excel_chat.py

import gradio as gr
from mistralai.client import MistralClient
from mistralai.models.chat_completion import ChatMessage
import os
import pandas as pd
import numpy as np
from groq import Groq
import anthropic
from users_management import update_json, users
from code_df_custom import load_excel
import zipfile
from openai import *
import time

#users = ['maksG', 'AlmaA', 'YchK']

def ask_llm(query, user_input, client_index, user, keys):
    messages = [
        {
            "role": "system",
            "content": f"You are a helpful assistant. Only show your final response to the **User Query**! Do not provide any explanations or details: \n# User Query:\n{query}."
        },
        {
            "role": "user",
            "content": user_input,
        }
    ]
    
    systemC = messages[0]["content"]
    messageC = [{
        "role": "user",
        "content": [{
            "type": "text",
            "text": user_input
        }]
    }]

    try: 
        
        if "Mistral" in client_index:
            client = MistralClient(api_key=os.environ[user['api_keys']['mistral']])
            model_map = {
                "Mistral Tiny": "mistral-tiny",
                "Mistral Small": "mistral-small-latest",
                "Mistral Medium": "mistral-medium",
            }
            chat_completion = client.chat(messages=messages, model=model_map[client_index])
    
        elif "Claude" in client_index:
            client = anthropic.Anthropic(api_key=os.environ[user['api_keys']['claude']])
            model_map = {
                "Claude Sonnet": "claude-3-sonnet-20240229",
                "Claude Opus": "claude-3-opus-20240229",
            }
            response = client.messages.create(
                model=model_map[client_index],
                max_tokens=350,
                temperature=0,
                system=systemC,
                messages=messageC
            )
            return response.content[0].text
    
        elif "GPT 4o" in client_index:
            client = OpenAI(api_key=os.environ["OPENAI_YCHK"])
            response = client.chat.completions.create(
                model="gpt-4o",
                messages=messageC
            )
            return response.choices[0][message][content].text
    
        elif "Perplexity" in client_index:
            client = OpenAI(api_key=os.environ["PERPLEXITY_ALMAA"], base_url="https://api.perplexity.ai")
            model_map = {
                    "Perplexity Llama3 70b": "llama-3-70b-instruct",
                    "Perplexity Llama3 8b": "llama-3-8b-instruct",
                    "Perplexity Llama3 Sonar Small": "llama-3-sonar-small-32k-chat",
                    "Perplexity Llama3 Sonar Large": "llama-3-sonar-large-32k-chat"
            }
            
            response = client.chat.completions.create(
                model=model_map[client_index],
                messages=messageC
            )
    
            responseContent = str(response.choices[0].message.content)
            print(responseContent)        
            return responseContent,keys
            
        elif "Groq" in client_index:
            try:
                client = Groq(api_key= os.getenv(keys[0]))
                model_map = {
                    "Groq Mixtral": "mixtral-8x7b-32768",
                    "Groq Llama3 70b": "llama3-70b-8192",
                    "Groq Llama3 8b": "llama3-8b-8192"
                }
                chat_completion = client.chat.completions.create(
                    messages=messages,
                    model=model_map[client_index],
                )
                response = chat_completion.choices[0].message.content
            except Exception as e:
                print("Change key")
                if keys[0] == keys[1][0]:
                    keys[0] = keys[1][1]
                elif keys[0] == keys[1][1]:
                    keys[0] = keys[1][2]
                else:
                    keys[0] = keys[1][0]
                    
                client = Groq(api_key= os.getenv(keys[0]))
                chat_completion = client.chat.completions.create(
                    messages=messages,
                    model='llama3-8b-8192',
                )
                response = chat_completion.choices[0].message.content
        else:
            raise ValueError("Unsupported client index provided")

