You can now filter by “failed” status in the
statuscommand (thanks Ludvig).
The output of
gwf status -f summaryis now more consistent, has colors, and shows all possible states instead of only non-zero ones (thanks, Jakob Grove).
Fixed bug where a failed target would not be included as dependency when submitting an upstream target.
When accounting is enabled in the Slurm backend (on by default), gwf will now requests job states from both
sacct, with the status from squeue taking precedence. This prevents strange-looking output from
gwf statuswhen a target has been submitted, but its state has not yet been flushed to the database by Slurm.
You can now specify which parts of a workflow to touch with the gwf touch command. E.g.
gwf touch Foowill only touch Foo and its dependencies (and their dependencies).
Fixed loading of entrypoints on Python 3.7.
This is a major release. Many internals have been changed, specifically on the scheduler, which is now much more robust and consistent, and to backends, which now can report whether a target failed or not. The local backend has also been rewritten from scratch. In the future it will be able to support e.g. target cancellation.
For supported backends (currently slurm and local), the status command can now show failed targets. These will be rescheduled as if they were incomplete. If accounting is not enabled in slurm, this functionality must be turned off with gwf config set backend.slurm.accounting_enabled no.
The configuration directory will now always be placed next to the workflow file in use.
The status command is now consistent with the output of `gwf run`n.
The scheduler will not re-run targets that are not “should run”.
Targets with only outputs will now be scheduled if an output file is missing. Otherwise, it will not be scheduled.
And many, many other tiny things!
Workflows can no longer be included into each other. All namespacing has been removed.
Templates returning tuples are no longer supported. Templates must now return an
APIs for loading workflow, using the scheduler, creating and loading backends etc. have been changed. This only affects plugins and custom backends.
The gwf init command has been removed. Instead, gwf will ask if a skeleton should be created in the current directory, if it can’t find an existing workflow file.
Spec hashing is now turned off by default since it can be confusing. It can be turned on with gwf config set use_spec_hashes yes.
Spec hashes will now be computed automatically on the first run with spec hashes enabled.
gwf touchwill now also update spec hashes.
Various tiny fixes to the scheduler, nothing user visible.
Status output regression reported in #399 has been fixed.
You can now protect an output file from being deleted by the
gwf cleancommand with the
protectargument when creating a target. Specify a list of the files to be protected and
gwf cleanwill ignore them (even
gwf clean --all). This was implemented before, but it was broken, so this is considered a fix.
You can now load modules next to your workflow file, e.g.,
Including one workflow into another workflow had very few real use cases, so the feature has been dropped.
Will now recursively look for a workflow file in parent directories. This means that you can now run e.g.
gwf statusin any subdirectory below the directory containing your workflow file.
Fixed bug in the scheduling code which caused workflows that ran just fine to report “shouldrun” for some targets.
The scheduler now outputs a data structure with all information from the scheduling process and with separate types for each “reason”.
gwf statuscommand is now more informative about the state of each target.
Drop support for Python <3.7.
Use a nicer theme for the website!
Removed the automatic update checking as it was fragile and usually not useful.
If no backend is configured, gwf will not try to guess which backend to use based on which commands are available on the system.
Pretty output for the
--format pretty. Still supports the JSON output format as the default.
clean_logssetting which defaults to yes (the current behaviour). Setting this to no will not remove old logs when
gwf runis executed.
Re-run targets when their spec changes (#180). Can be disabled with
gwf config set use_spec_hashes no. Enabled by default.
Remove use of
time.clock()as it was deprecated.
Fix import of
Replaced use of
Trying to cancel a target that is not running or completed will now not fail.
Documented the Slurm target options
Slurm backend now accepts a
grestarget option that maps directly to Slurm’s
cleancommand now asks for confirmation when no targets are specified. This can be avoided by using the
The Slurm backend will now get all jobs from
squeueand not only jobs belonging to the current user. This used to be the default behaviour, but was unintentionally changed when refactoring the Slurm backend.
The output of the
cancelcommand is now correct.
This release is the first version of gwf to support Python 3.8. We’re now testing on nightly version of Python so that future Python releases should be supported as soon as they’re available.
Automatically checks for updates on a regular basis. A warning will be shown when a new version is available. This feature can be disabled with
gwf config set check_updates no.
