import numpy as np 
import pandas as pd 

from datetime import datetime, date

import matplotlib.pyplot as plt 
import seaborn

import yfinance as yf

# import holoviews as hvs
import panel as pn
# import hvplot as hv
# import hvplot.pandas

pn.extension("bokeh", template="bootstrap")

import os

from dotenv import load_dotenv
import holoviews as hv

# from datasets import load_dataset
from utils import load_hf_dataset  # ,upload_to_hf_dataset, download_from_hf_dataset

# Load environment variables from .env file
load_dotenv()

# Get the Hugging Face API token from the environment; either set in .env file or in the environment directly in GitHub
HF_TOKEN = os.getenv("HF_TOKEN")

# Get the name of the GuruFocus dataset for TradingView, Finviz, MarketBeat and TipRanks to read from
dataset_name_TradingView_input = os.getenv("dataset_name_TradingView_input")


@pn.cache()
def get_tradingview(current_datetime):
    # Load lastest TradingView DataSet from HuggingFace Dataset which is always america.csv
    # download_from_hf_dataset("america.csv", dataset_name_TradingView_input, HF_TOKEN)
    return load_hf_dataset("america.csv", HF_TOKEN, dataset_name_TradingView_input)


@pn.cache()
def get_stock_history(ticker, startdate, enddate, current_datetime):
  return yf.Ticker(ticker).history(start=startdate, end=enddate)
  

# Reading gurufocus,finviz from crawling pipelines with GitHub Action
current_datetime = datetime.now().strftime("%Y-%m-%d")
DF = get_tradingview(current_datetime)
# DF.Industry.unique()


def get_plot_Sector(selected_Sector,  startdate , enddate, num_stocks, method_selector):
  # selected_Industry_or_Sector= "Major Banks"
  # ticker_list = list(DF.query('Sector==@selected_Sector & `Market Capitalization`>100e9').Ticker)
  ticker_list = list(DF.query('Sector==@selected_Sector').sort_values('Market Capitalization', ascending=False).Ticker[0:num_stocks])
  len(ticker_list)
  # ticker_list = ['AAPL', 'AMZN', 'GOOGL','TSLA', 'NVDA' , 'AMD']

  # start = datetime(2020, 1, 1)
  # end = datetime.now()
  # end = datetime(2020,1,1)
  method = method_selector #'pearson'#, 'kendall', 'spearman'

  tickers=[]
  for ticker in ticker_list:     
    if not '/' in ticker:
      # df = yf.Ticker(ticker).history(start=startdate , end=enddate)
      df = get_stock_history(ticker, startdate, enddate, current_datetime)
      df['Ticker'] = ticker    
      tickers.append(df)
  DFtotal = pd.concat(tickers)
  DFtotal = DFtotal.reset_index()
  DFtotal = DFtotal[['Date', 'Close', 'Ticker']]
  df_pivot = DFtotal.pivot(index='Date', columns='Ticker', values='Close').reset_index()
  df_pivot = df_pivot.drop('Date', axis=1)
  corr_df = df_pivot.corr(method=method)

  # Convert correlation matrix to format suitable for heatmap
  corr_reset = corr_df.reset_index()
  melted_corr = pd.melt(corr_reset, id_vars='Ticker', var_name='Variable', value_name='Correlation')

  # Create heatmap using holoviews
  hv.extension('bokeh')

  heatmap = hv.HeatMap(melted_corr, kdims=['Ticker', 'Variable'], vdims='Correlation')
  heatmap = heatmap.opts(
    title=f'Heat Map correlation based on {method} method for {selected_Sector} sector from {startdate.year} - {enddate.year}',
    width=800,
    height=600,
    tools=['hover'],
    colorbar=True,
    cmap='RdYlGn',
    xrotation=45,
    labelled=[],
    fontsize={'title': 12, 'labels': 10},
  )
  return heatmap

# Create dropdown widget for industry selection
industry_selector = pn.widgets.Select(
  name='Select Industry',
  options=list(DF.Industry.unique()),
  value='Major Banks'
)
num_stocks = pn.widgets.IntSlider(
  name='Number of Stocks',
  start=5,
  end=30,
  value=10,
  step=1
)

method_selector = pn.widgets.Select(
  name='Correlation Method',
  options=['pearson', 'kendall', 'spearman'],
  value='pearson'
)
# Create dropdown widget for industry selection
sector_selector = pn.widgets.Select(
  name='Select Sector',
  options=list(DF.Sector.unique()),
  value='Health Technology'
)

date_start = pn.widgets.DatePicker(
    name ="Start Date",
    description='Select a Date',
    start= date(2000, 1, 1)
)

date_end = pn.widgets.DatePicker(
    name ="End Date",# value=datetime(2000, 1, 1),
    description='Select a Date',
    end= date.today() #date(2023, 9, 1)
)

date_start.value = date(2020,1,1)
date_end.value = date.today()


tab_Sector_content = pn.Row(pn.Column(date_start, date_end, sector_selector, num_stocks, method_selector ), pn.bind(get_plot_Sector, sector_selector, date_start, date_end, num_stocks, method_selector))
# tab_Industry_content = pn.Row(pn.Column(date_start, date_end, industry_selector ), pn.bind(get_plot_Industry,industry_selector, date_start, date_end))
pn.Tabs(
    ("Sector", tab_Sector_content),
    # ("Tab Industry", tab_Industry_content)
).servable('Stock Correlation Heatmap')

# pn.Row(pn.Column(date_start, date_end, industry_selector ), pn.bind(get_plot,industry_selector, date_start, date_end)).servable('US Stock Correlation Heatmap')