mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-23 03:06:55 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

config and attr access improvements

Browse source
This commit is contained in:
Nick Sweeting 2024-08-20 18:31:21 -07:00
parent 4ae186dfca
commit 0285aa52a0
No known key found for this signature in database
15 changed files with 203 additions and 187 deletions
Download patch file Download diff file Expand all files Collapse all files

39
archivebox/extractors/favicon.py
View file

@ -11,27 +11,18 @@ from ..util import (
domain,
dedupe,
)
from ..config import (
TIMEOUT,
SAVE_FAVICON,
FAVICON_PROVIDER,
CURL_BINARY,
CURL_ARGS,
CURL_EXTRA_ARGS,
CURL_VERSION,
CHECK_SSL_VALIDITY,
CURL_USER_AGENT,
)
from ..config import CONFIG
from ..logging_util import TimedProgress
@enforce_types
def should_save_favicon(link: Link, out_dir: Optional[str]=None, overwrite: Optional[bool]=False) -> bool:
out_dir = out_dir or Path(link.link_dir)
def should_save_favicon(link: Link, out_dir: str | Path | None=None, overwrite: bool=False) -> bool:
assert link.link_dir
out_dir = Path(out_dir or link.link_dir)
if not overwrite and (out_dir / 'favicon.ico').exists():
return False
return SAVE_FAVICON
return CONFIG.SAVE_FAVICON
@enforce_types
def get_output_path():
@ -39,24 +30,26 @@ def get_output_path():
@enforce_types
def save_favicon(link: Link, out_dir: Optional[Path]=None, timeout: int=TIMEOUT) -> ArchiveResult:
def save_favicon(link: Link, out_dir: str | Path | None=None, timeout: int=CONFIG.TIMEOUT) -> ArchiveResult:
"""download site favicon from google's favicon api"""
out_dir = out_dir or link.link_dir
out_dir = Path(out_dir or link.link_dir)
assert out_dir.exists()
output: ArchiveOutput = 'favicon.ico'
# later options take precedence
options = [
*CURL_ARGS,
*CURL_EXTRA_ARGS,
*CONFIG.CURL_ARGS,
*CONFIG.CURL_EXTRA_ARGS,
'--max-time', str(timeout),
'--output', str(output),
*(['--user-agent', '{}'.format(CURL_USER_AGENT)] if CURL_USER_AGENT else []),
*([] if CHECK_SSL_VALIDITY else ['--insecure']),
*(['--user-agent', '{}'.format(CONFIG.CURL_USER_AGENT)] if CONFIG.CURL_USER_AGENT else []),
*([] if CONFIG.CHECK_SSL_VALIDITY else ['--insecure']),
]
cmd = [
CURL_BINARY,
CONFIG.CURL_BINARY,
*dedupe(options),
FAVICON_PROVIDER.format(domain(link.url)),
CONFIG.FAVICON_PROVIDER.format(domain(link.url)),
]
status = 'failed'
timer = TimedProgress(timeout, prefix=' ')
@ -72,7 +65,7 @@ def save_favicon(link: Link, out_dir: Optional[Path]=None, timeout: int=TIMEOUT)
return ArchiveResult(
cmd=cmd,
pwd=str(out_dir),
cmd_version=CURL_VERSION,
cmd_version=CONFIG.CURL_VERSION,
output=output,
status=status,
**timer.stats,