The Slurm backend will now retry failed operations, e.g. if it fails to get the state of the queue from Slurm or if submitting a job failed.
Support for Python 3.8.
When running gwf cancel without specifying targets, only submitted/running targets will be cancelled.
Fixed a deadlock in the local backend which caused the workers to freeze.
Links to examples and ReadTheDocs (we’re now hosting the docs ourselves).
This release contains a few large features such as named inputs and outputs
and the introduction of the
Workflow.map method for easily generating
multiple targets from a template and improving readability.
There are also several minor improvements that contribute to the overall
user experience ranging from speed improvements for the
logs command to
improved debugging output.
The documentation has also been restructured and improved.
Named inputs and outputs. The
Workflow.targetcan now be either a string, list or dictionary. See the documentation for more details.
Tutorial now explains what happens if a target fails.
Documentation now has an official list of gwf plugins.
statuscommand now has a
--summaryoption the summarizes the status of an entire workflow.
All input and output paths are now checked for non-printable characters such as newlines and tabs. This can cause problems that are very hard to find and fix (e.g. gwf reporting that a file is missing even though it seems to be there). Paths containing such characters now result in an error.
logscommand is now much faster since it no longer builds the entire graph.
The target in the
statusoutput is now sorted in creation order, instead of alphabetically.
Cleaner output formatting, especially when running with
Improved log messages from scheduler when running with
-v debug. The messages are now more specific and helpful.
Documentation has been restructured to be more readable and have less redundant information.
Crash when running
gwf initwithout an existing configuration file.
--forceflag for the
cancelcommand now actually forces cancellation.
We now respect the
--no-colorflag completely and implement the
NO_COLORenvironment variable standard described here.
Crash when Slurm returns unknown job state (#244).
Users can now run
gwf initto bootstrap a new gwf project (c78193).
Add option to protect output files in a target from being removed when
gwf cleanis being run (2f51ed).
Ensure job script end with a newline (#239).
Ignore missing log files when cleaning on run (#237).
Backend for Sun Grid Engine (SGE). The backend does not support all target options supported by the Slurm backend, so workflows can not necessarily run with the SGE backend without changes. See the documentation for a list of supported options.
gwf statuscommand now accepts multiple
-s/--statusflags and will show targets matching any of the given states. E.g.
gwf status -s completed -s runningwill show all completed and running targets.
A new command
gwf touchhas been introduced. The command touches all files in the workflow in order, creating missing files and updating timestamps, such that gwf thinks that the workflow has been run.
When specifying the workflow attribute in the workflow path, e.g.
gwf -f workflow.py:foo, the filename part can now be left out and will default to workflow.py. For example,
gwf -f :foowill access the
fooworkflow object in workflow.py.
Documentation describing advanced patterns for gwf workflows.
This release contains a bunch of new features and plenty of bug fixes. Most
noteworthy is the removal of the progress bars in the status command. The status
bars were often confusing and didn’t communicate much more than a simple
“percentage completion”. The status command now outputs a table with target
name, target status, and percentage completion (see the tutorial for examples).
Additionally, the status command now shows all targets by default (not only
endpoints). For users who wish to only see endpoints, there’s now a
We aim to make gwf a good cluster citizen. Thus, logs from targets that no
no longer exist in the workflow will now be removed when running
This ensures that gwf doesn’t unnecessarily accumulate logs over time.
Add missing import to documentation for function templates (4eddcac).
Remove reference to
Remove broken badges in README (e352f09).
Remove pre-1.0 upgrade documentation (bfa03da6).
Fixed bug in scheduler that caused an exception when a target’s input file did not exist, but the output file did (reported by Jonas Berglund) (92301ef3).
Dots have been removed from logging output to make copy-pasting target names easier (f33f7195).
Now uses pipenv to fix development environment.
Improved coloring of logging output when running with
Remove status bars in
gwf statuscommand (47cb7b50).
Added undocumented API which allows core and plugins to register validation functions for configuration keys. This fixes issues like #226 (c8c57d7c7).
gwf cleancommand now shows how much data will be removed (d81f143f1).
Remove log files for targets that are no longer defined in the workflow (beb912bd).
