mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-13 22:54:27 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

refactor: Unpack extractors tuple instead of using the index to access the relevant information

Browse source
This commit is contained in:
Cristian 2020-11-10 12:38:29 -05:00
parent 71655220ad
commit 508a0bb06e
1 changed files with 12 additions and 9 deletions
Download patch file Download diff file Expand all files Collapse all files

21
archivebox/core/utils.py
View file

@ -25,16 +25,19 @@ def get_icons(snapshot: Snapshot) -> str:
exclude = ["favicon"]
# Missing specific entry for WARC
for extractor in EXTRACTORS:
result = archive_results.filter(extractor=extractor[0], status="succeeded")
for extractor, _ in EXTRACTORS:
result = archive_results.filter(extractor=extractor, status="succeeded")
path, exists = link.archive_path, result.exists()
try:
if extractor[0] not in exclude:
output += output_template.format(link.archive_path, canon[f"{extractor[0]}_path"],
result.exists(), extractor[0], icons.get(extractor[0], "?"))
if extractor[0] == "wget":
extractor = "warc"
output += output_template.format(link.archive_path, canon[f"{extractor}_path"],
result.exists(), extractor, icons.get(extractor, "?"))
if extractor not in exclude:
output += output_template.format(path, canon[f"{extractor}_path"],
exists, extractor, icons.get(extractor, "?"))
if extractor == "wget":
# warc isn't technically it's own extractor, so we have to add it after wget
output += output_template.format(path, canon[f"warc_path"],
exists, "warc", icons.get("warc", "?"))
except Exception as e:
print(e)