app.py

import asyncio
import csv
import datetime
import json
import logging
import os
import os.path
import random
import re
import threading
import time

import discord
import gradio as gr
import gradio_client
import gspread
import numpy as np
import pandas as pd
import requests
from apscheduler.executors.pool import ThreadPoolExecutor
from apscheduler.schedulers.background import BackgroundScheduler
from discord import Color, Embed
from discord.ext import commands, tasks
from gradio_client import Client
from gspread_dataframe import get_as_dataframe, set_with_dataframe
from gspread_formatting.dataframe import format_with_dataframe
from huggingface_hub import HfApi, list_liked_repos, list_metrics, list_models
from tabulate import tabulate
from datetime import datetime

DISCORD_TOKEN = os.environ.get("DISCORD_TOKEN", None)
intents = discord.Intents.all() 
bot = commands.Bot(command_prefix='!', intents=intents)

#logger = logging.getLogger(__name__)
#logging.basicConfig(level=logging.DEBUG)


""""""
XP_PER_MESSAGE = 10 # 100k messages = 1M exp = lvl 100
""""""
service_account = json.loads(os.environ.get('KEY'))
file_path = 'service_account.json'
with open(file_path, 'w') as json_file:
    json.dump(service_account, json_file)
gspread_bot = gspread.service_account(filename='service_account.json')
worksheet = gspread_bot.open("levelbot").sheet1
worksheet2 = gspread_bot.open("hf_discord_verified_users_test").sheet1
""""""
bot_ids = [1136614989411655780, 1166392942387265536, 1158038249835610123, 1130774761031610388, 1155489509518098565, 1155169841276260546, 1152238037355474964, 1154395078735953930]
""""""
api = HfApi()
""""""
#csv_file = 'data.csv'
global_df = pd.DataFrame()
worksheet2_df = pd.DataFrame()
print(type(global_df))
community_global_df = pd.DataFrame() 
data = pd.read_csv("https://docs.google.com/spreadsheets/d/1hQSsIg1Y9WdBF_CdNM1L1rUUREoxKqRTe3_ILo-WK8w/export?format=csv&gid=0")
data2 = pd.read_csv("https://docs.google.com/spreadsheets/d/1T_twNONDSZ8zw7aDcYhw92DlawqPktti1xFw9ssJ13Y/export?format=csv&gid=0")


@bot.event
async def on_ready():
    try:
        global global_df
        global worksheet2_df
        await asyncio.sleep(1.1)
        print(f'Logged in as {bot.user.name}')
        print(f"XP_PER_MESSAGE: {XP_PER_MESSAGE}")
    
        """import data from google sheets -> HF Space df (doesn't make API call this way, as it's read-only)""" 
        data['discord_user_id'] = data['discord_user_id'].astype(str)
        global_df = data
        print(f"csv successfully retrieved: \n {global_df}")



        worksheet2_df = data2.copy()
        worksheet2_df['discord_user_id'] = worksheet2_df['discord_user_id'].astype(str)
        worksheet2_df['id_recover_test'] = worksheet2_df['id_recover_test'].astype(str)  
        print(worksheet2_df.dtypes)

        # updates both leaderboards
        remove_huggingfolks.start()
        await periodic_api_test()
        print(f"------------------------------------------------------------------------")
    except Exception as e:
        print(f"on_message Error: {e}")


def update_google_sheet():
    """save data from HF Space -> google sheets (makes 2 API calls)"""
    try:
        print("Updating google sheets...")
        name = "levelbot"
        #worksheet = gspread_bot.open(name).sheet1
        global_df['discord_user_id'] = global_df['discord_user_id'].astype(str)
        # could just do this / format
        set_with_dataframe(worksheet, global_df)
        # very important, otherwise gspread will force format IDs in scientific notation by default (making them unusable)
        worksheet.format(f"A1:A{len(global_df.index)+1}", {
        	"numberFormat": {
        		"type": 'TEXT',
        	},
        })
        timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
        print(f"Google sheet {name} successfully updated at {timestamp}! \n{global_df}")
        


        worksheet2.format(f"E1:E{len(worksheet2_df.index)+1}", {
        	"numberFormat": {
        		"type": 'TEXT',
        	},
        })
        worksheet2.format(f"A1:A{len(worksheet2_df.index)+1}", {
        	"numberFormat": {
        		"type": 'TEXT',
        	},
        })              
        name2 = "hf_discord_verified_users_test"
        # could just do this / format
        worksheet2['discord_user_id'] = worksheet2['discord_user_id'].astype(str)
        worksheet2['id_recover_test'] = worksheet2['id_recover_test'].astype(str)
        print(worksheet2_df.dtypes)
        set_with_dataframe(worksheet2, worksheet2_df)
        # very important, otherwise gspread will force format IDs in scientific notation by default (making them unusable)
  
