Pin tiledb version as work-around #454
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
ebezzi
approved these changes
May 2, 2023
Codecov Report
@@ Coverage Diff @@
## main #454 +/- ##
=======================================
Coverage 88.51% 88.51%
=======================================
Files 50 50
Lines 2770 2770
=======================================
Hits 2452 2452
Misses 318 318
Flags with carried forward coverage won't be shown. Click here to find out more.
📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
pablo-gar
approved these changes
May 2, 2023
mlin
added a commit
that referenced
this pull request
Jun 22, 2023
* Pin tiledb version as work-around (#454) * temporarily pin required tiledb version * revert release.json URL while infra changes are fluid * Incorporate comms changes (#456) * Final edits to gget tutorial (#451) * revert R API release.json URL (#458) the new URL is not yet ready and Python API has already been reverted * symlink the new notebooks * Fix docsite version and disable searchbar (#460) * update release boostrap URL to public permalink (#459) * Remove pre-release from README.md (#462) * Enable anonymous access to S3 bucket (#275) * Enable anonymous access to S3 bucket * R + unit tests * Pin tiledbsoma==1.2.3 * R styler * Update api/python/cellxgene_census/tests/test_open.py Co-authored-by: Andrew Tolopko <[email protected]> * remove R, upgrade pyproject * remove R * add newline * add negative test --------- Co-authored-by: Andrew Tolopko <[email protected]> * Bump static census version in R tests (#472) * [r] update `get_presence_matrix()` and vignette to use zero-based matrix view (#475) * wip * wip * update census_dataset_presence.Rmd * add acceptance test run (#485) * rerun notebooks for census build 2023-05-15 ("stable") Census release (#484) * Updated gget cellxgene tutorial to reflect workflow updates [formatting corrected] (#490) * Created using Colaboratory * Created using Colaboratory * Created using Colaboratory * correct formating --------- Co-authored-by: Laura Luebbert <[email protected]> * Add docsite version number from the library version (#481) * Add docsite version number from the library version * revert odd commit * pin tiledbsoma==1.2.4 (#493) * [docs] Add autosummary (#492) * autosummary * Add module desc + fix links * [R] close census objects (#486) * Using the new stateful open/close in R tiledbsoma -- mainly docs/vignettes/tests, but a few of the helper implementations too. * Refactored `open_soma()` to facilitate sharing/reuse of `SOMATileDBContext`, and do that hroughout the tests. * Updated vignettes to reflect recent changes to the Python notebooks. * However, the vignettes are now using too much memory to build in GHA. Still troubleshooting this, but for: disabled building them in GHA in order to un-break our CI. * Fix runs-on to use matrix strategy in py-unittests (#494) * Fix runs-on to use matrix strategy in py-unittests * try ARM64 runner * roll back ARM64 runner * add if * Revert bad commit * Enable anonymous access in R (#471) * Enable anonymous access to S3 bucket * R + unit tests * Pin tiledbsoma==1.2.3 * R styler * Update api/python/cellxgene_census/tests/test_open.py Co-authored-by: Andrew Tolopko <[email protected]> * fix * fix * reset credentials --------- Co-authored-by: Andrew Tolopko <[email protected]> * Add Databricks install instructions to FAQ (#488) * [docs] Fix the Census link in navbar (#491) * [r] use `stable` by default & add alias resolution message (#502) Completes #482 For parity with python #435 Also adapts to several recent breaking API changes in tiledbsoma * PyTorch DataLoader (#499) * Add PyTorch DataLoader support * Introduce this code under a new "experimental" sub-package, with new pytest "experimental" marker for unit tests. * Add initial PyTorch example code for LR model training. Not a notebook yet, but under notebooks dir for now. * chore: update lifecycle tags (#509) * Update lifecycle tags for non-experimental Python API to "maturing" * Update lifecycle tags for experimental Python API to "experimental" * export public names for experimental ml package * bump python api tiledbsoma version (#510) bump tiledbsoma to 1.2.5, which includes updated api doc lifecycle tags * minor clarifications for the pypi.org release process (#512) * cache most R dependencies to speed up r-check CI (#517) #309 -- Cache most R dependencies instead of always building the latest versions of all of them. (Then immediately afterwards, still install the latest tiledb & tiledbsoma from r-universe) * [r] Add