feat(docker): include pcg installation #854
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## main #854 +/- ##
=========================================
Coverage 100.00% 100.00%
=========================================
Files 142 142
Lines 6031 6034 +3
=========================================
+ Hits 6031 6034 +3 ☔ View full report in Codecov by Sentry. |
akhileshh
changed the title
akhileshh
force-pushed
the
akhilesh/zutils-pcg-docker
branch
from
6ff9fc9 to
6236bfb
Compare
There was a problem hiding this comment.
In case you want to try this approach with the graph tool installation in the zutils image, I would request to cleanup the image from all temporary data, caches and unused apt packages.
@ranlu showed me how to do this for the Pytorch base image: (https://github.com/ZettaAI/zetta_utils/blob/main/docker/pytorch_base/Dockerfile#L122-L140)
| && apt update && apt install git g++ -y \ | ||
| && git clone --depth 1 --branch v2.18.3 https://github.com/CAVEconnectome/PyChunkedGraph.git pcg \ | ||
| && cd pcg \ | ||
| && pip3 install . -r requirements.in |
There was a problem hiding this comment.
either pip install --no-cache-dir or use the --mount=type=cache approach above
|
|
||
|
|
||
| FROM zutils AS zutils-pcg | ||
| RUN conda install graph-tool -c conda-forge -y \ |
There was a problem hiding this comment.
graph-tool-base avoids all the graph display related dependencies
There was a problem hiding this comment.
Cleanup for conda: conda clean -ya
|
Thanks! But I think I like your pr better. I have no idea how long re-building the base image is going to take. Don't have the time to experiment with that now. |
No description provided.