class OperatingRevenueAssembler():
def __init__(self):
self.base_xpath = '//html/body/div[@id="SysJustIFRAMEDIV"]/table/tr/td/form/table/tr/td/table'
self.content_screener = ContentScreener()
self.string_utils = StringUtils()
self.lxml_utils = LxmlUtils()
def assemble(self, param):
content, stock_symbol = param['content'], param['stock_symbol']
self.content_screener.screen(param)
html_object = self.__get_html_object(content)
relative_html_object = self.__traverse_to_relative_html_object(html_object)
column_name_list = self.__assemble_column_name_list(relative_html_object)
row_list = self.__assemble_row_list(relative_html_object)
return OperatingRevenueDao(column_name_list, row_list, stock_symbol, 'M')
def __get_html_object(self, content):
content = self.string_utils.normalize_string(content)
content = content.replace(u'<br>', u'')
return lxml.html.fromstring(content)
def __traverse_to_relative_html_object(self, html_object):
relative_html_object_list = html_object.xpath(self.base_xpath)
assert len(relative_html_object_list) > 0, 'invalid base_xpath (table_tags)'
return relative_html_object_list[0]
def __assemble_column_name_list(self, relative_html_object):
# traverse and sanity check
tr_tags = relative_html_object.xpath('./tr')
assert len(tr_tags) > 5, 'invalid tr_tags'
td_texts = tr_tags[5].xpath('./td/text()')
assert len(td_texts) == 7, 'invalid td_texts size, should be 7'
return [text.strip() for text in td_texts]
def __assemble_row_list(self, relative_html_object):
tr_tags = relative_html_object.xpath('./tr')
assert len(tr_tags) > 5, 'invalid tr_tags'
return [self.__assemble_row(tr_tag) for tr_tag in tr_tags[6:]]
def __assemble_row(self, relative_html_object):
td_tags = relative_html_object.xpath('./td')
td_texts = self.lxml_utils.get_text_list(td_tags)
assert len(td_texts) == 7, 'invalid td_texts size, should be 7'
row = []
# should be stmt_date
stmt_date = self.string_utils.from_local_string_to_date(td_texts[0])
row.append(stmt_date)
# should be number
for number_string in td_texts[1:]:
number = self.string_utils.normalize_number(number_string)
row.append(number)
return row
def __init__(self):
self.base_xpath = '//html/body/div[@id="SysJustIFRAMEDIV"]/table/tr/td/form/table/tr/td/table'
self.content_screener = ContentScreener()
self.string_utils = StringUtils()
self.lxml_utils = LxmlUtils()
def __init__(self):
self.base_xpath = '//html/body/div[@id="SysJustIFRAMEDIV"]/table/tr/td/form/table/tr/td/table'
self.content_screener = ContentScreener()
self.string_utils = StringUtils()
self.lxml_utils = LxmlUtils()
class FinancialStatementAssembler():
def __init__(self, base_xpath_param, column_name_pos=0):
self.base_xpath = self.__init_base_xpath(base_xpath_param)
self.column_name_pos = column_name_pos
self.content_screener = ContentScreener()
self.string_utils = StringUtils()
self.lxml_utils = LxmlUtils()
def __init_base_xpath(self, base_xpath_param):
if base_xpath_param == 'basic':
return '//html/body/div[@id="SysJustIFRAMEDIV"]/table/tr/td/table/tr/td/table'
elif base_xpath_param == 'form':
return '//html/body/div[@id="SysJustIFRAMEDIV"]/table/tr/td/form/table/tr/td/table'
def assemble(self, param):
content, stock_symbol = param['content'], param['stock_symbol']
self.content_screener.screen(param)
html_object = self.__get_html_object(content)
relative_html_object = self.__traverse_to_relative_html_object(
html_object)
column_name_list = self.__assemble_column_name_list(
relative_html_object)
row_list = self.__assemble_row_list(relative_html_object)
return {
'stock_symbol': stock_symbol,
'column_name_list': column_name_list,
'row_list': row_list,
}
def __get_html_object(self, content):
content = self.string_utils.normalize_string(content)
content = content.replace(u'<br>', u'')
return lxml.html.fromstring(content)
def __traverse_to_relative_html_object(self, html_object):
relative_html_object_list = html_object.xpath(self.base_xpath)
assert len(
