def get_history(self, currency):
self.MONTHS = self.FR_MONTHS if currency == 'EUR' else self.US_MONTHS
#checking if the card is still valid
if self.doc.xpath('//div[@id="errorbox"]'):
return
#adding a time delta because amex have hard time to put the date in a good interval
beginning_date = self.get_beginning_debit_date() - datetime.timedelta(
days=360)
end_date = self.get_end_debit_date()
guesser = ChaoticDateGuesser(beginning_date, end_date)
for tr in reversed(
self.doc.xpath(
'//div[@id="txnsSection"]//tr[@class="tableStandardText"]')
):
cols = tr.findall('td')
t = Transaction()
day, month = CleanText().filter(cols[self.COL_DATE]).split(' ', 1)
day = int(day)
month = self.MONTHS.index(month.rstrip('.')) + 1
date = guesser.guess_date(day, month)
rdate = None
try:
detail = cols[self.COL_TEXT].xpath(
'./div[has-class("hiddenROC")]')[0]
except IndexError:
pass
else:
m = re.search(r' (\d{2} \D{3,4})', (' '.join(
[txt.strip() for txt in detail.itertext()])).strip())
if m:
rday, rmonth = m.group(1).strip().split(' ')
rday = int(rday)
rmonth = self.MONTHS.index(rmonth.rstrip('.')) + 1
rdate = guesser.guess_date(rday, rmonth)
detail.drop_tree()
raw = (' '.join([
txt.strip() for txt in cols[self.COL_TEXT].itertext()
])).strip()
credit = CleanText().filter(cols[self.COL_CREDIT])
debit = CleanText().filter(cols[self.COL_DEBIT])
t.date = date
t.rdate = rdate or date
t.raw = re.sub(r'[ ]+', ' ', raw)
t.label = re.sub('(.*?)( \d+)? .*', r'\1', raw).strip()
t.amount = CleanDecimal(replace_dots=currency == 'EUR').filter(
credit or debit) * (1 if credit else -1)
if t.amount > 0:
t.type = t.TYPE_ORDER
else:
t.type = t.TYPE_CARD
yield t
def get_history(self, date_guesser):
seen = set()
lines = self.document.xpath('(//table[@class="ca-table"])[2]/tr')
for line in lines[1:]: # first line is balance
is_balance = line.xpath('./td/@class="cel-texte cel-neg"')
[date, label, _, amount
] = [self.parser.tocleanstring(td) for td in line.xpath('./td')]
t = Transaction(0)
t.set_amount(amount)
t.label = t.raw = label
if is_balance:
m = re.search('(\d+ [^ ]+ \d+)', label)
if not m:
raise BrokenPageError(
'Unable to read card balance in history: %r' % label)
t.date = parse_french_date(m.group(1))
t.amount = -t.amount
else:
day, month = map(int, date.split('/', 1))
t.date = date_guesser.guess_date(day, month)
t.type = t.TYPE_CARD
t.rdate = t.date
try:
t.id = t.unique_id(seen)
except UnicodeEncodeError:
print t
print t.label
raise
yield t
def get_history(self, date_guesser):
seen = set()
lines = self.document.xpath('(//table[@class="ca-table"])[2]/tr')
for line in lines[1:]: # first line is balance
is_balance = line.xpath('./td/@class="cel-texte cel-neg"')
[date, label, _, amount] = [self.parser.tocleanstring(td)
for td in line.xpath('./td')]
t = Transaction(0)
t.set_amount(amount)
t.label = t.raw = label
if is_balance:
m = re.search('(\d+ [^ ]+ \d+)', label)
if not m:
raise BrokenPageError('Unable to read card balance in history: %r' % label)
t.date = parse_french_date(m.group(1))
t.amount = -t.amount
else:
day, month = map(int, date.split('/', 1))
t.date = date_guesser.guess_date(day, month)
t.type = t.TYPE_CARD
t.rdate = t.date
try:
t.id = t.unique_id(seen)
except UnicodeEncodeError:
print t
print t.label
raise
yield t
def get_card_transactions(self, latest_date, ongoing_coming):
for item in self.doc.xpath('//table[@class="ca-table"][2]//tr[td]'):
if CleanText('./td[2]/b')(item):
# This node is a summary containing the 'date' for all following transactions.
raw_date = Regexp(CleanText('./td[2]/b/text()'), r'le (.*) :')(item)
if latest_date and parse_french_date(raw_date).date() > latest_date:
# This summary has already been fetched
continue
latest_date = parse_french_date(raw_date).date()
if latest_date < ongoing_coming:
# This summary is anterior to the ongoing_coming so we create a transaction from it
tr = FrenchTransaction()
tr.date = tr.rdate = latest_date
tr.raw = tr.label = CleanText('./td[2]/b/text()')(item)
tr.amount = -CleanDecimal.French('./td[position()=last()]')(item)
tr.type = FrenchTransaction.TYPE_CARD_SUMMARY
yield tr
else:
