def get_dataframe_by_definition(def_dict):
"""Return dataframe corresponding to definition dict."""
file_path = os.path.join(def_dict['folder'], def_dict['filename'])
if 'anchor' in def_dict.keys():
gen = read_sheet(file_path, def_dict['sheet'], def_dict['anchor'])
else:
gen = read_sheet(file_path, def_dict['sheet'])
try:
df = get_dataframe(gen)[Regions.names()]
except:
raise ValueError(file_path)
df.insert(0, 'varname', def_dict['varname'])
return df
def get_dataframe_by_definition(def_dict):
"""Return dataframe corresponding to definition dict."""
file_path = os.path.join(def_dict['folder'], def_dict['filename'])
if 'anchor' in def_dict.keys():
gen = read_sheet(file_path, def_dict['sheet'], def_dict['anchor'])
else:
gen = read_sheet(file_path, def_dict['sheet'])
try:
df = get_dataframe(gen)[Regions.names()]
except:
raise ValueError(file_path)
df.insert(0, 'varname', def_dict['varname'])
return df
from datetime import date
import pandas as pd
import os
from xls_read import read_sheet, read_by_definition, yearmon
from regions import Regions
filter_region_name = Regions.filter_region_name
reference_region_names = Regions.names()
rf_name = Regions.rf_name()
district_names = Regions.district_names()
summable_regions = Regions.summable_regions()
def get_dataframe(datapoints_stream):
"""Return dataframe corresponding to datapoints stream."""
list_of_dicts = [{'val':x[0], 'region':x[1], 'dates':x[2]} for x in datapoints_stream]
df = pd.DataFrame(list_of_dicts)
df = df.pivot(columns='region', values='val', index='dates')[reference_region_names]
df.index = pd.DatetimeIndex(df.index)
return df
def get_dataframe_by_definition(def_dict):
"""Return dataframe corresponding to definition dict."""
file_path = os.path.join(def_dict['folder'], def_dict['filename'])
if 'anchor' in def_dict.keys():
gen = read_sheet(file_path, def_dict['sheet'], def_dict['anchor'])
else:
gen = read_sheet(file_path, def_dict['sheet'])
try:
df = get_dataframe(gen)[Regions.names()]
except:
import os
import sys
import pandas as pd
from getter import get_dataframe_by_definition
from regions import Regions
filter_region_name = Regions.filter_region_name
reference_region_names = Regions.names()
from definitions import definitions
# -------------------------------------------------------------------------
# Defintions
# -------------------------------------------------------------------------
from config import XL_SAMPLE_FOLDER
source_def_sample = {
'varname': 'PPI_PROM_ytd',
'folder': XL_SAMPLE_FOLDER,
'filename': 'industrial_prices.xls',
'sheet': 'пром.товаров',
'anchor': 'B5',
'anchor_value': 96.6
}
def_dict_2 = {
'varname': 'SHIPMENTS',
'folder': XL_SAMPLE_FOLDER,
'filename': 'shipment.xls',