Upstream 23.2 updates #8
Open
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
* Fix introduced typo in partition information dictionary key. (PySlurm#241) * Added wait_finished method to job class (PySlurm#240). * Added test method for wait_finished method of the job class. * Added _load_single_job method to the job class to extract the slurm_load_job functionality. * Updated find_id and wait_finished to use _load_single_job. Co-authored-by: Jonathan Goodson <[email protected]>
* Fixed mem_per_cpu setting. * Fixed settings with mem_per_cp mentions.
* update the build-script to allow parallel builds Well, doesn't really do much right now since it's just one file being compiled - but maybe it will matter in the future. Also make it so the "install" command doesn't actually build anything, and just installs. Update the README to promote usage of the build.sh script Also remove the installation instruction for PyPi - it's not up there yet. * fix bdist_rpm not including pyslurm.pyx Just do a recursive include of all .pyx .pxd .pxi .h files
To be consistent with other functions and methods, we need to raise a ValueError instead of just returning the error code. This also allows client applications to more easily interpret the reason for the error.
* Fixes "incomplete type" complaints cython always complaint about "incomplete type" for pthread_mutex_t and sockaddr_storage. They must be atleast defined with a "pass", which makes the complaint go away. Also reformat the imports a bit. * Actually return the batch-script contents as a string. Previously it was just printed to stdout, so it wasn't really useful when wanting to process the contents any further in a script for example.
* Transition from sphinx to mkdocs for documentation * Tidy up docstrings, mkdocs config * Add logo to docs, hide navbar * Reformat docstring indentation * Missing newline * Fix more wrong indents * Add `mike` for versioning the docs * add site dir to gitignore * setup.py: Remove .so files on cleanup * header.pxi: Rename all "cpdef enum" to "cdef enum" cpdef didn't work correctly with Cython 3.0 onwards, which is needed for the docs. Setting this to "cdef enum" doesn't change any functionality in our case. * Remove embedsignature=True For building docs, we use Cython 3.0 now, which automatically provides full introspection functionality for cython functions * Build docs with Cython 3.0 --------- Co-authored-by: Michael Milton <[email protected]> Co-authored-by: Michael Milton <[email protected]>
Also use type annotations to document the return type
Mostly because this pretty much destroys syntax highlighting on github And use the return type within the "Returns" section in parenthesis. It must be with parenthesis to get it properly documented in mkdocs for now, see mkdocstrings/griffe#137
- update the pyslurm_bindgen.py tool accordingly
* Rework the Job-API * Rework the Node-API * Add valgrind suppression file * Rework slurmdbd Job API * rework tests directory structure, split into "unit" and "integration"
* update README - include logo - remove testing section (maybe it back separetely in the documentation) - remove workflow-badge (will only fail anyway, add back later when the building process has been updated) * update the documentation * move pyslurm.core.common package to pyslurm.utils * move pyslurm.core.db package to pyslurm.db * rename JobStats to JobStatistics
* Some more docstring enhancements * clean before building docs * enable autorefs and more markdown extensions
…Slurm#288) "PySlurm Developers" also just sounds a bit nicer actually
* set light/dark mode primary colors explicitly * Reuse logic to get the batch script of a job from the new Job API * improve job submit testing and fix bugs * JobSubmitDescription: fix some docstrings * change quality_of_service to just qos in pyslurm/db/job * fix date parse tests?
* Start maintaining a Changelog - include it in the docs * bump version to 23.2.1
- correctly use README.md instead of README.rst - folder for documentation is now called "docs" instead of "doc"
- Rework whole Partition API - avoid some duplicate logic in utils/uint.pyx - make the string "UNLIMITED" a constant - include new Partition API in the docs - add tests for new partition API
## Added
- Ability to modify Database Jobs
- New attributes for a Database Job:
- extra
- failed_node
- Now possible to initialize a pyslurm.db.Jobs collection with existing job
ids or pyslurm.db.Job objects
## Fixes
- Fixes a problem that prevented loading specific Jobs from the Database if
the following two conditions were met:
- no start/end time was specified
- the Job was older than a day
## Changes
- create a _get_exit_code helper function to reduce duplicate code in various places when checking for exit-code or exit-signal of a Job or JobStep
- handle modification of objects via `modify()` consistently - use correct list of node states when modifying the node state - remove `consumable_resource` attribute of Partition class again, logic is merged into the `select_type_parameters` attribute
Fixes PySlurm#302 - also makes it clearer in the docs for some attributes which type is expected inside the lists.
- All collections inherit from `list` instead of `dict` now (this change is mostly motivated by multi-cluster support that will probably be added in the future) - Added work-in-progress API for database `Associations` - renamed `JobSearchFilter` to `JobFilter`
- kinda revert the change from PySlurm#304 where all collections inherit from `list` - there is now a new base class `MultiClusterMap`, which acts as a dict but is capable of holding data from multiple clusters (for example when querying database jobs from multiple clusters in pyslurm.db.Jobs) - improve docs
* Improved docs + better custom repr * Add __or__, __ror__ and __ior__ to MultiClusterMap * Use from_ptr method when loading single Job/Step/Node
- and add testing for it
…rom the calculation. (PySlurm#313)
* Nodes: Some Bugfixes and new attributes - add idle_memory and allocated_tres properties - correctly return GRES in allocated_gres property - trailing whitespace fixes - add some new tests for GRES parsing
KILL_JOB_ARRAY was renamed to KILL_ARRAY_TASK. Remove it since it is not used anywhere in the code. See SchedMD/slurm@987557a Fixes: PySlurm#324
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.
No description provided.