relative_html_object_list) > 0, 'invalid base_xpath (table_tags)'
return relative_html_object_list[0]
def __assemble_column_name_list(self, relative_html_object):
# traverse and sanity check
tr_tags = relative_html_object.xpath('./tr')
assert len(tr_tags) > self.column_name_pos, 'invalid tr_tags'
td_texts = tr_tags[self.column_name_pos].xpath('./td/text()')
# the first entry should be account
column_name_list = [td_texts[0]]
# the rest should be stmt_date
for text in td_texts[1:]:
stmt_date = self.string_utils.from_local_string_to_date(text)
column_name_list.append(stmt_date)
return column_name_list
def __assemble_row_list(self, relative_html_object):
tr_tags = relative_html_object.xpath('./tr')
assert len(tr_tags) > self.column_name_pos + 1, 'invalid tr_tags'
return [
self.__assemble_row(tr_tag)
for tr_tag in tr_tags[self.column_name_pos + 1:]
]
def __assemble_row(self, relative_html_object):
td_tags = relative_html_object.xpath('./td')
td_texts = self.lxml_utils.get_text_list(td_tags)
# return empty row if empty
if not td_texts:
return []
# the first entry should be account
row = [td_texts[0].strip()]
# the rest should be stmt_date
for number_string in td_texts[1:]:
number = self.string_utils.normalize_number(number_string)
row.append(number)
return row
class OperatingRevenueAssembler():
def __init__(self):
self.base_xpath = '//html/body/div[@id="SysJustIFRAMEDIV"]/table/tr/td/form/table/tr/td/table'
self.content_screener = ContentScreener()
self.string_utils = StringUtils()
self.lxml_utils = LxmlUtils()
def assemble(self, param):
content, stock_symbol = param['content'], param['stock_symbol']
self.content_screener.screen(param)
html_object = self.__get_html_object(content)
relative_html_object = self.__traverse_to_relative_html_object(
html_object)
column_name_list = self.__assemble_column_name_list(
relative_html_object)
row_list = self.__assemble_row_list(relative_html_object)
return OperatingRevenueDao(column_name_list, row_list, stock_symbol,
'M')
def __get_html_object(self, content):
content = self.string_utils.normalize_string(content)
content = content.replace(u'<br>', u'')
return lxml.html.fromstring(content)
def __traverse_to_relative_html_object(self, html_object):
relative_html_object_list = html_object.xpath(self.base_xpath)
assert len(
relative_html_object_list) > 0, 'invalid base_xpath (table_tags)'
return relative_html_object_list[0]
def __assemble_column_name_list(self, relative_html_object):
# traverse and sanity check
tr_tags = relative_html_object.xpath('./tr')
assert len(tr_tags) > 5, 'invalid tr_tags'
td_texts = tr_tags[5].xpath('./td/text()')
assert len(td_texts) == 7, 'invalid td_texts size, should be 7'
return [text.strip() for text in td_texts]
def __assemble_row_list(self, relative_html_object):
tr_tags = relative_html_object.xpath('./tr')
assert len(tr_tags) > 5, 'invalid tr_tags'
return [self.__assemble_row(tr_tag) for tr_tag in tr_tags[6:]]
def __assemble_row(self, relative_html_object):
td_tags = relative_html_object.xpath('./td')
td_texts = self.lxml_utils.get_text_list(td_tags)
assert len(td_texts) == 7, 'invalid td_texts size, should be 7'
row = []
# should be stmt_date
stmt_date = self.string_utils.from_local_string_to_date(td_texts[0])
row.append(stmt_date)
# should be number
for number_string in td_texts[1:]:
number = self.string_utils.normalize_number(number_string)
row.append(number)
return row
def __init__(self, base_xpath_param, column_name_pos=0):
self.base_xpath = self.__init_base_xpath(base_xpath_param)
self.column_name_pos = column_name_pos
self.content_screener = ContentScreener()
self.string_utils = StringUtils()
self.lxml_utils = LxmlUtils()
class BalanceSheetSummaryAssembler():
def __init__(self, period):
self.base_xpath = '//html/body/div[@id="SysJustIFRAMEDIV"]/table/tr/td/table/tr/td/table/tr/td/table'
self.content_screener = ContentScreener()
self.string_utils = StringUtils()
self.lxml_utils = LxmlUtils()
self.period = period