        
        timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")
        print(f"Google sheet {name2} successfully updated at {timestamp}!")
        print(f"------------------------------------------------------------------------")
    except Exception as e:
        print(f"on_message Error: {e}")


executor = ThreadPoolExecutor(max_workers=1)
scheduler = BackgroundScheduler(executors={'default': executor})
scheduler.add_job(update_google_sheet, trigger='interval', minutes=1, max_instances=1)
scheduler.start()


def calculate_level(xp):
    return int(xp ** (1.0 / 3.0))


def calculate_xp(level):
    return (int(level ** 3))


processed_users = set()

async def periodic_api_test(): # needs rewrite, can do same thing and interact with csv -> push to google sheets
    try:
        # this should be gspread type function (do not mix with discord at all)
        global worksheet2_df
        # get 
        for index, user in enumerate(worksheet2_df['hf_user_name']):
            if user in processed_users:
                continue
            await asyncio.sleep(0.1)
            url = f"https://huggingface.co/api/users/{user}/overview"
            response = requests.get(url)
    
            if response.status_code == 200:
                data = response.json()
    
                likes = data["numLikes"]
                models = data["numModels"]
                datasets = data["numDatasets"]
                spaces = data["numSpaces"]
                discussions = data["numDiscussions"]
                papers = data["numPapers"]
                upvotes = data["numUpvotes"]

                worksheet2_df.loc[index:index, 'likes':'upvotes'] = [[likes, models, datasets, spaces, discussions, papers, upvotes]]
                pd.set_option('display.max_rows', 30)
                print(worksheet2_df)
                
    
                processed_users.add(user)
                print(processed_users)
            else:
                print(f"Failed to retrieve data for user {user}. Status code: {response.status_code}")
            
    except Exception as e:
        print(f"periodic_api_test Error: {e}")    

    
async def add_exp(member_id):
    try:
        """Uses member_id to create new record or update old one [member_id_column] ... [member_exp_column]"""
        await asyncio.sleep(0.1)
        global global_df
        
        guild = bot.get_guild(879548962464493619)
        member = guild.get_member(member_id) # bot.get_user == User, guild.get_member == Member (not the same thing!)

        lvl1 = guild.get_role(1171861537699397733)
        lvl2 = guild.get_role(1171861595115245699)
        lvl3 = guild.get_role(1171861626715115591)
        lvl4 = guild.get_role(1171861657975259206)
        lvl5 = guild.get_role(1171861686580412497)
        lvl6 = guild.get_role(1171861900301172736)
        lvl7 = guild.get_role(1171861936258941018)
        lvl8 = guild.get_role(1171861968597024868)
        lvl9 = guild.get_role(1171862009982242836)
        lvl10 = guild.get_role(1164188093713223721)
        lvl11 = guild.get_role(1171524944354607104)
        lvl12 = guild.get_role(1171524990257082458)
        lvl13 = guild.get_role(1171525021928263791)
        lvl14 = guild.get_role(1171525062201966724)
        lvl15 = guild.get_role(1171525098465918996)
        lvl16 = guild.get_role(1176826165546201099)
        lvl17 = guild.get_role(1176826221301092392)
        lvl18 = guild.get_role(1176826260643659776)
        lvl19 = guild.get_role(1176826288816791693)
        lvl20 = guild.get_role(1176826319447801896)

        lvl21 = guild.get_role(1195030831174008902)
        lvl22 = guild.get_role(1195030883351150592)
        lvl23 = guild.get_role(1196055555006009445)
        lvl24 = guild.get_role(1196055640917938216)
        lvl25 = guild.get_role(1196055712506318869)
        lvl26 = guild.get_role(1196055775924195378)
        lvl27 = guild.get_role(1196055837018435664)
        lvl28 = guild.get_role(1196055908267081849)
        lvl29 = guild.get_role(1196055970804150352)
        lvl30 = guild.get_role(1196056027720847380)
        
        lvls = {
            1: lvl1, 2: lvl2, 3: lvl3, 4: lvl4, 5: lvl5, 6: lvl6, 7: lvl7, 8: lvl8, 9: lvl9, 10: lvl10,
            11: lvl11, 12: lvl12, 13: lvl13, 14: lvl14, 15: lvl15, 16: lvl16, 17: lvl17, 18: lvl18, 19: lvl19, 20: lvl20,
            21: lvl21, 22: lvl22, 23: lvl23, 24: lvl24, 25: lvl25, 26: lvl26, 27: lvl27, 28: lvl28, 29: lvl29, 30: lvl30,
        }              

