mirror of
https://github.com/ArchiveBox/ArchiveBox.git
synced 2025-05-13 22:54:27 -04:00
feat: WARC link points to the first warc result in target path
This commit is contained in:
parent
b237e412df
commit
e594e6a75a
2 changed files with 10 additions and 4 deletions
|
@ -1,6 +1,7 @@
|
|||
from django.utils.html import format_html
|
||||
|
||||
from core.models import Snapshot, EXTRACTORS
|
||||
from pathlib import Path
|
||||
|
||||
|
||||
def get_icons(snapshot: Snapshot) -> str:
|
||||
|
@ -34,9 +35,10 @@ def get_icons(snapshot: Snapshot) -> str:
|
|||
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["warc_path"],
|
||||
exists, "warc", icons.get("warc", "?"))
|
||||
exists = list((Path(path) / canon["warc_path"]).glob("*.warc.gz"))
|
||||
if exists:
|
||||
output += output_template.format(exists[0], "",
|
||||
True, "warc", icons.get("warc", "?"))
|
||||
|
||||
except Exception as e:
|
||||
print(e)
|
||||
|
|
Loading…
Add table
Add a link
Reference in a new issue