# This node is a real transaction.
# Its 'date' is the date of the most recently encountered summary node.
tr = FrenchTransaction()
tr.date = latest_date
date_guesser = LinearDateGuesser(latest_date)
tr.rdate = DateGuesser(CleanText('./td[1]//text()'), date_guesser=date_guesser)(item)
tr.label = tr.raw = CleanText('./td[2]')(item)
tr.amount = CleanDecimal.French('./td[last()]')(item)
tr.type = FrenchTransaction.TYPE_DEFERRED_CARD
yield tr
def get_history(self, currency):
self.MONTHS = self.FR_MONTHS if currency == 'EUR' else self.US_MONTHS
#checking if the card is still valid
if self.doc.xpath('//div[@id="errorbox"]'):
return
#adding a time delta because amex have hard time to put the date in a good interval
beginning_date = self.get_beginning_debit_date() - datetime.timedelta(days=300)
end_date = self.get_end_debit_date()
guesser = ChaoticDateGuesser(beginning_date, end_date)
for tr in reversed(self.doc.xpath('//div[@id="txnsSection"]//tr[@class="tableStandardText"]')):
cols = tr.findall('td')
t = Transaction()
day, month = CleanText().filter(cols[self.COL_DATE]).split(' ', 1)
day = int(day)
month = self.MONTHS.index(month.rstrip('.')) + 1
date = guesser.guess_date(day, month)
rdate = None
try:
detail = cols[self.COL_TEXT].xpath('./div[has-class("hiddenROC")]')[0]
except IndexError:
pass
else:
m = re.search(r' (\d{2} \D{3,4})', (' '.join([txt.strip() for txt in detail.itertext()])).strip())
if m:
rday, rmonth = m.group(1).split(' ')
rday = int(rday)
rmonth = self.MONTHS.index(rmonth.rstrip('.')) + 1
rdate = guesser.guess_date(rday, rmonth)
detail.drop_tree()
raw = (' '.join([txt.strip() for txt in cols[self.COL_TEXT].itertext()])).strip()
credit = CleanText().filter(cols[self.COL_CREDIT])
debit = CleanText().filter(cols[self.COL_DEBIT])
t.date = date
t.rdate = rdate or date
t.raw = re.sub(r'[ ]+', ' ', raw)
t.label = re.sub('(.*?)( \d+)? .*', r'\1', raw).strip()
t.amount = CleanDecimal(replace_dots=currency == 'EUR').filter(credit or debit) * (1 if credit else -1)
if t.amount > 0:
t.type = t.TYPE_ORDER
else:
t.type = t.TYPE_CARD
yield t
def get_history(self):
#checking if the card is still valid
if self.document.xpath('//div[@id="errorbox"]'):
return
#adding a time delta because amex have hard time to put the date in a good interval
beginning_date = self.get_beginning_debit_date() - datetime.timedelta(
days=120)
end_date = self.get_end_debit_date()
guesser = ChaoticDateGuesser(beginning_date, end_date)
for tr in reversed(
self.document.xpath(
'//div[@id="txnsSection"]//tr[@class="tableStandardText"]')
):
cols = tr.findall('td')
t = Transaction(tr.attrib['id'])
day, month = self.parser.tocleanstring(cols[self.COL_DATE]).split(
' ', 1)
day = int(day)
month = self.MONTHS.index(month.rstrip('.')) + 1
date = guesser.guess_date(day, month)
try:
detail = self.parser.select(cols[self.COL_TEXT],
'div.hiddenROC', 1)
except BrokenPageError:
pass
else:
detail.drop_tree()
raw = (' '.join([
txt.strip() for txt in cols[self.COL_TEXT].itertext()
])).strip()
credit = self.parser.tocleanstring(cols[self.COL_CREDIT])
debit = self.parser.tocleanstring(cols[self.COL_DEBIT])
t.date = date
t.rdate = date
t.raw = re.sub(r'[ ]+', ' ', raw)
t.label = re.sub('(.*?)( \d+)? .*', r'\1', raw).strip()
t.set_amount(credit, debit)
if t.amount > 0:
t.type = t.TYPE_ORDER
else:
t.type = t.TYPE_CARD
yield t
def get_history(self, date_guesser, state=None):
seen = set()
lines = self.document.xpath('(//table[@class="ca-table"])[2]/tr')
debit_date = None
for i, line in enumerate(lines):
is_balance = line.xpath('./td/@class="cel-texte cel-neg"')
