to ensure the code maintenance. Thanks to docling
project- took some good bits around python linting
from the project.
These checks are also added in the markdown, so that
user can run these checks locally to recreate the
errors.
Signed-off-by: Anil Vishnoi <vishnoianil@gmail.com>
Group GPU image generate for GHCR and QUAY in one job,
and similarly group CPU only image generation for GHCR
and QUAY in another job. These two jobs can now run
concurrently.
By grouping the jobs for the same image type, it can
use the docker layer cache to run the second job faster.
If we separately build these images in four diffent and
concurrent jobs, they won't be able to use the cache.
If this still takes significantly long, we probably
have to generate image in it's own job, so that they can run
concurrently, without leveraging cache of docker layers.
Also added make file target to lint the github workflows.
Signed-off-by: Anil Vishnoi <vishnoianil@gmail.com>
* Publish docling-serve images to ghcr and quay container repositories
This commit:
- Creates container file to build the image for cpu only
and default "with gpu" images
- Adds workflow jobs to publish the images to ghcr.io and quay.io
- Makefile to build these images locally for linux/amd64 platform.
Signed-off-by: Anil Vishnoi <vishnoianil@gmail.com>
* make cpu-only version with poetry
Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
* Rename cpu/gpu image names and also update the supported platforms
Signed-off-by: Anil Vishnoi <vishnoianil@gmail.com>
---------
Signed-off-by: Anil Vishnoi <vishnoianil@gmail.com>
Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
Co-authored-by: Michele Dolfi <dol@zurich.ibm.com>
