Use openpyxl for .xlsx files (a7eb026d) · Commits · energyincities / pyehub

pyehub/excel_to_request_format.py

+7 −5

Original line number	Diff line number	Diff line
		@@ -18,6 +18,7 @@ import argparse
		import json
		from contextlib import suppress

		import pathlib
		import xlrd
		import openpyxl

		@@ -69,9 +70,11 @@ class Converter:
		"""
		self._excel_file = excel_file

		if self._excel_file.endswith('.xls'):
		# function to return the file extension
		file_extension = pathlib.Path(excel_file).suffix
		if file_extension == '.xls':
		self._kind = 'xls'
		elif self._excel_file.endswith('.xlsx'):
		elif file_extension == '.xlsx':
		self._kind = 'xlsx'
		else:
		raise FormatUnsupportedError(f'File must be of format ".xls" or ".xlsx".')
		@@ -81,7 +84,6 @@ class Converter:
		else:
		self._file = openpyxl.load_workbook(excel_file)


		def _get_columns(self, sheet_name, start=0):
		if self._kind=='xls':
		sheet = self._file.sheet_by_name(sheet_name) # xlrd style
		@@ -376,7 +378,7 @@ class _NewFormatConverter(Converter):
		'type': series_type,
		'stream': stream,
		'units': units,
		'data': [float(d) for d in data if d is not None],
		'data': [float(d) for d in data if d is not None and d != ''],
		}

		if source:

pyehub/network_to_request_format.py

+48 −12

Original line number	Diff line number	Diff line
		@@ -5,6 +5,9 @@ The Excel file has to be in one of the supported formats, which examples can
		be found in the data_formats directory.
		"""
		import xlrd
		import openpyxl
		import pathlib

		from contextlib import suppress


		@@ -18,13 +21,36 @@ def convert(excel_file):
		class Converter:
		def __init__(self, excel_file):
		self._excel_file = excel_file

		# function to return the file extension
		file_extension = pathlib.Path(excel_file).suffix
		if file_extension == '.xls':
		self._kind = 'xls'
		elif file_extension == '.xlsx':
		self._kind = 'xlsx'
		else:
		raise FormatUnsupportedError(f'File must be of format ".xls" or ".xlsx".')

		if self._kind=='xls':
		self._file = xlrd.open_workbook(excel_file)
		else:
		self._file = openpyxl.load_workbook(excel_file)

		def _get_columns(self, sheet_name, start=0):
		sheet = self._file.sheet_by_name(sheet_name)

		if self._kind=='xls':
		sheet = self._file.sheet_by_name(sheet_name) # xlrd style
		for colx in range(start, sheet.ncols):
		yield sheet.col_values(colx)[1:] # slicing the first row
		yield sheet.col_values(colx)[1:] # ipysheet didn't allow editable column names,
		# so added an index row to sheets and slicing it off here
		elif self._kind=='xlsx':
		sheet = self._file[sheet_name] # openpyxl style
		for i,colx in enumerate(sheet.columns): # Different writing with openpyxl
		if i<start: continue; # Skip until start line is reached (generator can't skip earlier)

		column = [c.value if c.value is not None else '' for c in colx]
		if all([c=='' for c in column]): # openpyxl captures rows and rows of useless NaN.
		continue;
		yield column[1:]

		def _get_capacities(self):
		raise NotImplementedError
		@@ -44,8 +70,8 @@ class Converter:


		class NewFormatConverter(Converter):
		def _get_columns(self, sheet_name, start=0):
		return super()._get_columns(sheet_name, start=1)
		def _get_columns(self, sheet_name, start=1):
		return super()._get_columns(sheet_name, start=start)

		def _get_capacities(self):
		capacities = []
		@@ -129,11 +155,21 @@ class NewFormatConverter(Converter):
		return links

		def _get_network(self):
		sheet = self._file.sheet_by_name('Network')

		if self._kind=='xls':
		sheet = self._file.sheet_by_name('Network') # xlrd style
		return{
		'fixed_network_investment_cost': float(sheet.cell(1, 1).value),
		'link_proportional_cost': float(sheet.cell(2, 1).value),
		} # shifting the sheets one row below required to change the row indices here
		else:
		sheet = self._file["Network"] # openpyxl style
		return{
		'fixed_network_investment_cost': float(sheet.cell(1, 1).value),
		'link_proportional_cost': float(sheet.cell(2, 1).value), } # shifting the sheets one row below required to change the row indices here
		'link_proportional_cost': float(sheet.cell(2, 1).value),
		} # shifting the sheets one row below required to change the row indices here
		# sheet = self._file.get_sheet_by_name('Network')