app.py

import io
import tempfile
from starlette import status
from functions import *
import pandas as pd
from fastapi import FastAPI, File, UploadFile, HTTPException
from pydantic import BaseModel
from fastapi.middleware.cors import CORSMiddleware
from src.api.speech_api import speech_translator_router
from functions import client as supabase
from urllib.parse import urlparse
import nltk
import time
import uuid


nltk.download('punkt_tab')

app = FastAPI(title="ConversAI", root_path="/api/v1")

app.add_middleware(
    CORSMiddleware,
    allow_origins=["*"],
    allow_credentials=True,
    allow_methods=["*"],
    allow_headers=["*"],
)

app.include_router(speech_translator_router, prefix="/speech")


@app.post("/signup")
async def sign_up(email, username, password):
    res, _ = supabase.auth.sign_up(
        {"email": email, "password": password, "role": "user"}
    )
    user_id = res[1].id
    r_ = createUser(user_id=user_id, username=username)
    print(r_)

    response = {
        "status": "success",
        "code": 200,
        "message": "Please check you email address for email verification",
    }

    return response


@app.post("/session-check")
async def check_session():
    res = supabase.auth.get_session()

    return res

@app.post("/get-user")
async def get_user(access_token):
    res = supabase.auth.get_user(jwt=access_token)
    return res


@app.post("/referesh-token")
async def refresh_token(refresh_token):
    res = supabase.auth.refresh_token(refresh_token)
    return res



@app.post("/login")
async def sign_in(email, password):
    try:
        res = supabase.auth.sign_in_with_password(
            {"email": email, "password": password}
        )
        user_id = res.user.id
        access_token = res.session.access_token
        refresh_token = res.session.refresh_token

        store_session_check = supabase.table("Stores").select("*").filter("StoreID", "eq", user_id).execute()
        store_id = None
        if store_session_check and store_session_check.data:
            store_id = store_session_check.data[0].get("StoreID")

        userData = supabase.table("ConversAI_UserInfo").select("*").filter("user_id", "eq", user_id).execute().data
        username = userData[0]["username"]

        if not store_id:
            response = (
                supabase.table("Stores").insert(
                    {
                        "AccessToken": access_token,
                        "StoreID": user_id,
                        "RefreshToken": refresh_token,
                    }
                ).execute()
            )

            message = {
                "message": "Success",
                "code": status.HTTP_200_OK,
                "username": username,
                "user_id": user_id,
                "access_token": access_token,
                "refresh_token": refresh_token,
            }
            return message

        elif store_id == user_id:
            raise HTTPException(
                status_code=status.HTTP_400_BAD_REQUEST,
                detail="You are already signed in. Please sign out first to sign in again."
            )

        else:
            raise HTTPException(
                status_code=status.HTTP_400_BAD_REQUEST,
                detail="Failed to sign in. Please check your credentials."
            )

    except HTTPException as http_exc:
        raise http_exc

    except Exception as e:
        raise HTTPException(
            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
            detail=f"An unexpected error occurred during sign-in: {str(e)}"
        )


@app.post('login_with_token')
async def login_with_token(token):
    try:
        res = supabase.auth.sign_in_with_id_token(token)
        print(res)
        user_id = res.user.id
        access_token = res.session.access_token
        refresh_token = res.session.refresh_token

        store_session_check = supabase.table("Stores").select("*").filter("StoreID", "eq", user_id).execute()
        store_id = None

        if store_session_check and store_session_check.data:
            store_id = store_session_check.data[0].get("StoreID")

        if not store_id:
            response = (
                supabase.table("Stores").insert(
                    {
                        "AccessToken": access_token,
                        "StoreID": user_id,
                        "RefreshToken": refresh_token,
                    }
                ).execute()
            )

            message = {
                "message": "Success",
                "code": status.HTTP_200_OK,
                "user_id": user_id,
                "access_token": access_token,
                "refresh_token": refresh_token
            }
            return message

        elif store_id == user_id:
            raise HTTPException(
                status_code=status.HTTP_400_BAD_REQUEST,
                detail="You are already signed in. Please sign out first to sign in again."
            )

        else:
            raise HTTPException(
                status_code=status.HTTP_400_BAD_REQUEST,
                detail="Failed to sign in. Please check your credentials."
            )

    except HTTPException as http_exc:
        raise http_exc

    except Exception as e:
        raise HTTPException(
            status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
            detail=f"An unexpected error occurred during sign-in: {str(e)}"
        )



