From 604f0596ad111811595505349b3d8090454f7087 Mon Sep 17 00:00:00 2001
From: Mathias Johansson <math.joh.bio@gmail.com>
Date: Tue, 13 Aug 2024 14:50:24 +0200
Subject: [PATCH] add new test

---
 .../assets/scripts/postprocess_gens_cnvkit.py |  9 ++++-----
 BALSAMIC/utils/io.py                          |  5 +++++
 tests/utils/test_utils.py                     | 20 +++++++++++++++++++
 3 files changed, 29 insertions(+), 5 deletions(-)

diff --git a/BALSAMIC/assets/scripts/postprocess_gens_cnvkit.py b/BALSAMIC/assets/scripts/postprocess_gens_cnvkit.py
index a50e354d0..2899091ff 100644
--- a/BALSAMIC/assets/scripts/postprocess_gens_cnvkit.py
+++ b/BALSAMIC/assets/scripts/postprocess_gens_cnvkit.py
@@ -1,5 +1,5 @@
 import click
-from BALSAMIC.utils.io import read_csv
+from BALSAMIC.utils.io import read_csv, write_list_of_strings
 
 def calculate_log2_ratio(purity, log2_ratio, ploidy):
     # Ensure that the inputs are within valid ranges
@@ -82,10 +82,9 @@ def create_gens_cov_file(
         log2_data.append(f"{row['chromosome']}\t{midpoint - 1}\t{midpoint}\t{log2}")
 
     # Write log2 data to output file
-    with open(output_file, "w") as log2_out:
-        for resolution in ["o", "a", "b", "c", "d"]:
-            for line in log2_data:
-                log2_out.write(f"{resolution}_{line}\n")
+    resolutions = ["o", "a", "b", "c", "d"]
+    resolution_log2_lines = [f"{resolution}_{line}" for resolution in resolutions for line in log2_data]
+    write_list_of_strings(resolution_log2_lines, output_file)
 
 
 if __name__ == "__main__":
diff --git a/BALSAMIC/utils/io.py b/BALSAMIC/utils/io.py
index 0b615265d..315899f93 100644
--- a/BALSAMIC/utils/io.py
+++ b/BALSAMIC/utils/io.py
@@ -73,6 +73,11 @@ def read_csv(csv_path: str, delimeter: str = ",") -> List[dict]:
         csv_reader = csv.DictReader(csv_file, delimiter=delimeter)
         return [row for row in csv_reader]
 
+def write_list_of_strings(list_of_strings: list[str], output_file: str):
+    """Writes a list of strings to a file, each on a new line."""
+    with open(output_file, 'w') as file:
+        for string in list_of_strings:
+            file.write(f"{string}\n")
 
 def read_yaml(yaml_path: str) -> dict:
     """Read data from a yaml file."""
diff --git a/tests/utils/test_utils.py b/tests/utils/test_utils.py
index 0cd721f78..503134c2e 100644
--- a/tests/utils/test_utils.py
+++ b/tests/utils/test_utils.py
@@ -49,6 +49,7 @@
     read_csv,
     read_vcf_file,
     read_yaml,
+    write_list_of_strings,
     write_finish_file,
     write_json,
     write_sacct_to_yaml,
@@ -455,6 +456,25 @@ def test_write_json(tmp_path, reference):
 
     assert len(list(tmp.iterdir())) == 1
 
+def test_write_list_of_strings(tmp_path):
+    """Test writing list of strings to file"""
+
+    # GIVEN a list of strings and an output file path
+    list_of_strings = [f"header1\theader2\theader3", f"row1_col1\trow1_col2\trow1_col3"]
+    tmp = tmp_path / "tmp"
+    tmp.mkdir()
+    output_file: Path = Path(tmp / "output.csv")
+
+    # WHEN writing list of strings
+    write_list_of_strings(list_of_strings, output_file.as_posix())
+
+    # THEN file should have been written
+    assert output_file.exists()
+
+    # AND contain the same information
+    read_written_file = read_csv(csv_path=output_file, delimeter="\t")
+    assert read_written_file == [{"header1": "row1_col1", "header2": "row1_col2", "header3": "row1_col3"}]
+    assert len(read_written_file) == 1
 
 def test_write_json_error(tmp_path: Path):
     """Test JSON write error."""