feat: integrate dev branch and update software environments (#25) #29
Conversation
btraven00
commented
May 21, 2025
btraven00
commented
- run from post-0.2.0 tag, main branch
- docs: use public repo URIs
- chore: add convenience target to build environments
- add top-level Makefile to prepare env
- feat: parametrize num of cores on the makefile
- chore: ignore common temporary outputs and image build artifacts
- update .eb files to easybuild 5.0
- remove remote storage
- do not run artifacts if not in main repo
- inject checksums to rmarkdown easyconfig
- update sklearn singularity definition
- feat: add microbenchmark for numpy operations
- chore: bump clustering-benchmarks to 1.1.6
- feat: templatize the definitions
- feat: mv output folders to timestamped names
- feat: add --yes flag
- docs: update README
…hmark#25) * run from post-0.2.0 tag, main branch * docs: use public repo URIs * chore: add convenience target to build environments * add top-level Makefile to prepare env * feat: parametrize num of cores on the makefile * chore: ignore common temporary outputs and image build artifacts * update .eb files to easybuild 5.0 * remove remote storage * do not run artifacts if not in main repo * inject checksums to rmarkdown easyconfig * update sklearn singularity definition * feat: add microbenchmark for numpy operations * chore: bump clustering-benchmarks to 1.1.6 * feat: templatize the definitions * feat: mv output folders to timestamped names * feat: add --yes flag * docs: update README
There was a problem hiding this comment.
this is not a request for changes, just a proactive way to block the merge until I check sometime next week
note to self: red tests; singularity recipes to be made reproducible and planned with a grid; what overrides, ; update/upgrade actions; check how microbench, overrides, smoketests etc are supposed to be used; agree on a registry with shared write rights etc
|
Can I cherrypick the |
|
As mentioned elsewhere, let me know which parts you want split and I'll do a new PR. |
Do you mean you want me to revert this minor version bump? https://github.com/omnibenchmark/clustering_example/pull/29/files#diff-a6cca9b20e49ee9b099e3d6c9b2a7fe6712bb85b90931cf71e2e3160b08fc5eaR188 TBH I didn't fully grasp the argument (what old runs are we discussing here?). Feel free to request the revert, but be aware that it's not guaranteed to build without it (hint: bioconductor versions) 😄 |
There was a problem hiding this comment.
we switch to clustering_[conda|apptainer|envmodules].yaml only, and they're short, all others are out
we test on PR and nightly (short and long), perhaps with timeout
we don't write artifacts
we don't test reduced-imallona branch because it's not relevant anymore
we switch to the new run_omnibenchmark action
There was a problem hiding this comment.
same as above, and replace by the run_ob action
There was a problem hiding this comment.
same as above, and replace by the run_ob action
There was a problem hiding this comment.
let's automate writing to the registry and tag the images with the last (?) commit or some other identifier to be able to track versions
decide on an oras registry and share credentials/omnibenchmarkteam-wise, check quotas
get mark to do the bitwarden/MLS omnibenchmark credentials storage/thing
There was a problem hiding this comment.
perhaps combine the R envs into one, and avoid compiling another R
There was a problem hiding this comment.
this goes to the omnibenchmark_paper repo and not to this one, here the makefile should be easier
| 2. Clone the benchmark definition / this repository with `git clone [email protected]:omnibenchmark/clustering_example.git` | ||
| 3. Move to the cloned repository `cd clustering_example` | ||
| 4. Run locally, somewhat in parallel `ob run benchmark -b CLUSTERING.YAML --local --threads 6`. Choose `Clustering.yml` specification based on whether running it with conda, easybuild, apptainer, etc. [More details about the available backends](https://github.com/omnibenchmark/clustering_example/blob/main/envs/README.md). | ||
| 1. Install omnibenchmark: `pip install omnibenchmark>=0.2.0` |
There was a problem hiding this comment.
conda is needed, lmod, envmodules etc - not only pip
|
|
||
| by Marek Gagolewski, modified by Izaskun Mallona | ||
| * All needed recipes can be found under `envs`: conda, apptainer, easybuild (lmod modules) |
There was a problem hiding this comment.
for the example this is way too complicated, let's move this to the paper repo
There was a problem hiding this comment.
same, delete, and make sure singularity is reproducible and called apptainer
There was a problem hiding this comment.
make sure the easyconfigs are merged/reproducible, discuss the conda dependency, perhaps mamba
There was a problem hiding this comment.
too many things to unpack - let's test and pick one of the two alternatives
but if versions are changing (package versions) then conda's and apptainer's should match
There was a problem hiding this comment.
same. btw we might need to document the bootstrap thing
There was a problem hiding this comment.
discussed why this sounds risky to me
There was a problem hiding this comment.
add versions, if not merged to fcps
|
after a review today, we'll be closing this PR in favor of #30 |