# It is possible to have three or four columns.
cols = [self.parser.tocleanstring(td) for td in line.xpath('./td')]
date = cols[0]
label = cols[1]
amount = cols[-1]
t = Transaction()
t.set_amount(amount)
t.label = t.raw = label
if is_balance:
m = re.search('(\d+ [^ ]+ \d+)', label)
if not m:
raise BrokenPageError(
'Unable to read card balance in history: %r' % label)
if state is None:
debit_date = parse_french_date(m.group(1))
else:
debit_date = state
# Skip the first line because it is balance
if i == 0:
continue
t.date = t.rdate = debit_date
# Consider the second one as a positive amount to reset balance to 0.
t.amount = -t.amount
state = t.date
else:
day, month = map(int, date.split('/', 1))
t.rdate = date_guesser.guess_date(day, month)
t.date = debit_date
t.type = t.TYPE_CARD
try:
t.id = t.unique_id(seen)
except UnicodeEncodeError:
self.logger.debug(t)
self.logger.debug(t.label)
raise
yield state, t
def get_history(self, date_guesser, state=None):
seen = set()
lines = self.document.xpath('(//table[@class="ca-table"])[2]/tr')
debit_date = None
for i, line in enumerate(lines):
is_balance = line.xpath('./td/@class="cel-texte cel-neg"')
# It is possible to have three or four columns.
cols = [self.parser.tocleanstring(td) for td in line.xpath('./td')]
date = cols[0]
label = cols[1]
amount = cols[-1]
t = Transaction()
t.set_amount(amount)
t.label = t.raw = label
if is_balance:
m = re.search('(\d+ [^ ]+ \d+)', label)
if not m:
raise BrokenPageError('Unable to read card balance in history: %r' % label)
if state is None:
debit_date = parse_french_date(m.group(1))
else:
debit_date = state
# Skip the first line because it is balance
if i == 0:
continue
t.date = t.rdate = debit_date
# Consider the second one as a positive amount to reset balance to 0.
t.amount = -t.amount
state = t.date
else:
day, month = map(int, date.split('/', 1))
t.rdate = date_guesser.guess_date(day, month)
t.date = debit_date
t.type = t.TYPE_CARD
try:
t.id = t.unique_id(seen)
except UnicodeEncodeError:
self.logger.debug(t)
self.logger.debug(t.label)
raise
yield state, t
def get_history(self, guesser):
debit_date = self.get_debit_date()
if debit_date is not None:
guesser.current_date = debit_date
for tr in reversed(
self.document.xpath(
'//div[@id="txnsSection"]//tr[@class="tableStandardText"]')
):
cols = tr.findall('td')
t = Transaction(tr.attrib['id'])
day, month = self.parser.tocleanstring(cols[self.COL_DATE]).split(
' ', 1)
day = int(day)
month = self.MONTHS.index(month.rstrip('.')) + 1
date = guesser.guess_date(day, month)
try:
detail = self.parser.select(cols[self.COL_TEXT],
'div.hiddenROC', 1)
except BrokenPageError:
pass
else:
detail.drop_tree()
raw = (' '.join([
txt.strip() for txt in cols[self.COL_TEXT].itertext()
])).strip()
credit = self.parser.tocleanstring(cols[self.COL_CREDIT])
debit = self.parser.tocleanstring(cols[self.COL_DEBIT])
t.date = date
t.rdate = date
t.raw = re.sub(r'[ ]+', ' ', raw)
t.label = re.sub('(.*?)( \d+)? .*', r'\1', raw).strip()
t.set_amount(credit, debit)
if t.amount > 0:
t.type = t.TYPE_ORDER
else:
t.type = t.TYPE_CARD
yield t
def get_history(self):
#checking if the card is still valid
if self.document.xpath('//div[@id="errorbox"]'):
return
#adding a time delta because amex have hard time to put the date in a good interval
