adding tests

.gitignore

@@ -152,7 +152,6 @@ data/main/embeddings.parquet
 slurm-*.out
 notebooks/tests/step_0_.h5ad
 data/main/gene_embeddings.parquet
-*.h5ad
 data/main/main_scenic+.parquet
 scdataloader.out
 data/main/9606.protein.links.v12.0.txt.gz
@@ -170,3 +169,8 @@ lightning_logs/
 collator_output.txt
 data/bias_sparse.npz
 todel/
+tests/*.ckpt
+tests/out_*.h5ad
+data/*.h5ad
+notebooks/**/*.h5ad
+notebooks/*.h5ad

Makefile

@@ -27,15 +27,14 @@ install:          ## Install the project in dev mode.
 .PHONY: fmt
 fmt:              ## Format code using black & isort.
 	$(ENV_PREFIX)isort scprint/
-	$(ENV_PREFIX)black -l 79 scprint/
-	$(ENV_PREFIX)black -l 79 tests/
+	$(ENV_PREFIX)black -l 110 scprint/
+	$(ENV_PREFIX)black -l 110 tests/
 
 .PHONY: lint
 lint:             ## Run pep8, black, mypy linters.
-	$(ENV_PREFIX)flake8 scprint/
-	$(ENV_PREFIX)black -l 79 --check scprint/
-	$(ENV_PREFIX)black -l 79 --check tests/
-	$(ENV_PREFIX)mypy --ignore-missing-imports scprint/
+	$(ENV_PREFIX)flake8 --ignore=E501,E203,E266,E265,W503 scprint/
+	$(ENV_PREFIX)black -l 120 --check scprint/
+	$(ENV_PREFIX)black -l 120 --check tests/
 
 .PHONY: test
 test: lint        ## Run tests and generate coverage report.

scprint/cli.py

@@ -156,11 +156,10 @@ def _run_subcommand(self, subcommand: str) -> None:
             preprocessor = Preprocessor(
                 do_postp=False,
                 force_preprocess=True,
-                skip_validate=True,
-                use_layer="counts",
             )
             adata = preprocessor(adata)
             conf = dict(self.config_init[subcommand])
+
             model = scPrint.load_from_checkpoint(
                 self.config_init[subcommand]["ckpt_path"], precpt_gene_emb=None
             )

scprint/model/flash_attn/activations.py

@@ -1,8 +1,5 @@
 # Copied from https://github.com/mlcommons/training_results_v1.1/blob/main/NVIDIA/benchmarks/bert/implementations/pytorch/model/layers/activations.py
-import math
-
 import torch
-import torch.nn as nn
 import torch.nn.functional as F
 
 # 1/sqrt(2*pi)-> 0.3989423

scprint/model/flash_attn/block.py

@@ -5,7 +5,6 @@
 
 import torch
 import torch.nn as nn
-import torch.nn.functional as F
 from torch import Tensor
 from torchvision.ops import StochasticDepth
 

scprint/model/model.py

@@ -3,13 +3,11 @@
 from torch import Tensor, optim, nn
 from lightning.pytorch.tuner.lr_finder import _LRCallback
 from lightning.pytorch.callbacks.lr_finder import LearningRateFinder
-import torch.distributed as dist
 import torch
 from galore_torch import GaLoreAdamW
 from math import factorial
 import lightning as L
 import os
-import numpy as np
 import copy
 from scipy.sparse import load_npz
 from huggingface_hub import PyTorchModelHubMixin

scprint/tasks/denoise.py

@@ -6,14 +6,10 @@
 from scdataloader.data import SimpleAnnDataset
 from scdataloader import Collator, Preprocessor
 from scprint.model import utils
-from scprint.tasks import compute_corr
 
-import bionty as bt
 from torch.utils.data import DataLoader
 import sklearn.metrics
-import pandas as pd
 
-from lightning.pytorch import Trainer
 import anndata as ad
 from anndata import AnnData