def assemble(self, param):
content, stock_symbol = param['content'], param['stock_symbol']
self.content_screener.screen(param)
html_object = self.__get_html_object(content)
relative_html_object = self.__traverse_to_relative_html_object(html_object)
column_name_list = self.__assemble_column_name_list(relative_html_object)
row_list = self.__assemble_row_list(relative_html_object)
short_period = None
if self.period == 'quarterly':
short_period = 'Q'
elif self.period == 'yearly':
short_period = 'Y'
return BalanceSheetSummaryDao(column_name_list, row_list, stock_symbol, short_period)
def __get_html_object(self, content):
content = self.string_utils.normalize_string(content)
content = content.replace(u'<br>', u'')
return lxml.html.fromstring(content)
def __traverse_to_relative_html_object(self, html_object):
relative_html_object_list = html_object.xpath(self.base_xpath)
assert len(relative_html_object_list) > 1, 'invalid base_xpath (table_tags)'
if self.period == 'quarterly':
return relative_html_object_list[0]
elif self.period == 'yearly':
return relative_html_object_list[1]
def __assemble_column_name_list(self, relative_html_object):
# traverse and sanity check
tr_tags = relative_html_object.xpath('./tr')
assert len(tr_tags) > 0, 'invalid tr_tags'
td_texts = tr_tags[0].xpath('./td/text()')
# the first entry should be account
column_name_list = [td_texts[0]]
# the rest should be stmt_date
for text in td_texts[1:]:
stmt_date = self.string_utils.from_local_string_to_date(text)
column_name_list.append(stmt_date)
return column_name_list
def __assemble_row_list(self, relative_html_object):
tr_tags = relative_html_object.xpath('./tr')
assert len(tr_tags) > 1, 'invalid tr_tags'
return [self.__assemble_row(tr_tag) for tr_tag in tr_tags[1:]]
def __assemble_row(self, relative_html_object):
td_tags = relative_html_object.xpath('./td')
td_texts = self.lxml_utils.get_text_list(td_tags)
# the first entry should be account
row = [td_texts[0]]
# the rest should be stmt_date
for number_string in td_texts[1:]:
number = self.string_utils.normalize_number(number_string)
row.append(number)
return row
def __init__(self):
self.base_xpath = '//html/body'
self.string_utils = StringUtils()
self.lxml_utils = LxmlUtils()
class StockSymbolAssembler():
def __init__(self):
self.base_xpath = '//html/body'
self.string_utils = StringUtils()
self.lxml_utils = LxmlUtils()
def assemble(self, param):
content = self.string_utils.normalize_string(param['content'])
html_object = lxml.html.fromstring(content)
relative_html_object = self.__traverse_to_relative_html_object(html_object)
release_date = self.__assemble_release_date(relative_html_object)
column_name_list = self.__assemble_column_name_list(relative_html_object)
row_list = self.__assemble_row_list(relative_html_object)
return StockSymbolDao(column_name_list, row_list, release_date)
def __traverse_to_relative_html_object(self, html_object):
relative_html_object_list = html_object.xpath(self.base_xpath)
assert len(relative_html_object_list) == 1, 'invalid base_xpath'
return relative_html_object_list[0]
def __assemble_release_date(self, relative_html_object):
# try to get release date
table_tags = relative_html_object.xpath('./table')
assert len(table_tags) > 0, 'invalid table_tags'
headline_tags = table_tags[0].xpath('./h2')
assert len(headline_tags) > 0, 'invalid headline_tags'
headline_texts = headline_tags[1].xpath('./strong/center')
groups = self.string_utils.match(u'^最近更新日期:(.*)$', headline_texts[0].text.strip())
assert len(groups) > 0, 'could not match ^最近更新日期:(.*)$'
release_date = self.string_utils.from_local_string_to_date(groups[0])
return release_date
def __assemble_column_name_list(self, relative_html_object):
# traverse and sanity check
tr_tags = relative_html_object.xpath('./table[@class="h4"]/tr')
assert len(tr_tags) > 0, 'invalid tr_tags'
# traverse and sanity check
original_column_name_list = tr_tags[0].xpath('./td/text()')