beginning_date = self.get_beginning_debit_date() - datetime.timedelta(days=120)
end_date = self.get_end_debit_date()
guesser = ChaoticDateGuesser(beginning_date, end_date)
for tr in reversed(self.document.xpath('//div[@id="txnsSection"]//tr[@class="tableStandardText"]')):
cols = tr.findall('td')
t = Transaction(tr.attrib['id'])
day, month = self.parser.tocleanstring(cols[self.COL_DATE]).split(' ', 1)
day = int(day)
month = self.MONTHS.index(month.rstrip('.')) + 1
date = guesser.guess_date(day, month)
try:
detail = self.parser.select(cols[self.COL_TEXT], 'div.hiddenROC', 1)
except BrokenPageError:
pass
else:
detail.drop_tree()
raw = (' '.join([txt.strip() for txt in cols[self.COL_TEXT].itertext()])).strip()
credit = self.parser.tocleanstring(cols[self.COL_CREDIT])
debit = self.parser.tocleanstring(cols[self.COL_DEBIT])
t.date = date
t.rdate = date
t.raw = re.sub(r'[ ]+', ' ', raw)
t.label = re.sub('(.*?)( \d+)? .*', r'\1', raw).strip()
t.set_amount(credit, debit)
if t.amount > 0:
t.type = t.TYPE_ORDER
else:
t.type = t.TYPE_CARD
yield t
def get_history(self, guesser):
debit_date = self.get_debit_date()
if debit_date is not None:
guesser.current_date = debit_date
for tr in reversed(self.document.xpath('//div[@id="txnsSection"]//tr[@class="tableStandardText"]')):
cols = tr.findall("td")
t = Transaction(tr.attrib["id"])
day, month = self.parser.tocleanstring(cols[self.COL_DATE]).split(" ", 1)
day = int(day)
month = self.MONTHS.index(month.rstrip(".")) + 1
date = guesser.guess_date(day, month)
try:
detail = self.parser.select(cols[self.COL_TEXT], "div.hiddenROC", 1)
except BrokenPageError:
pass
else:
detail.drop_tree()
raw = (" ".join([txt.strip() for txt in cols[self.COL_TEXT].itertext()])).strip()
credit = self.parser.tocleanstring(cols[self.COL_CREDIT])
debit = self.parser.tocleanstring(cols[self.COL_DEBIT])
t.date = date
t.rdate = date
t.raw = re.sub(r"[ ]+", " ", raw)
t.label = re.sub("(.*?)( \d+)? .*", r"\1", raw).strip()
t.set_amount(credit, debit)
if t.amount > 0:
t.type = t.TYPE_ORDER
else:
t.type = t.TYPE_CARD
yield t
def get_history(self, date_guesser):
i = 0
for tr in self.document.xpath('//table[@class="ca-table"]//tr'):
parent = tr.getparent()
while parent is not None and parent.tag != 'table':
parent = parent.getparent()
if parent.attrib.get('class', '') != 'ca-table':
continue
if tr.attrib.get('class', '') == 'tr-thead':
heads = tr.findall('th')
for i, head in enumerate(heads):
key = self.parser.tocleanstring(head)
if key == u'Crédit':
self.COL_CREDIT = i - len(heads)
elif key == u'Débit':
self.COL_DEBIT = i - len(heads)
elif key == u'Libellé':
self.COL_TEXT = i
if not tr.attrib.get('class', '').startswith('ligne-'):
continue
cols = tr.findall('td')
# On loan accounts, there is a ca-table with a summary. Skip it.
if tr.find('th') is not None or len(cols) < 3:
continue
t = Transaction(i)
col_text = cols[self.COL_TEXT]
if len(col_text.xpath('.//br')) == 0:
col_text = cols[self.COL_TEXT + 1]
raw = self.parser.tocleanstring(col_text)
date = self.parser.tocleanstring(cols[self.COL_DATE])
credit = self.parser.tocleanstring(cols[self.COL_CREDIT])
if self.COL_DEBIT is not None:
debit = self.parser.tocleanstring(cols[self.COL_DEBIT])
else:
debit = ''
day, month = map(int, date.split('/', 1))
t.date = date_guesser.guess_date(day, month)
