app.py

from huggingface_hub import InferenceClient
from resume import data
import markdowm as md
import gradio as gr
import base64
import datetime

client = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1")

# client = InferenceClient("mistralai/Mixtral-8x7B-Instruct-v0.1")


# Chatbot response function with integrated system message
def respond(
        message,
        history: list[tuple[str, str]],
        max_tokens=1024,
        temperature=0.4,
        top_p=0.95,
):
    # System message defining assistant behavior
    system_message = {
        "role": "system",
        "content":  f"Act as Tarun and respond to the user's questions professionally. Tarun is a dedicated BTech final-year student actively seeking a job. Your name is Tarun."
                    f"Here is Tarun’s background:```{data}```. Only answer questions using the information provided here, and strictly use only the links found in this data. If an answer isn’t available within this information, notify the user politely and suggest they reach out via LinkedIn for further assistance."
                    f"Responses should be clear, professional, and strictly in English. Avoid giving random or empty responses at all times."


    }

    messages = [system_message]

    # Adding conversation history
    for val in history:
        if val[0]:
            messages.append({"role": "user", "content": val[0]})
        if val[1]:
            messages.append({"role": "assistant", "content": val[1]})
            
    # print(f"{datetime.datetime.now()}::{{'role': 'user', 'content': val[0]}}->{{'role': 'user', 'content': val[1]}}")
    
    # Adding the current user input
    messages.append({"role": "user", "content": message})
    
    response = ""

    # Streaming the response from the API
    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
        
    print(f"{datetime.datetime.now()}::{messages[-1]['content']}->{response}\n")
    
def encode_image(image_path):
    with open(image_path, "rb") as image_file:
        return base64.b64encode(image_file.read()).decode('utf-8')

# Encode the images
github_logo_encoded = encode_image("Images/github-logo.png")
linkedin_logo_encoded = encode_image("Images/linkedin-logo.png")
website_logo_encoded = encode_image("Images/ai-logo.png")

# Gradio interface with additional sliders for control
with gr.Blocks(theme=gr.themes.Ocean(font=[gr.themes.GoogleFont("Roboto Mono")]), css='footer {visibility: hidden}') as main:
    gr.Markdown(md.title)
    with gr.Tabs():
        with gr.TabItem("My2.0", visible=True, interactive=True):
            gr.ChatInterface(respond,
                             chatbot=gr.Chatbot(height=500),
                             examples=["Tell me about yourself",
                                       'Can you walk me through some of your recent projects and explain the role you played in each?',
                                       "What specific skills do you bring to the table that would benefit our company's AI/ML initiatives?",
                                       "How do you stay updated with the latest trends and advancements in AI and Machine Learning?",
                                      ]
                            )
            gr.Markdown(md.description)

        with gr.TabItem("Resume", visible=True, interactive=True):
            gr.Markdown(data)
    
    gr.HTML(md.footer.format(github_logo_encoded, linkedin_logo_encoded, website_logo_encoded))

if __name__ == "__main__":
    main.launch(share=True)