coffee/uchardet
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/uchardet/uchardet.git synced 2026-02-07 02:09:59 +08:00
Code Issues Packages Projects Releases Wiki Activity

make the logfile usable

Browse Source
This commit is contained in:
Martin T. H. Sandsmark 2022-01-28 21:39:18 +01:00 committed by Jehan
parent 2a04e57c8f
commit 8d15d6b557
1 changed files with 3 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
script/BuildLangModel.py
View File

@ -255,6 +255,7 @@ def visit_pages(titles, depth, lang, logfd):
print("Discarding page {}.\n".format(title))
continue
logfd.write("\n{} (revision {})".format(title, page.revision_id))
logfd.flush()
process_text(page.content, lang)
try:
@ -277,6 +278,7 @@ logfd.write('\n- Maximum depth: {}'.format(options.max_depth))
if options.max_page is not None:
logfd.write('\n- Max number of pages: {}'.format(options.max_page))
logfd.write('\n\n== Parsed pages ==\n')
logfd.flush()
try:
visit_pages(lang.start_pages, 0, lang, logfd)
except requests.exceptions.ConnectionError:
@ -284,7 +286,7 @@ except requests.exceptions.ConnectionError:
exit(1)
logfd.write('\n\n== End of Parsed pages ==')
logfd.write('\n\n- Wikipedia parsing ended at: {}\n'.format(str(datetime.datetime.now())))
logfd.flush()
########### CHARACTERS ###########
# Character ratios.