README.md

@@ -1,12 +1,81 @@
+                `# Dify Documentation Expert (RAG Chatbot)
+
+A Retrieval-Augmented Generation (RAG) chatbot for Dify documentation, using MiniLM for semantic search and DistilBERT for answer generation. Built with Python, Flask, LangChain, and FAISS. Deployable on Vercel.
+
+---
+
+## Features
+- Scrapes and processes Dify documentation
+- Builds a FAISS vector store with MiniLM embeddings
+- Uses DistilBERT for question answering
+- Fast API with `/ask` and `/health` endpoints
+- Ready for serverless deployment (Vercel)
+
+---
+
+## Project Structure
+```
+Dify-Expert/
+├── api/
+│   └── dify_api.py         # Flask API (entry point for Vercel)
+├── dify_docs_scraper/
+│   ├── dify_rag.py         # RAG logic (vector store + QA)
+│   ├── dify_faiss_index/   # FAISS index files
+│   └── ...                 # Scraper, processing scripts, etc.
+├── requirements.txt        # Python dependencies
+├── vercel.json             # Vercel config
+├── README.md               # This file
+└── futurescope.md          # Project roadmap/ideas
+```
+
 ---
-title: Dify Expert
-emoji: 💬
-colorFrom: yellow
-colorTo: purple
-sdk: gradio
-sdk_version: 5.0.1
-app_file: app.py
-pinned: false
+
+## Quickstart
+
+### 1. Install dependencies
+```bash
+pip install -r requirements.txt
+```
+
+### 2. Build the vector store (if not already built)
+```bash
+python3 dify_docs_scraper/build_vector_store.py
+```
+
+### 3. Run locally (API)
+```bash
+python3 api/dify_api.py
+```
+
+### 4. Query the API
+```bash
+curl -X POST http://localhost:8000/ask \
+  -H "Content-Type: application/json" \
+  -d '{"question": "What is Dify?"}'
+```
+
+---
+
+## Deployment (Vercel)
+1. Push to GitHub.
+2. Connect repo to Vercel.
+3. Deploy (Vercel auto-detects Python from requirements.txt).
+4. Endpoints:
+   - `/ask` (POST)
+   - `/health` (GET)
+
+---
+
+## API Endpoints
+- `POST /ask` — `{ "question": "..." }` → `{ "answer": ..., "score": ..., "sources": [...] }`
+- `GET /health` — `{ "status": "ok" }`
+
+---
+
+## Roadmap
+See [`futurescope.md`](futurescope.md) for planned features and ideas.
+
 ---
 
-An example chatbot using [Gradio](https://gradio.app), [`huggingface_hub`](https://huggingface.co/docs/huggingface_hub/v0.22.2/en/index), and the [Hugging Face Inference API](https://huggingface.co/docs/api-inference/index).
+## License
+MIT

app.py

@@ -1,64 +1,20 @@
 import gradio as gr
-from huggingface_hub import InferenceClient
-
-"""
-For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-"""
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
-
-
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
-
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
-
-    messages.append({"role": "user", "content": message})
-
-    response = ""
-
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
-        stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
-
-        response += token
-        yield response
-
-
-"""
-For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
-"""
-demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
-        ),
-    ],
+from dify_rag import answer_question
+
+def chat_fn(message, history):
+    result = answer_question(message)
+    answer = result["answer"]
+    sources = result.get("sources", [])
+    if sources:
+        answer += "\n\nSources:\n" + "\n".join(f"- [{s.get('title', '')}]({s.get('url', '')})" for s in sources if s.get("url"))
+    return answer
+
+iface = gr.ChatInterface(
+    fn=chat_fn,
+    title="Dify Documentation Expert",
+    description="Ask anything about Dify documentation!",
+    examples=["What is Dify?", "How do I use code-based extensions?", "What is LLMOps?"]
 )
 
-
 if __name__ == "__main__":
-    demo.launch()
+    iface.launch()

dify_rag.py

@@ -0,0 +1,45 @@
+from huggingface_hub import hf_hub_download
+import os
+
+faiss_dir = os.path.join(os.path.dirname(__file__), "dify_faiss_index")
+os.makedirs(faiss_dir, exist_ok=True)
+
+faiss_path = os.path.join(faiss_dir, "index.faiss")
+pkl_path = os.path.join(faiss_dir, "index.pkl")
+
+if not os.path.exists(faiss_path):
+    hf_hub_download(repo_id="k01010/k01010_dify-faiss-index", filename="index.faiss", local_dir=faiss_dir)
+if not os.path.exists(pkl_path):
+    hf_hub_download(repo_id="k01010/k01010_dify-faiss-index", filename="index.pkl", local_dir=faiss_dir)
+
+from langchain_community.embeddings import HuggingFaceEmbeddings
+from langchain_community.vectorstores import FAISS
+from transformers import pipeline
+
+# Load FAISS vector store and QA pipeline ONCE at module level
+embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/multi-qa-MiniLM-L6-cos-v1")
+index_path = os.path.join(os.path.dirname(__file__), "dify_faiss_index")
+vector_store = FAISS.load_local(index_path, embeddings, allow_dangerous_deserialization=True)
+qa = pipeline("question-answering", model="distilbert-base-uncased-distilled-squad")
+
+# Main RAG answer function
+def answer_question(question, top_k=4):
+    retriever = vector_store.as_retriever(search_kwargs={"k": top_k})
+    docs = retriever.get_relevant_documents(question)
+    context = " ".join([doc.page_content for doc in docs])
+    result = qa(question=question, context=context)
+    return {
+        "answer": result["answer"],
+        "score": result["score"],
+        "context": context,
+        "sources": [getattr(doc, "metadata", {}) for doc in docs]
+    }
+
+if __name__ == "__main__":
+    # Simple CLI for testing
+    while True:
+        q = input("Ask a question (or 'exit'): ")
+        if q.lower() == "exit":
+            break
+        out = answer_question(q)
+        print(f"Answer: {out['answer']}\nScore: {out['score']:.2f}\nSources: {out['sources']}")

requirements.txt

@@ -1 +1,7 @@
-huggingface_hub==0.25.2
+flask
+gradio
+langchain-community
+transformers
+faiss-cpu
+torch
+huggingface_hub