Note in tutorial on how to terminate the local workers (a long with other updates to the tutorial) (34421498).
Bug when returning an
AnonymousTargetfrom a template function without specifying the working_dir in the constructor (#212). Thanks to Steffen Møller-Larsen for reporting this.
Bug when using
--format tableand no targets were found (#203).
Bug when cancelling a target running on the Slurm backend (#199).
Link to documentation in error message when unable to connect to local workers.
Fixed bug in the FileLogManager where the wrong exception was raised when no log was found.
Moved checking of file timestamps to the scheduler. This means that creating a
Graphobject will never touch the file system, and thus won’t raise an exception if a target depends on a file that doesn’t exist and that’s not provided a target. Instead, unresolved paths are added to
Graph.unresolved. They will then be checked by the scheduler (if necessary). For end users, this means that many commands have become substantially faster.
AnonymousTargetwhich represents an unnamed target.
Targetnow inherits from this class and templates may now return an
AnonymousTargetinstead of a tuple.
Added backend.slurm.log_mode option, see the documentation for the Slurm backend for usage (#202).
Very slow scheduling when using dry run with unsubmitted targets (#184, 93e71a).
Fixed cancellation with the Slurm backend (#183, 29445f).
Fixed wildcard filtering of targets (#185, 036e3d).
Move file cache construction out of
Graph(#186, 93e71a). This change is invisible to end-users, but speeds up the
Made filtering more intuitive in all commands.
infocommand now outputs JSON instead of invalid YAML.
infocommand outputs information for all targets in the workflow by default.
Backends must now specify a
log_managerclass attribute specifying which log manager to use for accessing target log files.
Backends should now be used as context managers to make sure that
Backend.close()is called when the backend is no longer needed, as it is no longer called automatically on exit.
Added filtering of targets by name in the
Added API documentation for the
SlurmBackendto make it easier for plugins to integrate with Slurm.
Documentation for log managers.
Documentation on how to handle large workflows.
First stable release of gwf! We strongly encourage users of pre-1.0 users to read the tutorial, since quite a lot of things have changed. We also recommend reading the guide for converting pre-1.0 workflows to version 1.0. However, users attempting to do this should be aware that the the template mechanism in 1.0 is slightly different and thus requires rewriting template functions.
Fixed a bug which caused gwf to fail when cancelling jobs when using the Slurm backend (8c1717).
Documentation in various places, especially the core API.
Documentation for maintainers.
Topic guide covering templates (b175fe).
Fixed a subtle bug in scheduling which caused problems when resubmitting a workflow where some targets were already running (a5d884).
Fixed a bug in the
SlurmBackendwhich caused gwf to crash if the Slurm queue contained a job with many dependencies (eb4446).
Added back the -e flag in the
Fixed a bug in the
SlurmBackendwhich caused running targets as unknown (33a6bd).
The Slurm backend’s database of tracked jobs is now cleaned on initialization to keep it from growing indefinitely (bd3f95).
Fixed a bug which caused the gwf logs command to always show stderr (01b267).
Fixed a bug which caused dependencies to be set incorrectly when two targets depended on the same target (4d9e07).
Improved error message when trying to create a target from an invalid template (d27d1f).
Improved error message when assigning a non-string spec to a target (2aca0a).
gwf logs command now outputs logs via a pager when the system supports it, unless –no-pager is used (01b267).
Added more tests to cover scenarios with included workflows when building the workflow graph (86a68d0).
Added a bunch of documentation (69e136, 51a0e7, 942b05).
Fixed bug in scheduling which was actually the cause of the incorrect scheduling that was “fixed” in 1.0b6. Also added documentation for
Updated documentation in a bunch of places, mostly styling.
A bug in
SlurmBackendwhich caused dependencies between targets to not be set correctly (6b71d2).
More improvements to and clean up of build process.
Updated some examples in the tutorial with current output from gwf (42c5da).
Logging output is now more consistent (b95af04).
Documentation for maintainers on how to merge in contributions and rolling a new release (fe1ee3).
Unset option passed to backend causes error (#166, dcff44).
Set import path to allow import of module in workflow file (64841c).
Vastly improved build and deploy process. We’re now actually building and testing with conda.