mirrors/ArchiveBox
1
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox.git synced 2025-05-13 14:44:29 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

search for node dependencies in output dir first

Browse source
This commit is contained in:
Nick Sweeting 2020-08-18 18:40:19 -04:00
parent 4566904db2
commit 7144e0bdce
3 changed files with 5 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

2
archivebox/config/__init__.py
View file

@ -780,7 +780,7 @@ globals().update(CONFIG)
os.environ["TZ"] = 'UTC' os.environ["TZ"] = 'UTC'
# add ./node_modules/.bin to $PATH so we can use node scripts in extractors # add ./node_modules/.bin to $PATH so we can use node scripts in extractors
NODE_BIN_PATH = str((Path(CONFIG["OUTPUT_DIR"]) / 'node_modules' / '.bin').resolve()) NODE_BIN_PATH = str((Path(CONFIG["OUTPUT_DIR"]).absolute() / 'node_modules' / '.bin'))
sys.path.append(NODE_BIN_PATH) sys.path.append(NODE_BIN_PATH)

4
archivebox/extractors/readability.py
View file

@ -17,7 +17,7 @@ from ..util import (
from ..config import ( from ..config import (
TIMEOUT, TIMEOUT,
SAVE_READABILITY, SAVE_READABILITY,
READABILITY_BINARY, DEPENDENCIES,
READABILITY_VERSION, READABILITY_VERSION,
) )
from ..logging_util import TimedProgress from ..logging_util import TimedProgress
@ -73,7 +73,7 @@ def save_readability(link: Link, out_dir: Optional[str]=None, timeout: int=TIMEO
temp_doc.close() temp_doc.close()
cmd = [ cmd = [
READABILITY_BINARY, DEPENDENCIES['READABILITY_BINARY']['path'],
temp_doc.name temp_doc.name
] ]

4
archivebox/extractors/singlefile.py
View file

@ -15,7 +15,7 @@ from ..util import (
from ..config import ( from ..config import (
TIMEOUT, TIMEOUT,
SAVE_SINGLEFILE, SAVE_SINGLEFILE,
SINGLEFILE_BINARY, DEPENDENCIES,
SINGLEFILE_VERSION, SINGLEFILE_VERSION,
CHROME_BINARY, CHROME_BINARY,
) )
@ -43,7 +43,7 @@ def save_singlefile(link: Link, out_dir: Optional[str]=None, timeout: int=TIMEOU
# SingleFile CLI Docs: https://github.com/gildas-lormeau/SingleFile/tree/master/cli # SingleFile CLI Docs: https://github.com/gildas-lormeau/SingleFile/tree/master/cli
cmd = [ cmd = [
SINGLEFILE_BINARY, DEPENDENCIES['SINGLEFILE_BINARY']['path'],
'--browser-executable-path={}'.format(CHROME_BINARY), '--browser-executable-path={}'.format(CHROME_BINARY),
'--browser-args="{}"'.format(json.dumps(browser_args[1:])), '--browser-args="{}"'.format(json.dumps(browser_args[1:])),
link.url, link.url,