[Docker] Build docker for Parmetis Partition

[jalencato]

Issue #, if available:

Description of changes:

Successfully run on one and multiple instances:

Ways to build docker:

cd ~/graphstorm/docker
bash ./build_docker_parmetis.sh ~/graphstorm/ cyrus_test latest
docker run --network=host --memory="192g" --shm-size 192g -v ~/efs/mount/efs/:/efs -d --name test cyrus_test:latest
cd ~/graphstorm

To run parmetis partition:

python3 graphstorm/python/graphstorm/gpartition/dist_partition_graph.py --input-path /efs/ml-100k-gsp/ \
    --metadata-filename updated_row_counts_metadata.json \
    --output-path /tmp/local-part-test/ --num-parts 4 \
    --dgl-tool-path ~/dgl/tools/ \
    --ip-list ip_list.txt --partition-algorithm parmetis

As our previous discussion in the dev meeting, we decide to put the doc in another link. The command is here for reference about how to run the parmetis partition.

Update the test in the internal regression test pipeline.

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[thvasilo]

A few comments added.

Also, we're missing SageMaker support? As is this will only run on EC2 clusters correct?

[jalencato]

A few comments added.

Also, we're missing SageMaker support? As is this will only run on EC2 clusters correct?

SageMaker support is not in this PR. This PR is only for EC2 clusters(Batch). But eventually we will support SageMaker

[thvasilo]

Can we take the chance and add some documentation?

Can add as an example here https://graphstorm.readthedocs.io/en/latest/scale/distributed.html#process-and-partition-a-graph

[jalencato]

Can we take the chance and add some documentation?

Can add as an example here https://graphstorm.readthedocs.io/en/latest/scale/distributed.html#process-and-partition-a-graph

As we previously discussed, we should add another separate part for all gpartition in the doc. What about let us put it in the next doc refactor PR. I suppose it should involve the docker building + gpartition process for parmetis/random + overview for the whole pipeline. (As the gpartition is closed to the gsprocessing, we should put them together to avoid to let users to be confused.)

And I just take another look at the doc mentioned above, the part there involves both simple processing and partition. If we are trying to put our partition part there, we should at least put the gsprocessing + gpartition command + overview about the distributed processing pipeline there. It will take a lot of space there, so I still suggest to keep our first idea, to put all the staff in the following new doc to make it more easy to understand for user.

[thvasilo]

Let's add a task/issue to track the need for gpartition-specific docs, and parmetis-partition to our integration tests.

A couple of minor changes left, LGTM.

[jalencato]

Task for refactoring and adding doc: #833