Mirrors/ArchiveBox
2
0
Fork 0
mirror of https://github.com/ArchiveBox/ArchiveBox synced 2024-11-10 06:34:16 +00:00
Code Issues Projects Releases Packages Wiki Activity

feat: Update dockerfile to support readability

Browse source
This commit is contained in:
Cristian 2020-08-11 11:52:43 -05:00
parent 2a68af1b94
commit 5dc7e63792
2 changed files with 8 additions and 3 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

8
Dockerfile
View file

@ -57,6 +57,11 @@ RUN wget -qO - https://github.com/gildas-lormeau/SingleFile/archive/master.zip >
&& npm install --prefix SingleFile-master/cli --production > /dev/null 2>&1 \ && npm install --prefix SingleFile-master/cli --production > /dev/null 2>&1 \
&& chmod +x SingleFile-master/cli/single-file && chmod +x SingleFile-master/cli/single-file
RUN wget -qO - https://github.com/pirate/readability-extractor/archive/master.zip > readability.zip \
&& unzip -q readability.zip \
&& npm install --prefix readability-extractor-master --production > /dev/null 2>&1 \
&& chmod +x readability-extractor-master/readability-extractor
# Run everything from here on out as non-privileged user # Run everything from here on out as non-privileged user
RUN groupadd --system archivebox \ RUN groupadd --system archivebox \
&& useradd --system --create-home --gid archivebox --groups audio,video archivebox && useradd --system --create-home --gid archivebox --groups audio,video archivebox
@ -74,7 +79,8 @@ EXPOSE 8000
ENV IN_DOCKER=True \ ENV IN_DOCKER=True \
CHROME_BINARY=google-chrome \ CHROME_BINARY=google-chrome \
CHROME_SANDBOX=False \ CHROME_SANDBOX=False \
SINGLEFILE_BINARY="$EXTRA_PATH/SingleFile-master/cli/single-file" SINGLEFILE_BINARY="$EXTRA_PATH/SingleFile-master/cli/single-file" \
READABILITY_BINARY="$EXTRA_PATH/readability-extractor-master/readability-extractor"
RUN env ALLOW_ROOT=True archivebox version RUN env ALLOW_ROOT=True archivebox version

3
archivebox/extractors/readability.py
View file

@ -50,8 +50,7 @@ def should_save_readability(link: Link, out_dir: Optional[str]=None) -> bool:
if is_static_file(link.url): if is_static_file(link.url):
return False return False
output = Path(out_dir or link.link_dir) / 'readability.json' output = Path(out_dir or link.link_dir) / 'readability'
print(output, SAVE_READABILITY)
return SAVE_READABILITY and (not output.exists()) return SAVE_READABILITY and (not output.exists())