        # xp = int64
        # member_name = string
        # discord_user_id = string

        member_found = False
        for index, cell_value in global_df.iloc[:, 0].items():
            if cell_value == str(member_id):
                # if found, update that row...
                member_found = True
                print(f"Record for {member} found at row {index + 1}, column 1")
        
                # increment the old experience value (better not to replace outright)
                old_xp = global_df.loc[index, 'discord_exp']
                new_xp = old_xp + XP_PER_MESSAGE
                global_df.loc[index, 'discord_exp'] = new_xp # do not change column name
                print(f"Record for {member} updated from {old_xp} to {global_df.loc[index, 'discord_exp']} (+{XP_PER_MESSAGE}) ")

          
                # level up
                current_level = calculate_level(new_xp)
                print(f"Current_level for {member}: {current_level}")
                if current_level >= 2 and current_level <=30:
                    current_role = lvls[current_level]
                    if current_role not in member.roles: # if we leveled up
                        await member.add_roles(current_role)
                        print(f"Level Up! Gave {member} {current_role}")
                        await member.remove_roles(lvls[current_level-1])
                        print(f"Removed {lvls[current_level-1]} from {member}")  
                        #print(f"{member} Level up! {current_level-1} -> {current_level}!")
                        if current_role in member.roles: # needs update; reference exp reward for verification
                            #await member.send(f"Level up! {current_level-1} -> {current_level}!")  
                            #if member_id == 811235357663297546:

                            row = global_df[global_df['discord_user_id'] == str(member_id)] # does this need to be string?
                            if not row.empty:
                                target_exp = row['discord_exp'].values[0]
                                rank = (global_df['discord_exp'] > target_exp).sum() + 1
                                print(f"The rank for discord_id {member_id} based on discord_exp is: {rank}")
                            else:
                                print(f"Discord ID {member_id} not found in the DataFrame.") 
                    
                            # send embed
                            embed = Embed(color=Color.blue())
                            embed.set_author(name=f"{member}", icon_url=member.avatar.url if member.avatar else bot.user.avatar.url)
                            embed.title = f"Level Up! `{current_level-1}` -> `{current_level}`"
                            msg = f'🤗 Congrats {member}! You just leveled up in the Hugging Face Discord server'
                            embed.description = f"{msg}."
                            embed.add_field(name="Leaderboard Ranking:", value=f"👑 **{rank}**\n\nhttps://discord.com/channels/879548962464493619/1197143964994773023", inline=True)
                            # can use ^ to track changes in leaderboard ranking and display in levelup message when climbing
                            # (green red arrow thingies)
                            # could also pull ranks above and below yourself (if they exist, i.e.)
                            # 1 nateraw
                            # 2 [you are here]
                            # 3 osanseviero
                            # also, ahead by how much exp, how many levels maybe
                            msg3 = "- Posting\n- Reacting / being reacted to\n- Being active on the Hugging Face Hub (verify to link your Hub + Discord accounts!)"
                            embed.add_field(name="How to Level Up:", value=msg3, inline=True)
                            verification_link = "https://discord.com/channels/879548962464493619/900125909984624713"
                            embed.add_field(name="Verify Here:", value=verification_link, inline=True)

                            #lunar = bot.get_user(811235357663297546)
                            await member.send(embed=embed)                           
                            print(f"Sent levelup embed to {member}")
                            #You can verify your account to earn 100 points! To verify, do A.                            

                            
                print(f"------------------------------------------------------------------------")
        if not member_found:
            # if not, create new record
            print(f"creating new record for {member}")            
            
            xp = 10 # define somewhere else?
            member_id = str(member_id)
            member_name = str(member.name)
            current_level = calculate_level(xp)
            row_data = [member_id, member_name, xp, current_level]
            global_df.loc[len(global_df.index)] = row_data 
            
            # initial role assignment
            if current_level == 1:
                if lvl1 not in member.roles:
                    await member.add_roles(lvl1)
                    print(f"Gave {member} {lvl1}")  # can log this better       
            print(f"------------------------------------------------------------------------")
    except Exception as e:
        print(f"add_exp Error: {e}")   


@bot.event
async def on_message(message):
    try:
        if message.author.id not in bot_ids: # could change to if author does not have bot role (roleid)
            if "!help_xp" not in message.content:
                print(f"adding exp from message {message.author}")
                await asyncio.sleep(0.1)
                await add_exp(message.author.id)
        await bot.process_commands(message)
    except Exception as e:
        print(f"on_message Error: {e}")

        
@bot.event
async def on_reaction_add(reaction, user):
    try:
        if user.id not in bot_ids:
            print(f"adding exp from react {user}")
            await asyncio.sleep(0.1)            
            await add_exp(user.id)
            await asyncio.sleep(0.1)
            await add_exp(reaction.message.author.id)
    except Exception as e:
        print(f"on_reaction_add Error: {e}")


