Skip to content

Conversation

@thalassemia
Copy link
Contributor

@thalassemia thalassemia commented Dec 8, 2024

  • Use uv for Python package management and installation of the model
    • This resulted in a much simpler and more reproducible installation process
    • Update all containerized workflows (Google Cloud, Sherlock) to use uv
    • Update all packages to latest compatible versions and use Python 3.12
  • Move required data files out of data into reconstruction/ecoli/flat
  • Exclude data (now only required for testing) and other large folders like .git, notebooks, reconstruction/sim_data, and reconstruction/ecoli/scripts from container images for faster builds and leaner images
  • Try Altair plotting backend instead of hvplot, which requires hundreds of MBs in dependencies
  • Use Autograd instead of Jax for much lighter install, saving hundreds of MBs
  • Reformulate RNA degradation Km optimization as a minimization problem (see benchmarks here)
  • Go back to 4-core ParCa job that is submitted to SLURM queue on Sherlock (faster now that I've specified that jobs can run on the dev and normal partitions in addition to mcovert and owners)
  • Use a different seed for rich media Jenkins test because of numerical instability with default seed 0
  • Fix unique index generation so there is not even a slim chance of conflicts unless you generate more than 2**59 of a single unique molecule

TODO:

  • Wait for PR on stochastic-arrow to be merged and new wheels published
  • Try to fix migration tests. If too difficult, can get rid of them
  • Clean up repository. Now that the ParCa is fully functional, it is about time to scrub the pickle files that are adding a lot of bloat to the repo.

Also auto-delete temp files if error
@thalassemia thalassemia marked this pull request as ready for review December 12, 2024 01:22
@github-actions
Copy link
Contributor

Package Line Rate Complexity Health
. 30% 0
analysis 100% 0
analysis.antibiotics_colony 11% 0
analysis.antibiotics_colony.subgen_gene_plots 20% 0
analysis.causality_network 18% 0
analysis.colony 11% 0
analysis.multidaughter 100% 0
analysis.multiexperiment 100% 0
analysis.multigeneration 23% 0
analysis.multiseed 17% 0
analysis.multivariant 17% 0
analysis.single 15% 0
composites 35% 0
composites.ecoli_configs 90% 0
composites.environment 26% 0
debug 17% 0
ecoli 9% 0
ecoli.compendium 18% 0
ecoli.dataclasses 31% 0
ecoli.dataclasses.process 22% 0
ecoli.dataclasses.state 32% 0
ecoli.scripts 43% 0
ecoli.scripts.fold_changes 15% 0
ecoli.scripts.metabolism_kinetics 36% 0
ecoli.scripts.metabolite_concentrations 26% 0
ecoli.scripts.nca 10% 0
ecoli.scripts.protein_half_lives 29% 0
ecoli.scripts.rna_half_lives 21% 0
experiments 35% 0
io 52% 0
library 47% 0
library.cell_wall 14% 0
processes 43% 0
processes.antibiotics 27% 0
processes.chemotaxis 14% 0
processes.environment 29% 0
processes.listeners 78% 0
processes.membrane 27% 0
processes.spatiality 19% 0
processes.stubs 45% 0
tests 22% 0
tests.io 24% 0
tests.utils 25% 0
utils 28% 0
utils._netflow 45% 0
variants 36% 0
Summary 28% (8451 / 30006) 0

@thalassemia thalassemia merged commit 959462c into master Dec 12, 2024
6 checks passed
@thalassemia thalassemia deleted the tweaks branch December 12, 2024 03:04
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants