From b15d744bc446cdea0bdd3995295c68afa7eb1e6a Mon Sep 17 00:00:00 2001
From: d33bs <dave.bunten@cuanschutz.edu>
Date: Tue, 23 Jul 2024 16:10:35 -0600
Subject: [PATCH] linting

---
 .pre-commit-config.yaml                       |  3 +-
 .../create_image_and_parquet_data.py          | 75 -------------------
 .../create_mask_data.py                       |  4 +-
 3 files changed, 4 insertions(+), 78 deletions(-)
 delete mode 100644 tests/data/cytotable/NF1_cellpainting_data_shrunken/create_image_and_parquet_data.py

diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml
index 3f5f266..f2b1cb5 100644
--- a/.pre-commit-config.yaml
+++ b/.pre-commit-config.yaml
@@ -23,7 +23,8 @@ repos:
           (?x)^(
               .*\.lock |
               .*\.json |
-              .*\.ipynb
+              .*\.ipynb |
+              .*\.cppipe
           )$
   - repo: https://github.com/executablebooks/mdformat
     rev: 0.7.17
diff --git a/tests/data/cytotable/NF1_cellpainting_data_shrunken/create_image_and_parquet_data.py b/tests/data/cytotable/NF1_cellpainting_data_shrunken/create_image_and_parquet_data.py
deleted file mode 100644
index 9451a4f..0000000
--- a/tests/data/cytotable/NF1_cellpainting_data_shrunken/create_image_and_parquet_data.py
+++ /dev/null
@@ -1,75 +0,0 @@
-"""
-Creates and shrunken dataset for testing puproses
-based on coSMicQC/tests/data/cytotable/NF1_cellpainting_data (Plate 2)
-
-This file may be processed using the following command from the root
-of the project repository:
-`poetry run python \
-tests/data/cytotable/NF1_cellpainting_data_shrunken/create_image_and_parquet_data.py`
-"""
-
-
-import pathlib
-import shutil
-
-import pandas as pd
-
-source_data_path = "tests/data/cytotable/NF1_cellpainting_data/"
-target_data_path = "tests/data/cytotable/NF1_cellpainting_data_shrunken/"
-source_image_data_path = source_data_path + "Plate_2_images"
-source_parquet_path = source_data_path + "Plate_2_with_image_data.parquet"
-target_image_data_path = target_data_path + "Plate_2_images"
-target_parquet_path = target_data_path + "Plate_2_with_image_data_shrunken.parquet"
-
-# create target image dir
-pathlib.Path(target_image_data_path).mkdir(exist_ok=True)
-
-# read source data
-source_df = pd.read_parquet(source_parquet_path)
-
-# get a sample of 5 from the source data
-sampled_df = source_df.sample(n=5)
-
-# send the sampled df to parquet file
-sampled_df.to_parquet(target_parquet_path)
-
-
-def check_and_copy_file(filename: str):
-    """
-    Checks for files in target dir and copies them if they don't already exist
-    """
-    source_path = f"{source_image_data_path}/{filename}"
-    target_path = f"{target_image_data_path}/{filename}"
-
-    # Check if the file already exists in the target directory
-    if not pathlib.Path(target_path).is_file():
-        # Copy the file if it doesn't exist
-        shutil.copy(source_path, target_path)
-        return f"Copied {filename} to {target_image_data_path}"
-    else:
-        return f"{filename} already exists in {target_image_data_path}"
-
-
-# apply the file copy and collect status information
-sampled_df["Image_FileName_DAPI_status"] = sampled_df["Image_FileName_DAPI"].apply(
-    check_and_copy_file
-)
-sampled_df["Image_FileName_GFP_status"] = sampled_df["Image_FileName_GFP"].apply(
-    check_and_copy_file
-)
-sampled_df["Image_FileName_RFP_status"] = sampled_df["Image_FileName_RFP"].apply(
-    check_and_copy_file
-)
-
-sampled_df
-
-# show the results using a mask on the dataframe for status
-print(
-    sampled_df[
-        [
-            "Image_FileName_DAPI_status",
-            "Image_FileName_GFP_status",
-            "Image_FileName_RFP_status",
-        ]
-    ]
-)
diff --git a/tests/data/cytotable/NF1_cellpainting_data_shrunken/create_mask_data.py b/tests/data/cytotable/NF1_cellpainting_data_shrunken/create_mask_data.py
index f16eee3..d360508 100644
--- a/tests/data/cytotable/NF1_cellpainting_data_shrunken/create_mask_data.py
+++ b/tests/data/cytotable/NF1_cellpainting_data_shrunken/create_mask_data.py
@@ -10,9 +10,9 @@
 tests/data/cytotable/NF1_cellpainting_data_shrunken/create_mask_data.py`
 """
 
-import subprocess
-import pathlib
 import os
+import pathlib
+import subprocess
 
 # create a dir for segmentation masks
 pathlib.Path("tests/data/cytotable/NF1_cellpainting_data_shrunken/Plate_2_masks").mkdir(