@tasks.loop(minutes=1)
async def remove_huggingfolks():
    try:
        # remove huggingfolks
        global community_global_df

        community_global_df = global_df.copy()
        
        guild = bot.get_guild(879548962464493619)
        role = discord.utils.get(guild.roles, id=897376942817419265)
        members_with_role = [member.id for member in guild.members if role in member.roles]

        for member_id in members_with_role:
            community_global_df = community_global_df[community_global_df.iloc[:, 0] != str(member_id)]

        #print(community_global_df)

        community_global_df.drop(community_global_df.columns[0], axis=1, inplace=True)
        community_global_df['discord_exp'] = pd.to_numeric(community_global_df['discord_exp'])
        community_global_df = community_global_df.nlargest(len(community_global_df), 'discord_exp')        
        
        top_30_exp = community_global_df.nlargest(30, 'discord_exp')
        
        top_30_exp['D'] = ['🥇','🥈','🥉','','','','','','','','','','','','','','','','','','','','','','','','','','','']
        top_30_rows = top_30_exp.values.tolist()

        #print(top_30_rows)
        channel = bot.get_channel(1197143964994773023)
        message = await channel.fetch_message(1197148293164187678)

        # put into message / leaderboard
        new_table = tabulate(top_30_rows, headers=["Name", "Experience", "Level", "Rank"], tablefmt="plain")
        await message.edit(content=f"Updated Leaderboard:\n```\n{new_table}\n```")
        print(f"Updated discord leaderboard!")
        
    except Exception as e:
        print(f"remove_huggingfolks Error: {e}")


@bot.command(name='xp_help')
async def xp_help(ctx):
    try:
        help_message = "How to earn Discord / Hub exp: Post messages, react, Like, discuss, create repos and papers"
        await ctx.author.send(help_message)
    except Exception as e:
        print(f"on_message Error: {e}")
        

@bot.command()
async def count_users_with_role(ctx, role_id):
    role = discord.utils.get(ctx.guild.roles, id=int(role_id))
    count = sum(1 for member in ctx.guild.members if role in member.roles)
    await ctx.send(f"Number of users with the role '{role.name}': {count}")



# embeds with user pfps? 
# name, pfp, time in server....

# weekly do different count
# count number of messages per user for every channel (total messages)
# fix sheet if necessary
# add emojis for some color
# check if members are still in the server 


""""""
DISCORD_TOKEN = os.environ.get("DISCORD_TOKEN", None)
def run_bot():
    bot.run(DISCORD_TOKEN)
threading.Thread(target=run_bot).start()


URL = "https://docs.google.com/spreadsheets/d/1hQSsIg1Y9WdBF_CdNM1L1rUUREoxKqRTe3_ILo-WK8w/edit#gid=0"
csv_url = URL.replace('/edit#gid=', '/export?format=csv&gid=')


def get_data():
    try:
        return community_global_df   
    except Exception as e:
        print(f"on_message Error: {e}")


demo = gr.Blocks()
with demo:
    try:
        dataframe1 = pd.read_csv(csv_url)
        column_values_unique = sorted(dataframe1.iloc[:, 3].unique())
        dataframe2 = pd.DataFrame({'Levels': column_values_unique})
        counts = {}
        for value in data.iloc[:, 3]:
            counts[value] = counts.get(value, 0) + 1
        dataframe2['Members'] = dataframe2['Levels'].map(counts)
    
        print("\nDataframe 2:")
        print(dataframe2)
    
        TITLE = """<h1 align="center" id="space-title">🤗 Hugging Face Level Leaderboard</h1>"""
        gr.HTML(TITLE)
        with gr.Tabs(elem_classes="tab-buttons") as tabs:
            with gr.TabItem("🏅 Level leaderboard", elem_id="level-table", id=0):
                #gr.Markdown("# 📈 Experience Leaderboard")
                with gr.Row():
                    with gr.Column():
                        gr.DataFrame(get_data, every=5, height=500, interactive=False, col_count=(3, "fixed"), column_widths=["100px","100px","100px"])
    
                    with gr.Column():
                        gr.BarPlot(
                            value=dataframe2,
                            x="Levels",
                            y="Members",
                            title="Level Distribution",
                            height=450,
                            width=450,
                            interactive=False
                        )                   
            #with gr.TabItem("📈 Members of the Week", elem_id="week-table", id=1):
         
            #with gr.TabItem("📈 Hub-only leaderboard", elem_id="hub-table", id=2):
    except Exception as e:
        print(f"on_message Error: {e}")
demo.queue().launch()