t.rdate = t.date
t.raw = raw
# On some accounts' history page, there is a <font> tag in columns.
if col_text.find('font') is not None:
col_text = col_text.find('font')
t.category = unicode(col_text.text.strip())
t.label = re.sub('(.*) (.*)', r'\2', t.category).strip()
sub_label = col_text.find('br').tail
if sub_label is not None and (
len(t.label) < 3 or t.label == t.category
or len(re.findall('[^\w\s]', sub_label)) /
float(len(sub_label)) <
len(re.findall('\d', t.label)) / float(len(t.label))):
t.label = unicode(sub_label.strip())
# Sometimes, the category contains the label, even if there is another line with it again.
t.category = re.sub('(.*) .*', r'\1', t.category).strip()
t.type = self.TYPES.get(t.category, t.TYPE_UNKNOWN)
# Parse operation date in label (for card transactions for example)
m = re.match('(?P<text>.*) (?P<dd>[0-3]\d)/(?P<mm>[0-1]\d)$',
t.label)
if not m:
m = re.match('^(?P<dd>[0-3]\d)/(?P<mm>[0-1]\d) (?P<text>.*)$',
t.label)
if m:
if t.type in (t.TYPE_CARD, t.TYPE_WITHDRAWAL):
t.rdate = date_guesser.guess_date(
int(m.groupdict()['dd']),
int(m.groupdict()['mm']),
change_current_date=False)
t.label = m.groupdict()['text'].strip()
# Strip city or other useless information from label.
t.label = re.sub('(.*) .*', r'\1', t.label).strip()
t.set_amount(credit, debit)
yield t
i += 1
def get_history(self, date_guesser):
for tr in self.document.xpath('//table[@class="ca-table"]//tr'):
parent = tr.getparent()
while parent is not None and parent.tag != 'table':
parent = parent.getparent()
if parent.attrib.get('class', '') != 'ca-table':
continue
if tr.attrib.get('class', '') == 'tr-thead':
heads = tr.findall('th')
for i, head in enumerate(heads):
key = self.parser.tocleanstring(head)
if key == u'Débit':
self.COL_DEBIT = i - len(heads)
if key == u'Crédit':
self.COL_CREDIT = i - len(heads)
if key == u'Libellé':
self.COL_TEXT = i
if not tr.attrib.get('class', '').startswith('ligne-'):
continue
cols = tr.findall('td')
# On loan accounts, there is a ca-table with a summary. Skip it.
if tr.find('th') is not None or len(cols) < 3:
continue
t = Transaction()
col_text = cols[self.COL_TEXT]
if len(col_text.xpath('.//br')) == 0:
col_text = cols[self.COL_TEXT+1]
raw = self.parser.tocleanstring(col_text)
date = self.parser.tocleanstring(cols[self.COL_DATE])
credit = self.parser.tocleanstring(cols[self.COL_CREDIT])
if self.COL_DEBIT is not None:
debit = self.parser.tocleanstring(cols[self.COL_DEBIT])
else:
debit = ''
day, month = map(int, date.split('/', 1))
t.date = date_guesser.guess_date(day, month)
t.rdate = t.date
t.raw = raw
# On some accounts' history page, there is a <font> tag in columns.
if col_text.find('font') is not None:
col_text = col_text.find('font')
t.category = unicode(col_text.text.strip())
t.label = re.sub('(.*) (.*)', r'\2', t.category).strip()
sub_label = col_text.find('br').tail
if sub_label is not None and (len(t.label) < 3 or t.label == t.category or len(re.findall('[^\w\s]', sub_label))/float(len(sub_label)) < len(re.findall('\d', t.label))/float(len(t.label))):
t.label = unicode(sub_label.strip())
# Sometimes, the category contains the label, even if there is another line with it again.
t.category = re.sub('(.*) .*', r'\1', t.category).strip()
t.type = self.TYPES.get(t.category, t.TYPE_UNKNOWN)
# Parse operation date in label (for card transactions for example)
m = re.match('(?P<text>.*) (?P<dd>[0-3]\d)/(?P<mm>[0-1]\d)$', t.label)
if not m:
m = re.match('^(?P<dd>[0-3]\d)/(?P<mm>[0-1]\d) (?P<text>.*)$', t.label)
if m:
if t.type in (t.TYPE_CARD, t.TYPE_WITHDRAWAL):
t.rdate = date_guesser.guess_date(int(m.groupdict()['dd']), int(m.groupdict()['mm']), change_current_date=False)
t.label = m.groupdict()['text'].strip()
# Strip city or other useless information from label.
t.label = re.sub('(.*) .*', r'\1', t.label).strip()
t.set_amount(credit, debit)
yield t
