mirror/archivebox
1
0
Fork 0
mirror of https://github.com/pirate/ArchiveBox.git synced 2025-08-17 20:01:44 +02:00
Code Issues Releases Wiki Activity

add more readability error output

Browse Source
This commit is contained in:
Nick Sweeting
2023-10-20 04:14:28 -07:00
parent f3e81c1487
commit 82d8662c74
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
archivebox/extractors/readability.py
View File

@@ -71,7 +71,7 @@ def save_readability(link: Link, out_dir: Optional[str]=None, timeout: int=TIMEO
result = run(cmd, cwd=out_dir, timeout=timeout) result = run(cmd, cwd=out_dir, timeout=timeout)
try: try:
result_json = json.loads(result.stdout) result_json = json.loads(result.stdout)
assert result_json and 'content' in result_json assert result_json and 'content' in result_json, 'Readability output is not valid JSON'
except json.JSONDecodeError: except json.JSONDecodeError:
raise ArchiveError('Readability was not able to archive the page', result.stdout + result.stderr) raise ArchiveError('Readability was not able to archive the page', result.stdout + result.stderr)
@@ -85,7 +85,7 @@ def save_readability(link: Link, out_dir: Optional[str]=None, timeout: int=TIMEO
# "Downloaded: 76 files, 4.0M in 1.6s (2.52 MB/s)" # "Downloaded: 76 files, 4.0M in 1.6s (2.52 MB/s)"
output_tail = [ output_tail = [
line.strip() line.strip()
for line in (result.stdout + result.stderr).decode().rsplit('\n', 3)[-3:] for line in (result.stdout + result.stderr).decode().rsplit('\n', 5)[-5:]
if line.strip() if line.strip()
] ]
hints = ( hints = (