    
        # Return the response, handling the structure specific to Groq and Mistral clients.
        return chat_completion.choices[0].message.content,keys if client_index != "Claude" else chat_completion
        
    except (BadRequestError) as e:
        
        model_id = "meta-llama/Meta-Llama-3-70B-Instruct"
        access_token = os.getenv("HUGGINGFACE_SPLITFILES_API_KEY")

        tokenizer = AutoTokenizer.from_pretrained(
            model_id,
            padding_side="left",
            token = access_token
        )

        user_input_tokenized = tokenizer.encode(user_input)
        messages = []

        while len(user_input_tokenized) > max_token:

            user_input_divided = tokenizer.decode(user_input_tokenized[:max_token])
            messages.append([
            {
                "role": "system",
                "content": f"You are a helpful assistant. Only show your final response to the **User Query**! Do not provide any explanations or details: \n# User Query:\n{query}."
            },
            {
                "role": "user",
                "content": user_input_divided,
            }])

            user_input_tokenized = user_input_tokenized[max_token:]

        responses = []

        print(len(messages))
        for msg in messages:

            responses.append(client.chat.completions.create(
              model=model_map["Perplexity Llama3 70b"],
              messages=msg
            ))

        response = ""
        for resp in responses:
            response += " " + resp.choices[0].message.content

        return response

    except (RateLimitError) as e:

        #if model_user in keys:
            #Swap those keys
        #    return f()

        #else:
            #get eepy
        time.sleep(60)
        return ask_llm(query, user_input, client_index, user, keys)

    except Exception as e:
        print(e)
        return "unhandled error",keys if client_index != "Claude" else chat_completion





def filter_df(df, column_name, keywords):
    if len(keywords)>0:
        if column_name in df.columns:
            contains_keyword = lambda x: any(keyword.lower() in (x.lower() if type(x)==str else '') for keyword in keywords)
            filtered_df = df[df[column_name].apply(contains_keyword)]
        else:
            contains_keyword = lambda row: any(keyword.lower() in (str(cell).lower() if isinstance(cell, str) else '') for keyword in keywords for cell in row)
            filtered_df = df[df.apply(contains_keyword, axis=1)]
    else:
        filtered_df = df
    return filtered_df

def chat_with_mistral(source_cols, dest_col, prompt, excel_file, url, search_col, keywords, client, user):
    # API Keys for Groq :
    KEYS = ['GROQ_API_KEY1', 'GROQ_API_KEY2', 'GROQ_API_KEY3']
    GroqKey = KEYS[0]
    gloabal_keys = [GroqKey, KEYS]
    
    new_prompts, new_keywords, new_user, conf_file_path = update_json(user, prompt, keywords)
    print(f'xlsxfile = {excel_file}')
    df = pd.read_excel(excel_file)
    df[dest_col] = ""
    if excel_file:
        file_name = excel_file.split('.xlsx')[0] + "_with_" + dest_col.replace(' ', '_') + ".xlsx"
    elif url.endswith('Docs/', 'Docs'):
        file_name = url.split("/Docs")[0].split("/")[-1] + ".xlsx"
    else:
        file_name = "meeting_recap_grid.xlsx"

    print(f"Keywords: {keywords}")

    filtred_df = filter_df(df, search_col, keywords)

    cpt = 1
    for index, row in filtred_df.iterrows():
        concatenated_content = "\n\n".join(f"{column_name}: {str(row[column_name])}" for column_name in source_cols)
        if not concatenated_content == "\n\n".join(f"{column_name}: nan" for column_name in source_cols):

            try:
                llm_answer,gloabal_keys = ask_llm(prompt[0], concatenated_content, client, user, gloabal_keys)
            except Exception:
                llm_answer = "unhandled global error"
            
            print(f"{cpt}/{len(filtred_df)}\nQUERY:\n{prompt[0]}\nCONTENT:\n{concatenated_content[:200]}...\n\nANSWER:\n{llm_answer}")
            df.at[index, dest_col] = llm_answer
            cpt += 1
            # progress((index+1)/len(df),desc=f'Request {index+1}/{len(df)}')

    df.to_excel(file_name, index=False)

    zip_file_path = 'config_file.zip'
    
    with zipfile.ZipFile(zip_file_path, 'w') as zipf:
        zipf.write(conf_file_path, os.path.basename(conf_file_path))
        
    return file_name, df.head(5), new_prompts, new_keywords, new_user, zip_file_path


def get_columns(file,progress=gr.Progress()):
    if file is not None:
        #df = pd.read_excel(file)
        filename, df = load_excel(file)
        columns = list(df.columns)
        return gr.update(choices=columns), gr.update(choices=columns), gr.update(choices=columns), gr.update(choices=columns + [""]), gr.update(choices=columns + ['[ALL]']), df.head(5), filename, df
    else:
        return gr.update(choices=[]), gr.update(choices=[]), gr.update(choices=[]), gr.update(choices=[]), gr.update(choices=[]), pd.DataFrame(), '', pd.DataFrame()