@app.post("/set-session-data")
async def set_session_data(access_token, refresh_token):
    res = supabase.auth.set_session(access_token, refresh_token)

    return res


@app.post("/logout")
async def sign_out(user_id):
    try:
        supabase.table("Stores").delete().eq(
            "StoreID", user_id
        ).execute()
        res = supabase.auth.sign_out()
        response = {"message": "success"}

        return response
    except Exception as e:
        raise HTTPException(status_code=400, detail=str(e))


@app.post("/oauth")
async def oauth(provider):
    res = supabase.auth.sign_in_with_oauth({"provider": provider})
    return res


@app.post("/newChatbot")
async def newChatbot(chatbotName: str, username: str):
    currentBotCount = len(listTables(username=username)["output"])
    limit = supabase.table("ConversAI_UserConfig").select("chatbotLimit").eq("user_id", username).execute().data[0][
        "chatbotLimit"]
    if currentBotCount >= int(limit):
        return {
            "output": "CHATBOT LIMIT EXCEEDED"
        }
    supabase.table("ConversAI_ChatbotInfo").insert({"user_id": username, "chatbotname": chatbotName}).execute()
    chatbotName = f"convai${username}${chatbotName}"
    return createTable(tablename=chatbotName)


@app.post("/addPDF")
async def addPDFData(vectorstore: str, pdf: UploadFile = File(...)):
    source = pdf.filename
    pdf = await pdf.read()
    with tempfile.NamedTemporaryFile(delete=False, suffix='.pdf') as temp_file:
        temp_file.write(pdf)
        temp_file_path = temp_file.name
    start = time.time()
    text = extractTextFromPdf(temp_file_path)
    textExtraction = time.time()
    os.remove(temp_file_path)
    username, chatbotname = vectorstore.split("$")[1], vectorstore.split("$")[2]
    df = pd.DataFrame(supabase.table("ConversAI_ChatbotInfo").select("*").execute().data)
    currentCount = df[(df["user_id"] == username) & (df["chatbotname"] == chatbotname)]["charactercount"].iloc[0]
    limit = supabase.table("ConversAI_UserConfig").select("tokenLimit").eq("user_id", username).execute().data[0][
        "tokenLimit"]
    newCount = currentCount + len(text)
    if newCount < int(limit):
        supabase.table("ConversAI_ChatbotInfo").update({"charactercount": str(newCount)}).eq("user_id", username).eq(
            "chatbotname", chatbotname).execute()
        uploadStart = time.time()
        output = addDocuments(text=text, source=source, vectorstore=vectorstore)
        uploadEnd = time.time()
        uploadTime = f"VECTOR UPLOAD TIME: {uploadEnd - uploadStart}s" + "\n"
        timeTaken = f"TEXT EXTRACTION TIME: {textExtraction - start}s" + "\n"
        tokenCount = f"TOKEN COUNT: {len(text)}" + "\n"
        tokenizer = nltk.tokenize.RegexpTokenizer(r"\w+")
        wordCount = f"WORD COUNT: {len(tokenizer.tokenize(text))}" + "\n"
        newText = ("=" * 75 + "\n").join([timeTaken, uploadTime, wordCount, tokenCount, "TEXT: \n" + text + "\n"])
        fileId = str(uuid.uuid4())
        with open(f"{fileId}.txt", "w") as file:
            file.write(newText)
        with open(f"{fileId}.txt", "rb") as f:
            supabase.storage.from_("ConversAI").upload(file = f, path = os.path.join("/", f.name), file_options={"content-type": "text/plain"})
        os.remove(f"{fileId}.txt")
        output["supabaseFileName"] = f"{fileId}.txt"
        return output
    else:
        return {
            "output": "DOCUMENT EXCEEDING LIMITS, PLEASE TRY WITH A SMALLER DOCUMENT."
        }


@app.post("/scanAndReturnText")
async def returnText(pdf: UploadFile = File(...)):
    source = pdf.filename
    pdf = await pdf.read()
    start = time.time()
    text = getTextFromImagePDF(pdfBytes=pdf)
    end = time.time()
    timeTaken = f"{end - start}s"
    return {
        "source": source,
        "extractionTime": timeTaken,
        "output": text 
    }


