mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-22 10:55:18 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

more detailed parsing and indexing cli output

Browse source
This commit is contained in:
Nick Sweeting 2019-03-27 15:16:53 -04:00
parent b9839500b2
commit 88721512d4
2 changed files with 13 additions and 12 deletions
Download patch file Download diff file Expand all files Collapse all files

5
archivebox/index.py
View file

@ -76,10 +76,11 @@ def load_links_index(out_dir: str=OUTPUT_DIR, import_path: Optional[str]=None) -
# merge existing links in out_dir and new links # merge existing links in out_dir and new links
all_links = list(validate_links(existing_links + new_links)) all_links = list(validate_links(existing_links + new_links))
num_new_links = len(all_links) - len(existing_links)
if import_path and parser_name: if import_path and parser_name:
log_parsing_finished(num_new_links, parser_name) num_parsed = len(raw_links)
num_new_links = len(all_links) - len(existing_links)
log_parsing_finished(num_parsed, num_new_links, parser_name)
return all_links, new_links return all_links, new_links

20
archivebox/logs.py
View file

@ -37,12 +37,13 @@ def log_parsing_started(source_file: str):
**ANSI, **ANSI,
)) ))
def log_parsing_finished(num_new_links: int, parser_name: str): def log_parsing_finished(num_parsed: int, num_new_links: int, parser_name: str):
end_ts = datetime.now() end_ts = datetime.now()
_LAST_RUN_STATS.parse_end_ts = end_ts _LAST_RUN_STATS.parse_end_ts = end_ts
print(' > Adding {} new links to index (parsed import as {})'.format( print(' > Parsed {} links as {}'.format(num_parsed, parser_name))
print(' > Adding {} new links to collection: {}'.format(
num_new_links, num_new_links,
parser_name, OUTPUT_DIR,
)) ))
@ -95,12 +96,10 @@ def log_archiving_paused(num_links: int, idx: int, timestamp: str):
timestamp=timestamp, timestamp=timestamp,
total=num_links, total=num_links,
)) ))
print(' To view your archive, open: {}/index.html'.format(OUTPUT_DIR.replace(REPO_DIR + '/', ''))) print(' To view your archive, open:')
print(' Continue where you left off by running:') print(' {}/index.html'.format(OUTPUT_DIR))
print(' {} {}'.format( print(' Continue archiving where you left off by running:')
pretty_path(sys.argv[0]), print(' archivebox {}'.format(timestamp))
timestamp,
))
def log_archiving_finished(num_links: int): def log_archiving_finished(num_links: int):
end_ts = datetime.now() end_ts = datetime.now()
@ -121,7 +120,8 @@ def log_archiving_finished(num_links: int):
print(' - {} links skipped'.format(_LAST_RUN_STATS.skipped)) print(' - {} links skipped'.format(_LAST_RUN_STATS.skipped))
print(' - {} links updated'.format(_LAST_RUN_STATS.succeeded)) print(' - {} links updated'.format(_LAST_RUN_STATS.succeeded))
print(' - {} links had errors'.format(_LAST_RUN_STATS.failed)) print(' - {} links had errors'.format(_LAST_RUN_STATS.failed))
print(' To view your archive, open: {}/index.html'.format(OUTPUT_DIR.replace(REPO_DIR + '/', ''))) print(' To view your archive, open:')
print(' {}/index.html'.format(OUTPUT_DIR))
def log_link_archiving_started(link_dir: str, link: Link, is_new: bool): def log_link_archiving_started(link_dir: str, link: Link, is_new: bool):