Update chatbot using api

README.md

@@ -59,6 +59,7 @@ FFood is a food website that lets customers order food in a quick and convenient
 - View Food by Categories.
 - View Food Details.
 - Rate Food.
+- Chat bot
 - Search Food by keyword.
 - Search Food by image.
 - User cart Management.

api/chatbot/Dockerfile

@@ -0,0 +1,13 @@
+# Use an official Python runtime as a parent image
+FROM tiangolo/uvicorn-gunicorn-fastapi:python3.9
+
+COPY requirements.txt requirements.txt
+RUN pip3 install -r requirements.txt
+RUN export PROTOCOL_BUFFERS_PYTHON_IMPLEMENTATION=python
+
+COPY . .
+
+EXPOSE 8100
+
+# Run app.py when the container launches
+CMD ["python3", "app.py"]

api/chatbot/app.py

@@ -0,0 +1,35 @@
+import minsearch
+import json
+from fastapi import FastAPI, HTTPException
+from fastapi.middleware.cors import CORSMiddleware
+import uvicorn
+import asyncio
+import nest_asyncio
+from chatbot_rag import rag, read_json
+
+nest_asyncio.apply()
+
+app = FastAPI()
+file = 'food_description.json'
+index = read_json(file)
+print()
+
+# Add CORS Middleware
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=["*"],
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+@app.get("/rag/")
+async def rag_endpoint(query: str):
+    try:
+        answer = rag(query, index)
+        return {"answer": answer}
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=str(e))
+
+if __name__ == '__main__':
+    uvicorn.run('app:app', port=8100, host='0.0.0.0', loop='asyncio')

api/chatbot/chatbot_rag.py

@@ -0,0 +1,92 @@
+import minsearch
+import json
+from g4f.client import Client
+import g4f
+
+_providers = [
+    g4f.Provider.Aichat,
+    g4f.Provider.ChatBase,
+    g4f.Provider.Bing,
+    g4f.Provider.GptGo,
+    g4f.Provider.You,
+    g4f.Provider.Yqcloud,
+]
+
+def llm(prompt):
+    client = Client()
+    print("Creating chat completion...")
+    chat_completion = client.chat.completions.create(
+        model="gpt-3.5-turbo",
+        messages=[{"role": "user", "content": prompt}], 
+        ignored=["Ylokh", "GptGo", "AItianhu", "Aibn", "Myshell", "FreeGpt"],
+        stream=True
+    )
+
+    response = ""
+    print("Waiting for completion...")
+    for completion in chat_completion:
+        response += completion.choices[0].delta.content or ""
+
+    return {"response": response}
+
+def read_json(file):
+    with open(file, 'rt', encoding='utf-8') as f_in:  # Specify UTF-8 encoding
+        docs_raw = json.load(f_in)
+    print(f"Read {len(docs_raw)} documents from {file}.")
+
+    documents = []
+    for course_dict in docs_raw:
+        for doc in course_dict['items']:
+            doc['category'] = course_dict['category']
+            documents.append(doc)
+
+    index = minsearch.Index(
+        text_fields=["question", "description", "name"],
+        keyword_fields=["category"]
+    )
+    index.fit(documents)
+    print("Index has been initialized.")
+    return index
+
+def search(query, index):
+
+    if index is None:
+        raise RuntimeError("Index has not been initialized. Call read_json() first.")
+
+    boost = {'question': 3.0, 'section': 0.5}
+
+    results = index.search(
+        query=query,
+        filter_dict={'category': 'FFood'},
+        boost_dict=boost,
+        num_results=5
+    )
+
+    return results
+
+def build_prompt(query, search_results):
+    prompt_template = """
+You're a vietnamese food ordering assistant. Answer the QUESTION based on the CONTEXT from the food description database.
+Use only the facts from the CONTEXT when answering the QUESTION.
+
+QUESTION: {question}
+
+CONTEXT: 
+{context}
+""".strip()
+
+    context = ""
+
+    for doc in search_results:
+        context = context + f"name: {doc['name']}\nquestion: {doc['question']}\nanswer: {doc['description']}\n\n"
+
+    prompt = prompt_template.format(question=query, context=context).strip()
+    return prompt
+
+def rag(query, index):
+    print(f"Searching for: {query}")
+    search_results = search(query, index)
+    print(f"Found {len(search_results)} search results.")
+    prompt = build_prompt(query, search_results)
+    answer = llm(prompt)
+    return answer