maxwellbade

def fraud_rate(df, agg, cols=None, threshold=None, limit=None, days=None):
    if isinstance(cols, str):
        groupcols = [cols]
    elif cols is None:
        groupcols = [] 
    else:
        try:
            groupcols = list(cols)
        except:
            raise TypeError('Unable to convert cols to a list.')

maxwellbade

def multiple_defect(nds, agg, cols, n, minrate=0.05):
    import itertools
    combos = list(itertools.combinations(cols, n))
    num_cols = ['col' + str(i) for i in range(1, n + 1)]
    res = []
    for combo in combos:
        fr = fraud_rate(nds, agg, combo)
        fr = fr.reset_index()
        fr.columns = num_cols + ['fraud', 'nonfraud', 'fraudrate','fpr']
        fr = fr.assign(groupcols = ', '.join(combo))

maxwellbade

#random pivot df for a plot

create_date_df = df[['Microsoft Account Created Date', 'fraud']]

create_date_df = pd.pivot_table(create_date_df
               , index=['Microsoft Account Created Date']
               , columns = 'fraud'
               , aggfunc=len
              ).reset_index()

maxwellbade

dirpath = '/Users/max.bade/Downloads/pancakeswap/'
output = '/Users/max.bade/Downloads/pancakeswap/output/pancakeswap.csv'
csvout_lst = []
files = [os.path.join(dirpath, fname) for fname in os.listdir(dirpath)]

# for filename in sorted(files):
for filename in cakes:
    
    def cross_join(left, right):
        new_rows = []

maxwellbade

cake_4_1 = json.load(open('/Users/max.bade/Downloads/pancakeswap/4_1.txt'))
cake_4_2 = json.load(open('/Users/max.bade/Downloads/pancakeswap/4_2.txt'))
cake_4_3 = json.load(open('/Users/max.bade/Downloads/pancakeswap/4_3.txt'))
cake_4_4 = json.load(open('/Users/max.bade/Downloads/pancakeswap/4_4.txt'))
cake_4_5 = json.load(open('/Users/max.bade/Downloads/pancakeswap/4_5.txt'))
cake_4_6 = json.load(open('/Users/max.bade/Downloads/pancakeswap/4_6.txt'))
cake_4_7 = json.load(open('/Users/max.bade/Downloads/pancakeswap/4_7.txt'))
cake_4_8 = json.load(open('/Users/max.bade/Downloads/pancakeswap/4_8.txt'))
cake_4_9 = json.load(open('/Users/max.bade/Downloads/pancakeswap/4_9.txt'))
cake_4_10 = json.load(open('/Users/max.bade/Downloads/pancakeswap/4_10.txt'))

maxwellbade

df = pd.read_csv(output,)
df.columns = ['index_duplicate','buy_symbol','sell_symbol','#_trades','trade_amount_$']
del df['index_duplicate']
df['num_trades_x_tradeamount'] = df['#_trades'] * df['trade_amount_$']
# df = df.sort_values(by='num_trades_x_tradeamount',ascending=False).reset_index()
# del df['index']

start = pd.to_datetime("4/1/2021")
df['trade_date'] = pd.Series(np.arange(len(df)) // 1000) \
              .apply(lambda x: pd.Timedelta(days=x)) + start

maxwellbade

filtered_coins = df.groupby(['Week_Number','pair'])['pct_ch'].agg('sum').reset_index()
filtered_coins.columns = ['Week_Number','pair','total_weekly_pct_ch']
filtered_coins = filtered_coins[filtered_coins['Week_Number'] == 18][['Week_Number','pair','total_weekly_pct_ch']].sort_values(by='total_weekly_pct_ch',ascending=False).reset_index()
filtered_coins = filtered_coins[filtered_coins['total_weekly_pct_ch']>=50]
print(filtered_coins.shape)
print(filtered_coins.pair.nunique())
filtered_coins.head(10)

maxwellbade

fig = px.bar(filtered_coins
             ,x='pair'
             ,y='total_weekly_pct_ch'
             ,color='pair'
#              ,size='total_weekly_pct_ch'
#              ,facet_row='pair'
#              ,facet_row_spacing=.005
#              ,facet_col='Week_Number'
#              ,facet_col_wrap=2
#              ,log_y=True

maxwellbade

fig = px.line(df2
             ,x='trade_date'
             ,y='pct_ch'
             ,color='pair'
             ,facet_row='pair'
             ,facet_row_spacing=.005
#              ,facet_col='Week_Number'
#              ,facet_col_wrap=2
#              ,log_y=True
             ,range_y=[0, 8000]

maxwellbade

coins_to_remove = ['BUSD','WBNB','ETH','USDT','USDC','BTC','Cake']
df = df[~df.buy_symbol.isin(coins_to_remove)]
print(df.shape)
print('unique pairs: ',df.pair.nunique())
df.head(20)
df2 = df
df2 = df2[df2['sum_pct_change'] >= 500]
print(df2.pair.nunique())
df2.head()