@app.post("/addText")
async def addText(vectorstore: str, text: str, source: str | None = None):
    username, chatbotname = vectorstore.split("$")[1], vectorstore.split("$")[2]
    df = pd.DataFrame(supabase.table("ConversAI_ChatbotInfo").select("*").execute().data)
    currentCount = df[(df["user_id"] == username) & (df["chatbotname"] == chatbotname)]["charactercount"].iloc[0]
    newCount = currentCount + len(text)
    limit = supabase.table("ConversAI_UserConfig").select("tokenLimit").eq("user_id", username).execute().data[0][
        "tokenLimit"]
    if newCount < int(limit):
        supabase.table("ConversAI_ChatbotInfo").update({"charactercount": str(newCount)}).eq("user_id", username).eq(
            "chatbotname", chatbotname).execute()
        uploadStart = time.time()
        output = addDocuments(text=text, source=source, vectorstore=vectorstore)
        uploadEnd = time.time()
        uploadTime = f"VECTOR UPLOAD TIME: {uploadEnd - uploadStart}s" + "\n"
        tokenCount = f"TOKEN COUNT: {len(text)}" + "\n"
        tokenizer = nltk.tokenize.RegexpTokenizer(r"\w+")
        wordCount = f"WORD COUNT: {len(tokenizer.tokenize(text))}" + "\n"
        newText = ("=" * 75 + "\n").join([uploadTime, wordCount, tokenCount, "TEXT: \n" + text + "\n"])
        fileId = str(uuid.uuid4())
        with open(f"{fileId}.txt", "w") as file:
            file.write(newText)
        with open(f"{fileId}.txt", "rb") as f:
            supabase.storage.from_("ConversAI").upload(file = f, path = os.path.join("/", f.name), file_options={"content-type": "text/plain"})
        os.remove(f"{fileId}.txt")
        output["supabaseFileName"] = f"{fileId}.txt"
        return output
    else:
        return {
            "output": "WEBSITE EXCEEDING LIMITS, PLEASE TRY WITH A SMALLER DOCUMENT."
        }


class AddQAPair(BaseModel):
    vectorstore: str
    question: str
    answer: str


@app.post("/addQAPair")
async def addQAPairData(addQaPair: AddQAPair):
    username, chatbotname = addQaPair.vectorstore.split("$")[1], addQaPair.vectorstore.split("$")[2]
    df = pd.DataFrame(supabase.table("ConversAI_ChatbotInfo").select("*").execute().data)
    currentCount = df[(df["user_id"] == username) & (df["chatbotname"] == chatbotname)]["charactercount"].iloc[0]
    qa = f"QUESTION: {addQaPair.question}\tANSWER: {addQaPair.answer}"
    newCount = currentCount + len(qa)
    limit = supabase.table("ConversAI_UserConfig").select("tokenLimit").eq("user_id", username).execute().data[0][
        "tokenLimit"]
    if newCount < int(limit):
        supabase.table("ConversAI_ChatbotInfo").update({"charactercount": str(newCount)}).eq("user_id", username).eq(
            "chatbotname", chatbotname).execute()
        return addDocuments(text=qa, source="Q&A Pairs", vectorstore=addQaPair.vectorstore)
    else:
        return {
            "output": "WEBSITE EXCEEDING LIMITS, PLEASE TRY WITH A SMALLER DOCUMENT."
        }


