lv_recipe_chatbot/app.py

# AUTOGENERATED! DO NOT EDIT! File to edit: ../nbs/01_app.ipynb.

# %% auto 0
__all__ = ['ConversationBot', 'create_demo']

# %% ../nbs/01_app.ipynb 3
import copy
import os

import gradio as gr
from langchain import LLMChain, OpenAI, PromptTemplate
from langchain.agents import (
    AgentExecutor,
    AgentType,
    OpenAIFunctionsAgent,
    Tool,
    initialize_agent,
    load_tools,
)
from langchain.chains import ConversationChain
from langchain.chat_models import ChatOpenAI
from langchain.memory import ChatMessageHistory, ConversationBufferMemory
from langchain.prompts.chat import (
    ChatPromptTemplate,
    HumanMessagePromptTemplate,
    MessagesPlaceholder,
)
from PIL import Image

import constants
from .engineer_prompt import INIT_PROMPT
from lv_recipe_chatbot.ingredient_vision import (
    SAMPLE_IMG_DIR,
    BlipImageCaptioning,
    VeganIngredientFinder,
    format_image,
)
from .vegan_recipe_tools import vegan_recipe_edamam_search

# %% ../nbs/01_app.ipynb 16
class ConversationBot:
    memory_key: str = "chat_history"

    def __init__(
        self,
        vegan_ingred_finder: VeganIngredientFinder,
        img_cap: BlipImageCaptioning,
        verbose: bool = True,
    ):
        self.llm = ChatOpenAI(temperature=0.1, verbose=verbose)
        self.init_prompt = copy.deepcopy(INIT_PROMPT)
        self.img_cap = img_cap
        self.vegan_ingred_finder = vegan_ingred_finder
        self.verbose = verbose
        init_prompt_msgs = self.init_prompt.messages
        self.ai_prompt_questions = {
            "ingredients": init_prompt_msgs[1],
            "allergies": init_prompt_msgs[3],
            "recipe_open_params": init_prompt_msgs[5],
        }

    def respond(self, user_msg, chat_history):
        response = self._get_bot_response(user_msg, chat_history)
        chat_history.append((user_msg, response))
        return "", chat_history

    def init_agent_executor(self, chat_msgs):
        tools = [vegan_recipe_edamam_search]
        prompt = OpenAIFunctionsAgent.create_prompt(
            system_message=self.init_prompt.messages[0],
            extra_prompt_messages=chat_msgs
            + [MessagesPlaceholder(variable_name=self.memory_key)],
        )
        self.memory = ConversationBufferMemory(
            chat_memory=ChatMessageHistory(messages=chat_msgs),
            return_messages=True,
            memory_key=self.memory_key,
        )
        self.agent_executor = AgentExecutor(
            agent=OpenAIFunctionsAgent(llm=self.llm, tools=tools, prompt=prompt),
            tools=tools,
            memory=self.memory,
            verbose=True,
        )

    def reset(self):
        self.memory.clear()
        self.init_prompt = copy.deepcopy(INIT_PROMPT)

    def run_img(self, image: str):
        desc = self.img_cap.inference(format_image(image))
        answer = self.vegan_ingred_finder.list_ingredients(image)
        msg = f"""I uploaded an image that may contain vegan ingredients.
The description of the image is: `{desc}`.
The extracted ingredients are:
```
{answer}
```"""
        base_prompt = INIT_PROMPT.messages[2].prompt.template
        new_prompt = f"{msg}I may type some more ingredients below.\n{base_prompt}"
        self.init_prompt.messages[2].prompt.template = new_prompt
        return msg

    def _get_bot_response(self, user_msg: str, chat_history) -> str:
        if len(chat_history) < 2:
            return self.ai_prompt_questions["allergies"].prompt.template

        if len(chat_history) < 3:
            return self.ai_prompt_questions["recipe_open_params"].prompt.template

        if len(chat_history) < 4:
            user = 0
            ai = 1
            user_msgs = [msg_pair[user] for msg_pair in chat_history[1:]]
            f_init_prompt = self.init_prompt.format_prompt(
                ingredients=user_msgs[0],
                allergies=user_msgs[1],
                recipe_freeform_input=user_msg,
            )
            chat_msgs = f_init_prompt.to_messages()
            results = self.llm.generate([chat_msgs])
            chat_msgs.append(results.generations[0][0].message)
            # prepare the agent to takeover from this point
            self.init_agent_executor(chat_msgs)
            return self.agent_executor.run("Search for a vegan recipe with that query")

        response = self.agent_executor.run(input=user_msg)
        return response

# %% ../nbs/01_app.ipynb 20
def create_demo(bot: ConversationBot):
    sample_images = []
    all_imgs = [f"{SAMPLE_IMG_DIR}/{img}" for img in os.listdir(SAMPLE_IMG_DIR)]
    for i, img in enumerate(all_imgs):
        if i in [
            1,
            2,
            3,
        ]:
            sample_images.append(img)
    with gr.Blocks() as demo:
        gr_img = gr.Image(type="filepath")
        btn = gr.Button(value="Submit image")
        ingredients_msg = gr.Text(label="Ingredients from image")
        btn.click(bot.run_img, inputs=[gr_img], outputs=[ingredients_msg])
        gr.Examples(
            examples=sample_images,
            inputs=gr_img,
        )

        chatbot = gr.Chatbot(
            value=[(None, bot.ai_prompt_questions["ingredients"].prompt.template)]
        )

        msg = gr.Textbox()
        # clear = gr.Button("Clear")
        gr.Markdown(
            """
        **🔃Refresh the page to start from scratch🔃**  
        
        Recipe search tool powered by the [Edamam API](https://www.edamam.com/)  
        
        ![Edamam Logo](https://www.edamam.com/assets/img/small-logo.png)
        """
        )
        msg.submit(
            fn=bot.respond, inputs=[msg, chatbot], outputs=[msg, chatbot], queue=False
        )
        # clear.click(lambda: None, None, chatbot, queue=False).then(bot.reset)
        return demo