#!/usr/bin/env python
# coding: utf-8

# Let's find the top individual series by shot attempts.

# In[1]:


import PbPMethods2 as pm2
import pandas as pd


# In[2]:


dfs = []
for season in range(2007, 2017):
    dfs.append(pd.read_csv(pm2.get_gamebygame_data_filename(season)))
dfs = pd.concat(dfs)
dfs.head()


# Filter for the playoffs and add a variable with the round number

# In[3]:


playoffs = dfs.query('Game > 30000')
playoffs = playoffs[['Player', 'Team', 'Game', 'Season', 'TOION(60s)', 'CFON', 'CAON']]
playoffs['Round'] = (playoffs['Game'] - 30000) // 100
playoffs.head()


# Group by round to get gp and counts

# In[4]:


player_gp = playoffs.groupby(['Player', 'Team', 'Season', 'Round']).count()
player_gp.reset_index(inplace = True)
player_gp.rename(columns = {'Game': 'GP'}, inplace = True)
player_gp = player_gp[['Player', 'Team', 'Season', 'Round', 'GP']]
player_gp.head()


# In[5]:


player_rounds = playoffs.groupby(['Player', 'Team', 'Season', 'Round']).sum()
player_rounds.drop(['Game'], axis = 1, inplace = True)
player_rounds.reset_index(inplace = True)

#Join to get the gp column
player_rounds = player_rounds.merge(player_gp, on = ['Player', 'Team', 'Season', 'Round'], how = 'inner')

player_rounds['CF60'] = player_rounds['CFON'] / player_rounds['TOION(60s)']
player_rounds['CA60'] = player_rounds['CAON'] / player_rounds['TOION(60s)']
player_rounds['CD60'] = player_rounds['CF60'] - player_rounds['CA60']
player_rounds.sort_values(by = 'CD60', ascending = False, inplace = True)
player_rounds.head()


# Filter for at least six games (optional)

# In[6]:


temp = player_rounds.query('GP >= 6')
temp.reset_index(inplace = True, drop = True) #so index = ranks
temp.head(20)