@app.post("/addWebsite")
async def addWebsite(vectorstore: str, websiteUrls: list[str]):
    start = time.time()
    text = extractTextFromUrlList(urls = websiteUrls)
    textExtraction = time.time()
    username, chatbotname = vectorstore.split("$")[1], vectorstore.split("$")[2]
    df = pd.DataFrame(supabase.table("ConversAI_ChatbotInfo").select("*").execute().data)
    currentCount = df[(df["user_id"] == username) & (df["chatbotname"] == chatbotname)]["charactercount"].iloc[0]
    newCount = currentCount + len(text)
    limit = supabase.table("ConversAI_UserConfig").select("tokenLimit").eq("user_id", username).execute().data[0]["tokenLimit"]
    if newCount < int(limit):
        supabase.table("ConversAI_ChatbotInfo").update({"charactercount": str(newCount)}).eq("user_id", username).eq(
            "chatbotname", chatbotname).execute()
        uploadStart = time.time()
        output = addDocuments(text=text, source=urlparse(websiteUrls[0]).netloc, vectorstore=vectorstore)
        uploadEnd = time.time()
        uploadTime = f"VECTOR UPLOAD TIME: {uploadEnd - uploadStart}s" + "\n"
        timeTaken = f"TEXT EXTRACTION TIME: {textExtraction - start}s" + "\n"
        tokenCount = f"TOKEN COUNT: {len(text)}" + "\n"
        tokenizer = nltk.tokenize.RegexpTokenizer(r"\w+")
        wordCount = f"WORD COUNT: {len(tokenizer.tokenize(text))}" + "\n"
        links = "LINKS:\n" + "\n".join(websiteUrls) + "\n"
        newText = ("=" * 75 + "\n").join([timeTaken, uploadTime, wordCount, tokenCount, links, "TEXT: \n" + text + "\n"])
        fileId = str(uuid.uuid4())
        with open(f"{fileId}.txt", "w") as file:
            file.write(newText)
        with open(f"{fileId}.txt", "rb") as f:
            supabase.storage.from_("ConversAI").upload(file = f, path = os.path.join("/", f.name), file_options={"content-type": "text/plain"})
        os.remove(f"{fileId}.txt")
        output["supabaseFileName"] = f"{fileId}.txt"
        return output
    else:
        return {
            "output": "WEBSITE EXCEEDING LIMITS, PLEASE TRY WITH A SMALLER DOCUMENT."
        }


@app.post("/answerQuery")
async def answerQuestion(query: str, vectorstore: str, llmModel: str = "llama3-70b-8192"):
    username, chatbotName = vectorstore.split("$")[1], vectorstore.split("$")[2]
    output = answerQuery(query=query, vectorstore=vectorstore, llmModel=llmModel)
    response = (
        supabase.table("ConversAI_ChatHistory")
        .insert({"username": username, "chatbotName": chatbotName, "llmModel": llmModel, "question": query, "response": output["output"]})
        .execute()
    )
    return output


@app.post("/deleteChatbot")
async def delete(chatbotName: str):
    username, chatbotName = chatbotName.split("$")[1], chatbotName.split("$")[2]
    supabase.table('ConversAI_ChatbotInfo').delete().eq('user_id', username).eq('chatbotname', chatbotName).execute()
    return deleteTable(tableName=chatbotName)


@app.post("/listChatbots")
async def delete(username: str):
    return listTables(username=username)


@app.post("/getLinks")
async def crawlUrl(baseUrl: str):
    return {
        "urls": getLinks(url=baseUrl, timeout=30)
    }


@app.post("/getCurrentCount")
async def getCount(vectorstore: str):
    username, chatbotName = vectorstore.split("$")[1], vectorstore.split("$")[2]
    df = pd.DataFrame(supabase.table("ConversAI_ChatbotInfo").select("*").execute().data)
    return {
        "currentCount": df[(df['user_id'] == username) & (df['chatbotname'] == chatbotName)]['charactercount'].iloc[0]
    }


@app.post("/getYoutubeTranscript")
async def getYTTranscript(urls: str):
    return {
        "transcript": getTranscript(urls=urls)
    }


@app.post("/analyzeData")
async def analyzeAndAnswer(query: str, file: UploadFile = File(...)):
    extension = file.filename.split(".")[-1]
    try:
        if extension in ["xls", "xlsx", "xlsm", "xlsb"]:
            df = pd.read_excel(io.BytesIO(await file.read()))
            response = analyzeData(query=query, dataframe=df)
        elif extension == "csv":
            df = pd.read_csv(io.BytesIO(await file.read()))
            response = analyzeData(query=query, dataframe=df)
        else:
            response = "INVALID FILE TYPE"
        return {
            "output": response
        }
    except:
        return {
            "output": "UNABLE TO ANSWER QUERY"
        }


@app.post("/getChatHistory")
async def chatHistory(vectorstore: str):
    username, chatbotName = vectorstore.split("$")[1], vectorstore.split("$")[2]
    response = supabase.table("ConversAI_ChatHistory").select("timestamp", "question", "response").eq("username", username).eq("chatbotName", chatbotName).execute().data
    return response