update code structure for huggingface deployment

Dockerfile

@@ -1,13 +1,42 @@
-FROM tiangolo/uvicorn-gunicorn-fastapi:python3.9
+# Use the custom FastAPI image
+# FROM tiangolo/uvicorn-gunicorn-fastapi:python3.9
+# Use ubuntu as base image
+FROM ubuntu:20.04
+
+# Avoid prompts from apt
+ENV DEBIAN_FRONTEND=noninteractive
+
+# Update apt repositories and install Python and Pip
+RUN apt-get update && \
+    apt-get install -y python3-pip python3-dev
+
+# Check if the symbolic links already exist and create them if they don't
+RUN if [ ! -e /usr/bin/python ]; then ln -s /usr/bin/python3 /usr/bin/python; fi && \
+    if [ ! -e /usr/bin/pip ]; then ln -s /usr/bin/pip3 /usr/bin/pip; fi
 
 # Set default values for environment variables
 ENV OPENAI_ORG_ID=default_org_id
 ENV OPENAI_API_KEY=default_api_key
 
-COPY ./requirements.txt /code/requirements.txt
+# Set environment variables for Matplotlib and Fontconfig
+ENV MPLCONFIGDIR=/app/matplotlib_cache
+ENV FONTCONFIG_PATH=/app/fontconfig
 
+# Create the directories for Matplotlib cache and Fontconfig
+RUN mkdir -p /app/matplotlib_cache /app/fontconfig && \
+    chmod -R 777 /app/matplotlib_cache /app/fontconfig
+
+# Create a writable directory for Fontconfig cache
+RUN mkdir -p /app/fontconfig_cache && chmod -R 777 /app/fontconfig_cache
+
+# Set the environment variable so Fontconfig uses the writable directory
+ENV FONTCONFIG_PATH=/app/fontconfig_cache
+
+# Copy the requirements file and install dependencies
+COPY ./requirements.txt /code/requirements.txt
 RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
 
+# Copy your application source code and script
 COPY ./src /app
 
-CMD ["uvicorn", "index.main:app", "--host", "0.0.0.0", "--port", "80"]
+CMD ["uvicorn", "app.app.main:app", "--host", "0.0.0.0", "--port", "7860"]

README.md

@@ -1,3 +1,13 @@
+---
+title: Building your generative AI Apps in Python
+emoji: 🤖
+colorFrom: "green"
+colorTo: "green"
+sdk: docker
+sdk_version: 3.8.5
+app_file: app.py
+pinned: false
+---
 # 生成式AI後端開發部署實戰
 # Building your generative AI API in Python 
 

scripts/app.py

@@ -0,0 +1,44 @@
+import gradio as gr
+import requests
+
+is_gpt3 = True
+
+
+def echo_gptchatbot(message, history):
+    # Post the message to the server
+    response = requests.post(
+        "https://markchenx-udemy-genai-app-course.hf.space/api/chat/gpt3", data={"user_request": message}
+    )
+    # Return the response
+    llm_output = response.json()["result"]["content"]
+
+    return llm_output
+
+
+def echo_llamachatbot(message, history):
+    # Post the message to the server
+    response = requests.post(
+        "https://markchenx-udemy-genai-app-course.hf.space/api/chat/llama", data={"user_request": message}
+    )
+
+    # Return the response
+    llm_output = response.json()["result"][0]["generated_text"]
+
+    return llm_output
+
+
+# Create a Gradio interface with the chatbot
+if is_gpt3:
+    demo = gr.ChatInterface(
+        fn=echo_gptchatbot,
+        examples=["What is OpenAI?", "What is GPT-3?"],
+        title="GPT-3 Chatbot",
+    )
+else:
+    demo = gr.ChatInterface(
+        fn=echo_llamachatbot,
+        examples=["What is OpenAI?", "What is LLM?"],
+        title="LLM Chatbot - Llama 2 7B",
+    )
+
+demo.launch()

scripts/app_start.sh

@@ -1,3 +1,3 @@
-# Run two commands in the background at the same time
-cd ../src
-python3 -m uvicorn index.main:app --reload & python3 app/main.py 
+#!/bin/sh
+# Run the python module while starting the python app with ./app/main.py
+python3 -m uvicorn index.main:app --host 0.0.0.0 --port 8000 #& python3 src_app/gradio_main.py

src/app/main.py

@@ -1,44 +1,56 @@
-import gradio as gr
-import requests
+from fastapi import FastAPI, Form
+from fastapi.middleware.cors import CORSMiddleware
+from fastapi.staticfiles import StaticFiles
+from fastapi.responses import HTMLResponse
+from .ai.chatbot import gpt_chatbot, llama_chatbot
 
-is_gpt3 = False
+isProduction = False
 
+origins = ["*"]
 
-def echo_gptchatbot(message, history):
-    # Post the message to the server
-    response = requests.post(
-        "http://127.0.0.1:8000/api/chat/gpt3", data={"user_request": message}
+if isProduction:
+    app = FastAPI(
+        title="LLM API Endpoints",
+        docs_url=None,  # Disable docs (Swagger UI)
+        redoc_url=None,  # Disable redoc
     )
-    # Return the response
-    llm_output = response.json()["result"]["content"]
+    #app.mount("/static", StaticFiles(directory="static"), name="static")
+else:
+    app = FastAPI(title="LLM API Endpoints")
+    #app.mount("/static", StaticFiles(directory="static"), name="static")
 
-    return llm_output
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=origins,
+    allow_credentials=True,
+    allow_methods=["POST", "GET", "PUT", "DELETE"],
+    allow_headers=["*"],
+)
 
 
-def echo_llamachatbot(message, history):
-    # Post the message to the server
-    response = requests.post(
-        "http://127.0.0.1:8000/api/chat/llama", data={"user_request": message}
-    )
+# Create a homepage route
+@app.get("/")
+async def index():
+    return {"server ok": True}
 
-    # Return the response
-    llm_output = response.json()["result"][0]["generated_text"]
 
-    return llm_output
+@app.post("/api/chat/gpt3", tags=["OpenAI GPT-3"])
+async def gpt_chat(user_request: str = Form(...)):
+    """
+    Chat with LLM Backend - GPT-3
+    """
+    # Get the text content in the user request
+    result = gpt_chatbot(user_request=user_request)
 
+    return {"result": result}
 
-# Create a Gradio interface with the chatbot
-if is_gpt3:
-    demo = gr.ChatInterface(
-        fn=echo_gptchatbot,
-        examples=["What is OpenAI?", "What is GPT-3?"],
-        title="GPT-3 Chatbot",
-    )
-else:
-    demo = gr.ChatInterface(
-        fn=echo_llamachatbot,
-        examples=["What is OpenAI?", "What is LLM?"],
-        title="LLM Chatbot - Llama 2 7B",
-    )
 
-demo.launch()
+@app.post("/api/chat/llama", tags=["Llama 2 7B Chat"])
+async def llama_chat(user_request: str = Form(...)):
+    """
+    Chat with LLM Backend - Llama 2 7b Chat
+    """
+    # Get the text content in the user request
+    result = llama_chatbot(user_request=user_request)
+
+    return {"result": result}

src/test_main.py

@@ -1,5 +1,5 @@
 from fastapi.testclient import TestClient
-from index.main import app
+from app.main import app
 
 client = TestClient(app)