Salary PDF file content extraction for salary net amount and month

This commit is contained in:
Karl Voit 2018-05-31 12:12:02 +02:00
parent 51c06d475a
commit 83f62df07d

View file

@ -1,6 +1,6 @@
#!/usr/bin/env python3 #!/usr/bin/env python3
# -*- coding: utf-8 -*- # -*- coding: utf-8 -*-
PROG_VERSION = u"Time-stamp: <2018-05-21 13:56:48 vk>" PROG_VERSION = u"Time-stamp: <2018-05-31 12:11:00 vk>"
# TODO: # TODO:
@ -506,16 +506,6 @@ class GuessFilename(object):
logging.debug("derive_new_filename_from_old_filename called") logging.debug("derive_new_filename_from_old_filename called")
datetimestr, basefilename, tags, extension = self.split_filename_entities(oldfilename) datetimestr, basefilename, tags, extension = self.split_filename_entities(oldfilename)
# Paycheck
if extension == "PDF" and self.config.SALARY_STARTSTRING and self.config.SALARY_STARTSTRING in oldfilename:
year, month, day = re.match(self.DATESTAMP_REGEX, datetimestr).groups()
month = int(month)
if int(day) < 15:
# salary came after the new month has started; salary is from previous month
month = month - 1
print(' ' * 7 + colorama.Style.DIM + '→ PDF file password: ' + self.config.SALARY_PDF_PASSWORD + colorama.Style.RESET_ALL)
return datetimestr + ' ' + self.config.SALARY_DESCRIPTION + ' ' + self.NumToMonth(month) + ' - € -- detego private.pdf'
# Android screenshots: # Android screenshots:
# Screenshot_2013-03-05-08-14-09.png -> 2013-03-05T08.14.09 -- android screenshots.png # Screenshot_2013-03-05-08-14-09.png -> 2013-03-05T08.14.09 -- android screenshots.png
regex_match = re.match(self.ANDROID_SCREENSHOT_REGEX, oldfilename) regex_match = re.match(self.ANDROID_SCREENSHOT_REGEX, oldfilename)
@ -744,6 +734,19 @@ class GuessFilename(object):
try: try:
pdffile = PyPDF2.PdfFileReader(open(filename, "rb")) pdffile = PyPDF2.PdfFileReader(open(filename, "rb"))
# if PDF is encryped, try password stored in config file
# or quit this function if decryption is not successful
if pdffile.isEncrypted:
returncode = pdffile.decrypt(self.config.SALARY_PDF_PASSWORD)
if returncode < 1:
logging.error('PDF file is encrypted and could NOT be decrypted using ' +
'config.SALARY_PDF_PASSWORD. Skipping content analysis.')
return False
else:
logging.debug('PDF file is encrypted and could be decrypted using ' +
'config.SALARY_PDF_PASSWORD. Return code = ' + str(returncode))
# use first and second page of content only: # use first and second page of content only:
if pdffile.getNumPages() > 1: if pdffile.getNumPages() > 1:
content = pdffile.pages[0].extractText() + pdffile.pages[1].extractText() content = pdffile.pages[0].extractText() + pdffile.pages[1].extractText()
@ -760,6 +763,36 @@ class GuessFilename(object):
logging.warning('Could read PDF file content but it is empty (skipping content analysis)') logging.warning('Could read PDF file content but it is empty (skipping content analysis)')
return False return False
# Salary - NOTE: this is highly specific to the PDF file
# structure of the author's salary processing software.
# Therefore, this most likely does not work for your salary
# PDF file.
if extension == "PDF" and self.config.SALARY_STARTSTRING and self.config.SALARY_STARTSTRING in filename:
content = content.replace('\n', '') # there is a '\n' after each character
try:
# should parse starting sequence of
# "^.LOHN/GEHALTSABRECHNUNG JÄNNER 2018Klien..." and
# return "Jaenner"
month_of_salary = re.match(r'.LOHN/GEHALTSABRECHNUNG (.+) .+', content).group(1).capitalize().replace('ä', 'ae')
except:
logging.error('derive_new_filename_from_content(' + filename + '): I recognized pattern ' +
'for salary file but content format for extracting month must have changed.')
month_of_salary = 'FIXXME'
try:
# should extract "2.345,67" from following sequence
# ".+SZAbzüge1.234,56Netto2.345,67IBAN:.+"
net_salary = re.match(r'.+Netto(\d\.\d{3},\d{2})IBAN.+', content).group(1)
except:
logging.error('derive_new_filename_from_content(' + filename + '): I recognized pattern ' +
'for salary file but content format for extracting net salary must have changed.')
net_salary = 'FIXXME'
# print out password to stdout in order to give the user a
# hint when he wants to open the PDF in a PDF viewer
print(' ' * 7 + colorama.Style.DIM + '→ PDF file password: ' + self.config.SALARY_PDF_PASSWORD +
colorama.Style.RESET_ALL)
return datetimestr + ' ' + self.config.SALARY_DESCRIPTION + ' ' + month_of_salary + ' - ' + \
net_salary + '€ -- detego private.pdf'
# 2010-06-08 easybank - neue TAN-Liste -- scan private.pdf # 2010-06-08 easybank - neue TAN-Liste -- scan private.pdf
if self.fuzzy_contains_all_of(content, ["Transaktionsnummern (TANs)", "Ihre TAN-Liste in Verlust geraten"]) and \ if self.fuzzy_contains_all_of(content, ["Transaktionsnummern (TANs)", "Ihre TAN-Liste in Verlust geraten"]) and \
datetimestr: datetimestr: