Add workflow logic for weather forecast demo (#24)

chatbot-ui/Dockerfile

@@ -0,0 +1,24 @@
+# copied from https://github.com/bergos/embedding-server
+
+FROM python:3 AS base
+
+#
+# builder
+#
+FROM base AS builder
+
+WORKDIR /src
+
+COPY requirements.txt /src/
+RUN pip install --prefix=/runtime --force-reinstall -r requirements.txt
+
+COPY . /src
+
+FROM python:3-slim AS output
+
+COPY --from=builder /runtime /usr/local
+
+COPY /app /app
+WORKDIR /app
+
+CMD ["python", "run.py"]

chatbot-ui/app/run.py

@@ -0,0 +1,81 @@
+import gradio as gr
+
+import asyncio
+import httpx
+import async_timeout
+
+from loguru import logger
+from typing import Optional, List
+from pydantic import BaseModel
+from dotenv import load_dotenv
+
+import os
+load_dotenv()
+
+OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
+CHAT_COMPLETION_ENDPOINT = os.getenv("CHAT_COMPLETION_ENDPOINT", "https://api.openai.com/v1/chat/completions")
+
+class Message(BaseModel):
+    role: str
+    content: str
+
+async def make_completion(messages:List[Message], nb_retries:int=3, delay:int=30) -> Optional[str]:
+    """
+    Sends a request to the ChatGPT API to retrieve a response based on a list of previous messages.
+    """
+    header = {
+        "Content-Type": "application/json",
+        "Authorization": f"Bearer {OPENAI_API_KEY}"
+    }
+    try:
+        async with async_timeout.timeout(delay=delay):
+            async with httpx.AsyncClient(headers=header) as aio_client:
+                counter = 0
+                keep_loop = True
+                while keep_loop:
+                    logger.debug(f"Chat/Completions Nb Retries : {counter}")
+                    try:
+                        resp = await aio_client.post(
+                            url = CHAT_COMPLETION_ENDPOINT,
+                            json = {
+                                "model": "gpt-3.5-turbo",
+                                "messages": messages
+                            }
+                        )
+                        logger.debug(f"Status Code : {resp.status_code}")
+                        if resp.status_code == 200:
+                            return resp.json()["choices"][0]["message"]["content"]
+                        else:
+                            logger.warning(resp.content)
+                            keep_loop = False
+                    except Exception as e:
+                        logger.error(e)
+                        counter = counter + 1
+                        keep_loop = counter < nb_retries
+    except asyncio.TimeoutError as e:
+        logger.error(f"Timeout {delay} seconds !")
+    return None
+
+async def predict(input, history):
+    """
+    Predict the response of the chatbot and complete a running list of chat history.
+    """
+    history.append({"role": "user", "content": input})
+    print(history)
+    response = await make_completion(history)
+    history.append({"role": "assistant", "content": response})
+    messages = [(history[i]["content"], history[i+1]["content"]) for i in range(0, len(history)-1, 2)]
+    return messages, history
+
+"""
+Gradio Blocks low-level API that allows to create custom web applications (here our chat app)
+"""
+with gr.Blocks() as demo:
+    logger.info("Starting Demo...")
+    chatbot = gr.Chatbot(label="WebGPT")
+    state = gr.State([])
+    with gr.Row():
+        txt = gr.Textbox(show_label=False, placeholder="Enter text and press enter")
+    txt.submit(predict, [txt, state], [chatbot, state])
+
+demo.launch(server_name="0.0.0.0", server_port=8080)

chatbot-ui/requirements.txt

@@ -0,0 +1,6 @@
+gradio==4.39.0
+async_timeout==4.0.3
+loguru==0.7.2
+asyncio==3.4.3
+httpx==0.27.0
+python-dotenv==1.0.1