Asia/code cleanup

.gitignore

@@ -11,6 +11,9 @@ src/results/*
 src/logs/*
 idea/*
 src/lightning_logs/*
+apptainer/*
+*.yaml
+*.sif
 notebooks/.ipynb_checkpoints/RAE-dev-checkpoint.ipynb
 src/utils/data/__init__.py
 src/utils/evaluation/__init__.py

.isort.cfg

@@ -0,0 +1,2 @@
+[tool.isort]
+profile = "black"

README.md

@@ -1,7 +1,7 @@
 # D4L-Hackaton
 
 ## Setup
-We recommend creating ```venv``` or ```conda``` environment with ```python==3.12```. 
+We recommend creating ```venv``` or ```conda``` environment with ```python==3.12```.
 
 ### Conda and requirements.txt
 
@@ -29,14 +29,14 @@ python3 train_and_validate.py [ARGUMENTS]
 
 with possible options:
   * ```-h, --help```: show this help message and exit
-  * ```--method``` {"omivae", "babel", "advae"} 
+  * ```--method``` {"omivae", "babel", "advae"}
   * ```--config``` (default="standard"): Name of a configuration in src/config/{method} directory.
   * ```--cv-seed``` (default=42): Seed used to make k folds for cross validation.
   * ```--n-folds``` (default=5): Number of folds in cross validation.
   * ```--subsample-frac``` (default=None): Fraction of the data to use for cross validation.
   * ```--retrain``` (default=True): Retrain a model using the whole dataset.
 
-### 
+###
 
 ## Dataset
 We use dataset from NeurIPS competition: https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE194122 named GSE194122_openproblems_neurips2021_cite_BMMC_processed.h5ad.gz.
@@ -50,6 +50,17 @@ log-scaled counts, as well as Spearman correlation.
 probability assigned to the correct matching. The latter is a relative measure per dataset that accounts
 for non-identifiability among cells with the same identity.
 
+### Pre-commit
+
+Before you can run hooks, you need to have the pre-commit package manager installed.
+
+Using pip: `pip install pre-commit`.
+
+In a python project, add the following to your requirements.txt (or requirements-dev.txt): `pre-commit`.
+
+Then run `pre-commit install` to set up the git hook scripts.
+
+
 ### Viewing logs
 To see the logs from different runs from the `src` directory run
 

container_vaes.def

@@ -0,0 +1,32 @@
+BootStrap: docker
+From: ubuntu:22.04
+
+
+%files
+        ./requirements_vaes.txt
+
+%post
+        TZ=Europe/Berlin
+        ln -snf /usr/share/zoneinfo/$TZ /etc/localtime && echo $TZ > /etc/timezone
+
+        apt -y update && apt -y upgrade
+
+        # Install python
+        apt install -y python3-pip
+        pip3 install --upgrade pip
+        apt-get -y install git
+
+        echo "PYTHON VERSION"
+        python3 --version # 3.10.12
+        pip3 --version # 23.3.2
+        pip3 install -r requirements_vaes.txt
+
+        # Clean cache to reduce size of the image
+        apt clean
+        rm -r $(pip cache dir)/*
+
+%startscript
+        PATH=$PATH:$HOME/.local/bin
+
+%labels
+   Author Ja

mrun.py

@@ -0,0 +1,28 @@
+import glob
+import os
+from argparse import Namespace
+
+import mrunner
+from mrunner.helpers import client_helper
+from mrunner.helpers.client_helper import get_configuration
+
+from src.main import main
+
+# mrunner.settings.NEPTUNE_USE_NEW_API = True # comment if running locally
+
+if __name__ == "__main__":
+    params = get_configuration(
+        print_diagnostics=True, with_neptune=True, inject_parameters_to_gin=True
+    )
+    params.pop("experiment_id")
+    # verify params: combine_with_defaults(params, defaults=vars(parse_namespace([])))
+    namespace = Namespace(**params)
+    try:
+        client_helper.experiment_["neptune_path"] = os.environ["NEPTUNEPWD"]
+    except KeyError:
+        pass
+
+    # if not namespace.cpdag and namespace.sampling:
+    #     pass  # add constraints here
+
+    main(namespace)