# handle the first column name: '有價證券代號及名稱'
combined_column_name = original_column_name_list[0].strip()
assert combined_column_name == u'有價證券代號及名稱', 'should be 有價證券代號及名稱 in unicode'
# the chinese character '及' means 'and' so we need to seperate this column name
seperated_column_name_list = combined_column_name.split(u'及')
assert len(seperated_column_name_list) == 2
column_name_list = seperated_column_name_list + original_column_name_list[1:]
assert len(column_name_list) == 8, 'invalid column_name_list size, should be 8'
return column_name_list
def __assemble_row_list(self, relative_html_object):
# skip one row of column name list
tr_tags = relative_html_object.xpath('./table[@class="h4"]/tr')[1:]
row_list = []
for tr_tag in tr_tags:
row = self.__assemble_row(tr_tag)
# if there is only one cell '股票' in row, skip it
if row:
row_list.append(row)
return row_list
def __assemble_row(self, relative_html_object):
td_tags = relative_html_object.xpath('./td')
td_texts = self.lxml_utils.get_text_list(td_tags)
# if there is only one cell '股票', return None
if len(td_texts) == 1:
return None
# sanity check
assert len(td_texts) == 7
# handle the first cell: '有價證券代號及名稱'
# it should be seperated as stock symbol and stock name
combined_cell = td_texts[0].strip()
seperated_cell_list = combined_cell.split()
assert len(seperated_cell_list) == 2
# convert to datetime.date type
listing_date = self.string_utils.from_local_string_to_date(td_texts[2])
row = seperated_cell_list + [td_texts[1]] + [listing_date] + td_texts[3:]
return row
class BalanceSheetSummaryAssembler():
def __init__(self, period):
self.base_xpath = '//html/body/div[@id="SysJustIFRAMEDIV"]/table/tr/td/table/tr/td/table/tr/td/table'
self.content_screener = ContentScreener()
self.string_utils = StringUtils()
self.lxml_utils = LxmlUtils()
self.period = period
def assemble(self, param):
content, stock_symbol = param['content'], param['stock_symbol']
self.content_screener.screen(param)
html_object = self.__get_html_object(content)
relative_html_object = self.__traverse_to_relative_html_object(
html_object)
column_name_list = self.__assemble_column_name_list(
relative_html_object)
row_list = self.__assemble_row_list(relative_html_object)
short_period = None
if self.period == 'quarterly':
short_period = 'Q'
elif self.period == 'yearly':
short_period = 'Y'
return BalanceSheetSummaryDao(column_name_list, row_list, stock_symbol,
short_period)
def __get_html_object(self, content):
content = self.string_utils.normalize_string(content)
content = content.replace(u'<br>', u'')
return lxml.html.fromstring(content)
def __traverse_to_relative_html_object(self, html_object):
relative_html_object_list = html_object.xpath(self.base_xpath)
assert len(
relative_html_object_list) > 1, 'invalid base_xpath (table_tags)'
if self.period == 'quarterly':
return relative_html_object_list[0]
elif self.period == 'yearly':
return relative_html_object_list[1]
def __assemble_column_name_list(self, relative_html_object):
# traverse and sanity check
tr_tags = relative_html_object.xpath('./tr')
assert len(tr_tags) > 0, 'invalid tr_tags'
td_texts = tr_tags[0].xpath('./td/text()')
# the first entry should be account
column_name_list = [td_texts[0]]
# the rest should be stmt_date
for text in td_texts[1:]:
stmt_date = self.string_utils.from_local_string_to_date(text)
column_name_list.append(stmt_date)
return column_name_list
def __assemble_row_list(self, relative_html_object):
tr_tags = relative_html_object.xpath('./tr')
assert len(tr_tags) > 1, 'invalid tr_tags'
return [self.__assemble_row(tr_tag) for tr_tag in tr_tags[1:]]
def __assemble_row(self, relative_html_object):
td_tags = relative_html_object.xpath('./td')
td_texts = self.lxml_utils.get_text_list(td_tags)
# the first entry should be account
row = [td_texts[0]]
# the rest should be stmt_date
for number_string in td_texts[1:]:
number = self.string_utils.normalize_number(number_string)
row.append(number)
return row
