import get_functions as gf
# parametros (junto a los nombres en namesferia_[fecha].txt)
dia = "2018-11-03"
timezone = "America/Buenos_Aires"
hora_inicio = "05:00:00"
hora_fin = "22:00:00"
omit_accounts = ['1.2.150830', '1.2.151476', '1.2.667678',
'1.2.1095159'] # gob, propuesta, pamelaps y nodomardelplata
# %%
# lee nombres input
with open('data/input/namesferia_' + dia + '.txt', 'r') as f:
names_feria = pd.Series(f.read().splitlines())
# accounts de todas las cuentas Par
accounts = gf.get_accounts(prefix='moneda-par')
# STOP si no encuentra un nombre de la feria entre las cuentas par
wrong_names = names_feria[~names_feria.isin(accounts.name)]
if len(wrong_names) > 0:
sys.exit('NON-EXISTENT ACCOUNT NAMES: ' +
', '.join(str(x) for x in list(wrong_names)))
# accounts de participantes
accounts_feria = accounts.loc[accounts.name.isin(names_feria)]
# %% history de cada usuario (lee 80 tx max por usuario - 4*20)
history_user = [
gf.get_user_history(user_id=x, max_page_num=4)
for x in list(accounts_feria.id_user)
]
# %% dataframe de txs
import numpy as np
import pandas as pd
import pickle
import get_functions as gf
tk_id = '1.3.1236' # id de token MONEDAPAR
names_prefix = r'moneda-par' # prefijos de las cuentas de par
with open('data/input/names_mdq.txt','r') as f:
names_mdq = pd.Series(f.read().splitlines()) # nombres de mdq
omit_accounts = ['1.2.150830','1.2.151476','1.2.667678'] # id cuenta del gob, propuesta y pamelaps
dia = '2018-09-11'
accounts = gf.get_accounts(prefix=names_prefix)
accounts_mdq = accounts.loc[accounts.name.isin(names_mdq)]
accounts_history = pickle.load(open("data/accounts_history_"+dia+".p", "rb"))
accounts_history_mdq = accounts_history[accounts_mdq.id_user]
txs_accounts_mdq = [gf.get_user_txs_fromhistory(json_account_history=i,token_id=tk_id) for i in accounts_history_mdq]
txs_mdq = pd.concat(txs_accounts_mdq).drop_duplicates().sort_values('datetime', ascending=True)
# merge con nombres de cuentas
txs = pd.merge(txs_mdq, accounts, how='left', left_on='sender_id', right_on='id_user').drop('id_user',axis=1)
txs = txs.rename(columns={'name':'sender_name'})
txs = pd.merge(txs, accounts, how='left', left_on='recipient_id', right_on='id_user').drop('id_user',axis=1)
txs = txs.rename(columns = {'name':'recipient_name'})
# omite txs en las que participan omit_accounts
txs = txs.loc[~(txs.sender_id.isin(omit_accounts) | txs.recipient_id.isin(omit_accounts)),:]
# regenera index
txs.index = range(len(txs.index))
# data['time'] = pd.to_datetime(data['time'])
import get_functions as gf
#%% parameters
dia = '2018-10-24'
# history de transacciones
txs_full = pickle.load(open("output/raw/accounts_txs_history_"+dia+".p", "rb"))
#%% filtros
# solo MONEDAPAR
txs_par = txs_full.loc[txs_full.asset_name=="MONEDAPAR"]
# omite txs en las que participan omit accounts (ver gf.filter_usersintx)
txs_parf = gf.filter_usersintx(txs_par)
# omite casos especiales
txs_parf = gf.filter_specialtx(txs_parf)
# merge con otra data y correcion de datetime
txs_parf = gf.merge_txs_data(txs_parf, accounts_df=gf.get_accounts(), tokens_df=gf.token_data())
#%% variables year-month y year-month-day
ym = txs_parf.datetime.dt.to_period(freq='m')
ymd = txs_parf.datetime.dt.to_period(freq='d')
txs_parf.loc[:,'ym'] = ym.values
txs_parf.loc[:,'ymd'] = ymd.values
# ym y ymd de todo el periodo (por si falta alguno)
all_ym = pd.PeriodIndex(start=txs_parf.datetime.min(), end=dia, freq='m')
all_ym_df = pd.Series(np.full_like(all_ym,0), index=all_ym, dtype='int64')
all_ymd = pd.PeriodIndex(start=txs_parf.datetime.min(), end=dia, freq='d')
all_ymd_df = pd.Series(np.full_like(all_ymd,0), index=all_ymd, dtype='int64')
#%% number of txs (ym y ymd)
# ym
n_ym = txs_parf.groupby('ym').count().amount.add(all_ym_df, fill_value=0)
#%% parameters
dia = '2018-11-01'
# history de transacciones
txs_full = pickle.load(
open("output/raw/accounts_txs_history_" + dia + ".p", "rb"))
#%% filtros
# solo MONEDAPAR (obsoleto)
# txs_par = txs_full.loc[txs_full.asset_name=="MONEDAPAR"]
# omite txs en las que participan omit accounts (ver gf.filter_usersintx)
txs_parf = gf.filter_usersintx(txs_full)
# omite casos especiales
txs_parf = gf.filter_specialtx(txs_parf)
# merge con otra data y correcion de datetime
txs_parf = gf.merge_txs_data(txs_parf,
accounts_df=gf.get_accounts(),
tokens_df=gf.token_data())
#%% variables year-month y year-month-day
ym = txs_parf.datetime.dt.to_period(freq='m')
ymd = txs_parf.datetime.dt.to_period(freq='d')
txs_parf.loc[:, 'ym'] = ym.values
txs_parf.loc[:, 'ymd'] = ymd.values
# ym y ymd de todo el periodo (por si falta alguno)
all_ym = pd.PeriodIndex(start=txs_parf.datetime.min(), end=dia, freq='m')
all_ymd = pd.PeriodIndex(start=txs_parf.datetime.min(), end=dia, freq='d')
#%% active users by ym y ymd
# para cada ym y ymd, calcula unique de senders+recipientes
# lo hago loop porque no se con sql :(
# para ym