comp_bio_census_info.Rmd (#407) Also: - update all vignettes to recent tiledbsoma API evolution - temporarily move vignettes into `vignettes_wip/` pending a plan for how to build them outside of GitHub Actions Co-authored-by: Emanuele Bezzi <[email protected]> * fix pytorch multiprocessing result (#516) The first partition of data was being returned from each worker, apparently caused by use of a PyArrow array for passing the set of joinids to each worker's result iterator, possibly due to a bug in TileDB-SOMA. * Update release_process.md (#520) * highly variable gene annotation (#511) * initial implementation of highly_variable_genes * add test marks * add prebuffered iterator * lint * lint * docstrings * reduce expensive tests * fix typo * actually fix typo * add test for get_highly_variable_genes * lint * reduce memory use in tests * add example to docstring * fix anon access in small memory context * PR feedback * loess jitter * increase max loess noise max to 1e-6 * add tests * fix: pytorch unit test hangs (#522) * force use of multiprocessing spawn start method for pytorch * run experimental tests in all envs except 3.7 * Add support for Python 3.11 (#528) * Add support for Python 3.11 * Add 3.11 to test workflow * update notebook guidelines (#534) * update notebook guidelines * Update docs/census_notebook_guidelines.md Co-authored-by: Emanuele Bezzi <[email protected]> * Update docs/census_notebook_guidelines.md * Apply suggestions from code review Co-authored-by: Andrew Tolopko <[email protected]> --------- Co-authored-by: Emanuele Bezzi <[email protected]> Co-authored-by: Andrew Tolopko <[email protected]> * Add docs for experimental modules (#537) * Add modules to python-api.rst * Add README.md * Update docs/python-api.rst Co-authored-by: pablo-gar <[email protected]> * Update docs/python-api.rst Co-authored-by: pablo-gar <[email protected]> --------- Co-authored-by: pablo-gar <[email protected]> * demo notebook for HVG experimental API (#536) * reorg files * add HVG notebook * clean up notebook docs symlinks * lint * lint,try 2 * fix Ruff error * fix lint in thread regex * work around upstream lint * update target python version to match target container version * fix typing lint * update and fix config for pre-commit * PR feedback / fixes * Add experimental notebooks to docsite + fix headers (#541) * fix OOM on pytorch unit test (#542) skip failing test on 3.9 only * Add pytorch notebook (#551) * add pytorch notebook; minor pytorch api docstring updates * CR feedback * Update api/python/notebooks/experimental/pytorch.ipynb * Update api/python/notebooks/experimental/pytorch.ipynb * run full notebook --------- Co-authored-by: pablo-gar <[email protected]> Co-authored-by: Pablo E Garcia-Nieto <[email protected]> * fix incorrect pytorch obs soma_joinids (#555) - Use torch.from_numpy() instead of Torch.Tensor() to construct tensors. The latter created with dtype=float32, which caused the bug due to truncation of precision when casting to int64-to-float32-to-int64. In addition to fixing the bug, a data copy is eliminated by using this new method. - Rework the test fixture generation methods to allow for ranges of obs and var soma_joinids that may start at any arbitrary value. Necessary for testing the case that produced this bug. - Add asserts to ease paranoia. * Fix MacOS failing tests (#557) * Fix MacOS failing tests For the time being, we'll try to determine if the issue is specific to a Python version by removing 3.7. * try macos13 * comma * try if * exclude * typo --------- Co-authored-by: Bruce Martin <[email protected]> Co-authored-by: pablo-gar <[email protected]> Co-authored-by: Laura Luebbert <[email protected]> Co-authored-by: Andrew Tolopko <[email protected]> Co-authored-by: Emanuele Bezzi <[email protected]> Co-authored-by: Martin Kim <[email protected]> Co-authored-by: Pablo E Garcia-Nieto <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Primary change: pin the tiledb version dependency to >=0.21.3 to account for a missing dependency in tiledbsoma==1.2.2. See single-cell-data/TileDB-SOMA#1322 for details. This work-around can be removed when the issue is resolved upstream, and we bump our tiledbsoma pin.
Secondary change: reverted the URL change for the release.json bootstrap file, introduced in #443. This will be further updated when the final permalink, and infra changes to support that permalink, are in place.
NOTE: this PR did not revert the R bootstrap URL. That needs to be done at some point before it is released. CC: @mlin and @pablo-gar