Skip to content

docs: add per-data-type tutorials - #241

Merged
Zethson merged 15 commits into
mainfrom
feature/tutorials_usecases
Jun 24, 2026
Merged

docs: add per-data-type tutorials#241
Zethson merged 15 commits into
mainfrom
feature/tutorials_usecases

Conversation

@Zethson

@Zethson Zethson commented Jun 22, 2026

Copy link
Copy Markdown
Member

Adds runnable, GPU-executed tutorials for the paper's use cases under a new Guides by data type section:

  • scRNA-seq — sparse counts from CELLxGENE
  • Genetics (VCF / WGS) — 1000 Genomes → plink2 → cellink → annbatch
  • Microscopy images — scPortrait single-cell images in obsm
  • Distributed — multi-GPU sharding with DistributedSampler

Also restructures the quickstart to a small synthetic example and renames docs/notebooksdocs/tutorials.

Closes #240.

@review-notebook-app

Copy link
Copy Markdown

Check out this pull request on  ReviewNB

See visual diffs & provide feedback on Jupyter Notebooks.


Powered by ReviewNB

Add runnable guides for scRNA-seq, genetics (VCF/WGS), single-cell microscopy images, and distributed multi-GPU training, each executed end to end on GPU. Restructure the quickstart to a small synthetic example and rename docs/notebooks to docs/tutorials.
@Zethson
Zethson force-pushed the feature/tutorials_usecases branch from 96de0ae to 7352661 Compare June 22, 2026 12:54
@Zethson Zethson added the run-gpu-ci Signal that gpu ci should be run label Jun 22, 2026
@Zethson Zethson added this to the 0.2.0 milestone Jun 22, 2026
@github-actions github-actions Bot removed the run-gpu-ci Signal that gpu ci should be run label Jun 22, 2026
anndata passes arguments that scverse-misc 0.1 deprecates, so importing anndata to resolve the ImplicitModificationWarning filter raised under the `error` filter and broke pytest config parsing. The pattern is ordered before the category filters so it applies first. Also update the cruft skip path for the docs/notebooks -> docs/tutorials rename.
Comment thread docs/tutorials/distributed.ipynb
Comment thread docs/tutorials/images.ipynb
@codecov

codecov Bot commented Jun 22, 2026

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 91.91%. Comparing base (b7c1328) to head (7324be6).

Additional details and impacted files
@@            Coverage Diff             @@
##             main     #241      +/-   ##
==========================================
- Coverage   93.48%   91.91%   -1.58%     
==========================================
  Files          15       15              
  Lines        1397     1397              
==========================================
- Hits         1306     1284      -22     
- Misses         91      113      +22     

see 3 files with indirect coverage changes

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

Write the distributed tutorial's synthetic input as zarr v3 instead of h5ad, and build the images AnnData with no X at all (the image stack lives in obsm).
Comment thread docs/tutorials/distributed.ipynb
Comment thread docs/tutorials/genetics.ipynb
Comment thread docs/tutorials/images.ipynb
Comment thread docs/tutorials/quickstart.ipynb
@Zethson

Zethson commented Jun 22, 2026

Copy link
Copy Markdown
Member Author

@ilan-gold without going into details yet (sorry, I had forgotten to mark this as draft), do you generally like the structure as I'm suggesting it here? Is the high level content also okay?

@ilan-gold

Copy link
Copy Markdown
Collaborator

Oh yes, absolutely!

Use preload_nchunks=32 across the tutorials (bumping scrnaseq chunk_size to 512) so chunk_size * preload_nchunks stays well above batch_size, and filter anndata's zarr v3 autosharding warning emitted by the write_zarr cells.
Comment thread pyproject.toml Outdated
Drop the separate 'Guides by data type' caption and list every tutorial (quickstart, scRNA-seq, genetics, images, distributed) under one Tutorials section — distributed and the quickstart are not data types. Also restore the scrnaseq loader cell output.
Comment thread docs/index.md
@ilan-gold ilan-gold added the skip-gpu-ci Whether gpu ci should be skipped label Jun 22, 2026
@ilan-gold ilan-gold removed this from the 0.2.0 milestone Jun 23, 2026
Comment thread pyproject.toml Outdated
Comment thread docs/tutorials/distributed.ipynb
Comment thread docs/tutorials/quickstart.ipynb
Comment thread docs/tutorials/scrnaseq.ipynb
Zethson added 7 commits June 24, 2026 10:47
Signed-off-by: Lukas Heumos <lukas.heumos@posteo.net>
Signed-off-by: Lukas Heumos <lukas.heumos@posteo.net>
Signed-off-by: Lukas Heumos <lukas.heumos@posteo.net>
Signed-off-by: Lukas Heumos <lukas.heumos@posteo.net>
Signed-off-by: Lukas Heumos <lukas.heumos@posteo.net>
@Zethson
Zethson requested a review from ilan-gold June 24, 2026 13:07
Comment thread docs/tutorials/scrnaseq.ipynb
Signed-off-by: Lukas Heumos <lukas.heumos@posteo.net>
@Zethson
Zethson enabled auto-merge (squash) June 24, 2026 14:18
@Zethson
Zethson merged commit 28f3280 into main Jun 24, 2026
13 checks passed
@Zethson
Zethson deleted the feature/tutorials_usecases branch June 24, 2026 14:25
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

skip-gpu-ci Whether gpu ci should be skipped

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Add guide per data type

2 participants