Su 202503 #392
Open
Su 202503 #392
Conversation
This file contains hidden or 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
Codecov Report❌ Patch coverage is Additional details and impacted files@@ Coverage Diff @@
## develop #392 +/- ##
===========================================
+ Coverage 47.53% 47.57% +0.04%
===========================================
Files 239 239
Lines 22135 22138 +3
===========================================
+ Hits 10522 10533 +11
+ Misses 11613 11605 -8 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
the CaHVA and CaLVA begin and end are set according to the specific morphology, these shouldn't have default values
[BUG] Fix path resolution for database register & correct import
Remove code formatting for `pixi`
Tests don't run on some windows machines, because the combination of
long filenames in `simrun` (such as dendritic voltage traces) and long
pytest `tmpdir` directories (i.e.
`{temproot}/pytest-of-{user}/pytest-{num}/{testname}/`) quickly exceed
the `260` character limit imposed on most windows machines. Rather than
requiring users to mess with the registry, I implemented a different
default `basetmp` for windows.
Other fix: No more verbose error output for `spiketrain` creation. See
#360
Add pytest.mark.statistical for statistical tests. This is much improved over XFAIL. Short paths on windows (see #400) Download the barrel cortex model during testing. Idk why I didn't do this earlier, it's pretty important. Add timeout for entire test session (10 minutes) as well as single function (3 minutes). Reduce warnings during testing: - Don't call plt.show() on non-GUI backends - Don't use np.fromstring, but rather np.frombuffer - Don't check empty lists using any() Add pytest-timeout as dependency. This should prevent deadstalling tests (if they still pop up) from eating through our github minutes
# API Added `copy_tree` to our own implementation of `NTParameterSet` to be backwards compatible with `sumatra`'s implementation. this is used in network mapping. See 7ee9782 # Performance Vectorize network mapping (see PR #408) More conscious output during network mapping. Only main points are INFO level, the rest ist DEBUG level. # Compatibility macOS sometimes automatically creates `.DS_Store` files in folders for god knows what. This confuses databases: they think it's a key they need to construct metadata for. Databases now ignore these files. # Test fitting synaptic strength hovers around the 180s mark, which is also our test timeout. I set it so this one test is allowed to take longer.
## Bugfixes - Now properly fetch and resolve the exact parameter file name for re-running simulation trials (or in general, really) - Update error message when reading a ModelDataBase using DataBase when you haven't yet imported `compatibility`, thanks to @su-saka ## Data base initialization from `simrun` See PR #420 - Allow passing an integer to `repartition` to pass a target partition size for the voltage traces - Added health check that scans simulation results to see if they can be initialized and reproduced. htis was necessary to handle some missing data, throwing off the `init()` function. See PR #420 - Copying parameterfiles is now parallellized - You can pass `param_file_copy_kwargs` to specify where and how parameterfiles should be copied to the `DataBase` - Allows for using the recsite ID as subdirectory for dendritic vts, rather than the full label (useful for slightly scaled morphologies) - Allow specifying the file format for database `meta`s (for msgpack and parquet) in `config.db_settings.json`. Available formats are `json` and `msgpack`. See #416 - More robust fetching of dendritic voltage traces with `get_recsite_labels_from_dend_vt_filelist`. Avoids having to pass around a `suffix_dict` too. - Faster reading of voltage traces using `pd.read_csv()` instead of `np.readtxt()` - Refactor `load_dendritic_voltage_traces_helper` for readability. responsibility of handling of old/new/newer filename conventions is passed to `data_base.db_initializers.load_simrun_general.file_handling.get_recsite_labels_from_dend_vt_filelist` - Don't error anymore when soft references (e.g. refs to .hoc in .syn files) are not found: this is not a simulation-breaking issue, just a minor inconvenience. `.hoc` files are sometimes saved in the `.syn` files with relative paths, which makes it impossible to infer the exact location. - Speed up extracting paramfile reference from other paramfiles using regex instead of building all paramfiles ## Documentation See PR #421 ## Other changes - Deprecate the `tStim` argument in `simrun.run_new_simulations`. It is unused, and led to a decent amount of confusion for me. Stimulus onset is defined per population as an `offset` key in the network parameters. - Update `pixi` version in CI (building cythong packages with conda's `gcc` and Python not pinned down was broken otherwise due to missing `crypt.h` lib) - Allow passing user flags to compiling NEURON mechanisms. This turned out to not be necessary to fix compilation on newer systems (now we simply pin down `gcc` on linux), but is kept as a feature. - Add `BlenderSpike` as a run-dependency to the environment
This PR separates out the documentation dependencies from run/default dependencies, and updates them accordingly to fix the broken documentation build
remove unnecessary args and simplify the code to load existing results
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.
Refactor the RW class and add the option to provide custom exploration mode