Closed enryH closed 6 months ago
…#66)
:sparkles: test Snakemake workflow with more recent Python versions
:sparkles: remove snakemake upper limit
:bug: bump plac version
related to : https://github.com/snakemake/snakemake/issues/2276
:pushpin: update actions, make artefacts unique
check if local windows error with pandas can be reproduced in action (corr)
:art: dump counts for histograms
add for simulated missing values
remove duplication
:bug: do not omit last bin
:bug: assure that some values are set to NA
if all values are higher than the default threshold, the assertion on L17 is not met. Make sure some NAs (missing values) are set.
:art: write out corr and prepare for pandas 2.0
see if this works also with pandas 1.5.3
:wrench: test relaxing pandas restriction
:bug: drop batches with one sample for training DAE and VAE
for creating the latent representation, now a new DataLoader is needed.
:sparkles: splitup large global environement
separate environment for PIMMS models and R based models
global environment should still work
:bug: test if adding jupyter is sufficient to install further packages in R sesssion
only execute one job at a time in retry to see errors better
:bug: fix sampling to make it compatible with python >=3.11
:arrow_up: remove pytorch upper dependency
:white_check_mark: Test workflow v2 on Alzheimer dataset
once this passes, add ald analysis to website (for a reasonable subset of models)
maybe only showcase PIMMS models with a handful of other models
:bug: update path to execute run, speed-up
also remove two slowest models
:art: hide code in rendered notebooks of workflow, sort imports
hide code cells for generated report
isort imports
:sparkles: Functionality for plot source data (ALD study)
add some functionality required to collect source data for reporting on saved figures.
:sparkles: Run differential analysis workflow in CI on Alzheimer data
several adaptions to slightly different design between ALD and Alzheimer data
:bug: specify folder_experiment from global space
... and not as wildcard
:art: rename Snakefile_v2 to Snakefile_v2.smk
uncommon names should have a file ending specifying Snakefiles.
:sparkles: skript to build website (execution)
execution should work, but subfolders need their own index.rst
need to adapt script for updating main index.rst
See if everything runs for now.
:bug: do not exclude diff analysis folder in conf.py
:art::bug: make a strict hierarchy of headings per document
mapping titles in sphinx (cross-referencing) otherwise does not work
:art: collapse code in published notebooks
for better inspection of generated report for example
:art: annotate notebooks
add some comments and streamline cells.
:sparkles: Test tutorial on colab
:bug::art: format and check briefly colab workflow on dev branch
:art: hide more inputs, downscale tutorial runner
:memo: Update README
:bug: use larger image to test tutorial on colab
:memo: update READMEs and add some hints to config files
:memo::art: save some adhoc script used during revisons, add and cleanup nb list
:bug: go back old config indentation (and model configuration)
rerun in codespace for inspection
:bug: fix issue having same model with 2 configurations
had to set model id ("model_key") as index
:memo::sparkles: Allow users to download large HeLa protein groups dataset easily
…#66)
:sparkles: test Snakemake workflow with more recent Python versions
:sparkles: remove snakemake upper limit
:bug: bump plac version
related to : https://github.com/snakemake/snakemake/issues/2276
:pushpin: update actions, make artefacts unique
check if local windows error with pandas can be reproduced in action (corr)
:art: dump counts for histograms
add for simulated missing values
remove duplication
:bug: do not omit last bin
:bug: assure that some values are set to NA
if all values are higher than the default threshold, the assertion on L17 is not met. Make sure some NAs (missing values) are set.
:art: write out corr and prepare for pandas 2.0
see if this works also with pandas 1.5.3
:wrench: test relaxing pandas restriction
:bug: drop batches with one sample for training DAE and VAE
for creating the latent representation, now a new DataLoader is needed.
:sparkles: splitup large global environement
separate environment for PIMMS models and R based models
global environment should still work
:bug: test if adding jupyter is sufficient to install further packages in R sesssion
only execute one job at a time in retry to see errors better
:bug: fix sampling to make it compatible with python >=3.11
:arrow_up: remove pytorch upper dependency
:white_check_mark: Test workflow v2 on Alzheimer dataset
once this passes, add ald analysis to website (for a reasonable subset of models)
maybe only showcase PIMMS models with a handful of other models
:bug: update path to execute run, speed-up
also remove two slowest models
:art: hide code in rendered notebooks of workflow, sort imports
hide code cells for generated report
isort imports
:sparkles: Functionality for plot source data (ALD study)
add some functionality required to collect source data for reporting on saved figures.
:sparkles: Run differential analysis workflow in CI on Alzheimer data
several adaptions to slightly different design between ALD and Alzheimer data
:bug: specify folder_experiment from global space
... and not as wildcard
:art: rename Snakefile_v2 to Snakefile_v2.smk
uncommon names should have a file ending specifying Snakefiles.
:sparkles: skript to build website (execution)
execution should work, but subfolders need their own index.rst
need to adapt script for updating main index.rst
See if everything runs for now.
:bug: do not exclude diff analysis folder in conf.py
:art::bug: make a strict hierarchy of headings per document
mapping titles in sphinx (cross-referencing) otherwise does not work
:art: collapse code in published notebooks
for better inspection of generated report for example
:art: annotate notebooks
add some comments and streamline cells.
:sparkles: Test tutorial on colab
:bug::art: format and check briefly colab workflow on dev branch
:art: hide more inputs, downscale tutorial runner
:memo: Update README
:bug: use larger image to test tutorial on colab
:memo: update READMEs and add some hints to config files
:memo::art: save some adhoc script used during revisons, add and cleanup nb list
:bug: go back old config indentation (and model configuration)
rerun in codespace for inspection
:bug: fix issue having same model with 2 configurations
had to set model id ("model_key") as index
:memo::sparkles: Allow users to download large HeLa protein groups dataset easily