From bb41c55622d160c321e491646c131221d46cb323 Mon Sep 17 00:00:00 2001
From: lauraporta <29216006+lauraporta@users.noreply.github.com>
Date: Fri, 8 Nov 2024 15:48:35 +0000
Subject: [PATCH] Change usage of file pattern in reader

---
 calcium_imaging_automation/core/reader.py | 14 ++++++--------
 examples/example_usage.py                 | 15 ++-------------
 examples/example_usage.sh                 |  1 +
 3 files changed, 9 insertions(+), 21 deletions(-)

diff --git a/calcium_imaging_automation/core/reader.py b/calcium_imaging_automation/core/reader.py
index f1ccf9e..65d1630 100644
--- a/calcium_imaging_automation/core/reader.py
+++ b/calcium_imaging_automation/core/reader.py
@@ -6,11 +6,9 @@ class ReadAllPathsInFolder:
     def __init__(
         self,
         raw_data_folder: Path,
-        filetypes_of_interest: List[str],
         folder_read_pattern: str,
-        file_read_pattern: str,
+        file_read_pattern: list,
     ):
-        self.filetypes_of_interest = filetypes_of_interest
         self.folder_read_pattern = folder_read_pattern
         self.file_read_pattern = file_read_pattern
 
@@ -25,19 +23,19 @@ def get_folders_first_layer(self, file_path: Path) -> List[Path]:
     def get_files_paths(self, folder: Path) -> List[Path]:
         return [
             file
-            for filetype in self.filetypes_of_interest
-            for file in folder.rglob(f"*.{filetype}")
+            for pattern in self.folder_read_pattern
+            for file in folder.rglob(pattern)
         ]
 
-    def total_objects_by_filetype(self, folder: Path) -> dict:
+    def total_objects_by_format(self, folder: Path) -> dict:
         return {
             filetype: len(self.get_files_paths(folder))
-            for filetype in self.filetypes_of_interest
+            for filetype in self.folder_read_pattern.split(".")[-1]
         }
 
     def max_session_number(self, filetype="tif", max_allowed=5) -> int:
         total_tif_number = [
-            self.total_objects_by_filetype(dataset_path).get(filetype, 0)
+            self.total_objects_by_format(dataset_path).get(filetype, 0)
             for dataset_path in self.datasets_paths
         ]
 
diff --git a/examples/example_usage.py b/examples/example_usage.py
index 616138b..d3a48b5 100644
--- a/examples/example_usage.py
+++ b/examples/example_usage.py
@@ -9,7 +9,6 @@
 def main(
     raw_data_path: Path,
     output_path: Path,
-    filetypes_of_interest: list,
     folder_read_pattern: str,
     file_read_pattern: str,
 ):
@@ -25,7 +24,6 @@ def main(
 
     reader = ReadAllPathsInFolder(
         raw_data_path,
-        filetypes_of_interest,
         folder_read_pattern,
         file_read_pattern,
     )
@@ -53,13 +51,6 @@ def main(
     parser.add_argument(
         "output_path", type=Path, help="Path to the output data."
     )
-    parser.add_argument(
-        "--filetypes",
-        type=list,
-        nargs="+",
-        help="Filetypes of interest.",
-        default=["tif", "bin"],
-    )
     parser.add_argument(
         "--folder_read_pattern",
         type=str,
@@ -68,22 +59,20 @@ def main(
     )
     parser.add_argument(
         "--file_read_pattern",
-        type=str,
+        type=list,
         help="Glob pattern for reading files.",
-        default="*",
+        default=["*.tif", "*.bin"],
     )
 
     args = parser.parse_args()
     raw_data_path = args.raw_data_path
     output_path = args.output_path
-    file_types = args.filetypes
     folder_read_pattern = args.folder_read_pattern
     file_read_pattern = args.file_read_pattern
 
     main(
         raw_data_path,
         output_path,
-        file_types,
         folder_read_pattern,
         file_read_pattern,
     )
diff --git a/examples/example_usage.sh b/examples/example_usage.sh
index 890024f..9998c69 100755
--- a/examples/example_usage.sh
+++ b/examples/example_usage.sh
@@ -3,3 +3,4 @@ python ./examples/example_usage.py \
     /nfs/winstor/margrie/SimonWeiler/RawData/Invivo_imaging/3photon_rotation/shared/ \
     /ceph/margrie/laura/cimaut/ \
     --folder_read_pattern '2*' \
+    --file_read_pattern 'rotation_00001.tif, *.bin' \