added tests for a file list and edited help message addressing the rules for inputing a file list

Author: yucongalicechen

src/diffpy/labpdfproc/labpdfprocapp.py

@@ -17,7 +17,13 @@
 def get_args(override_cli_inputs=None):
     p = ArgumentParser()
     p.add_argument("mud", help="Value of mu*D for your " "sample. Required.", type=float)
-    p.add_argument("input", help="The filename or directory of the datafile to load. Required.")
+    p.add_argument(
+        "input",
+        help="The filename or directory of the datafile to load. Required. "
+        "Supports either a single input file or directory, or a file containing a list of files. "
+        "If providing a file list, please ensure all files are in the same directory as the file list, "
+        "and each filename is written line by line in the file list. ",
+    )
     p.add_argument(
         "-a",
         "--anode-type",

src/diffpy/labpdfproc/tests/conftest.py

@@ -39,4 +39,13 @@ def user_filesystem(tmp_path):
     with open(os.path.join(input_dir, "binary.pkl"), "wb") as f:
         f.write(binary_data)
 
+    file_list_dir = Path(tmp_path).resolve() / "file_list_dir"
+    file_list_dir.mkdir(parents=True, exist_ok=True)
+    with open(os.path.join(file_list_dir, "file_list.txt"), "w") as f:
+        f.write("good_data.chi \n good_data.xy \n good_data.txt")
+    with open(os.path.join(file_list_dir, "invalid_file_list.txt"), "w") as f:
+        f.write("good_data.chi \n non_existing_file.xy \n non_existing_file.txt")
+    with open(os.path.join(file_list_dir, "invalid_format_file_list.txt"), "w") as f:
+        f.write("good_data.chi good_data.xy \n non_existing_file.txt")
+
     yield tmp_path

src/diffpy/labpdfproc/tests/test_tools.py

@@ -1,3 +1,4 @@
+import os
 import re
 from pathlib import Path
 
@@ -15,7 +16,7 @@
 
 # Use cases can be found here: https://github.com/diffpy/diffpy.labpdfproc/issues/48
 
-# This test covers existing single input file or directory
+# This test covers existing single input file, directory, or a file list
 # We store absolute path into input_directory and file names into input_file
 params_input = [
     (["good_data.chi"], [".", "good_data.chi"]),
@@ -42,6 +43,7 @@
             ["good_data.chi", "good_data.xy", "good_data.txt", "unreadable_file.txt", "binary.pkl"],
         ],
     ),
+    (["file_list_dir/file_list.txt"], ["file_list_dir", ["good_data.chi", "good_data.xy", "good_data.txt"]]),
 ]
 
 
@@ -57,6 +59,28 @@ def test_set_input_files(inputs, expected, user_filesystem):
     assert set(actual_args.input_file) == set(expected_input_files)
 
 
+# This test is for existing single input file or directory absolute path not in cwd
+# Here we are in user_filesystem/input_dir, testing for a file or directory in user_filesystem
+params_input_not_cwd = [
+    (["good_data.chi"], [".", "good_data.chi"]),
+    (["."], [".", ["good_data.chi", "good_data.xy", "good_data.txt", "unreadable_file.txt", "binary.pkl"]]),
+]
+
+
+@pytest.mark.parametrize("inputs, expected", params_input_not_cwd)
+def test_set_input_files_not_cwd(inputs, expected, user_filesystem):
+    expected_input_directory = Path(user_filesystem) / expected[0]
+    expected_input_files = expected[1]
+    actual_input = [str(Path(user_filesystem) / inputs[0])]
+    os.chdir("input_dir")
+
+    cli_inputs = ["2.5"] + actual_input
+    actual_args = get_args(cli_inputs)
+    actual_args = set_input_files(actual_args)
+    assert actual_args.input_directory == expected_input_directory
+    assert set(actual_args.input_file) == set(expected_input_files)
+
+
 # This test covers non-existing single input file or directory, in this case we raise an error with message
 params_input_bad = [
     (["non_existing_file.xy"], "Please specify valid input file or directory."),

src/diffpy/labpdfproc/tools.py

@@ -15,21 +15,42 @@ def set_input_files(args):
     args argparse.Namespace
         the arguments from the parser
 
+    It is implemented as this:
+    If input is a file, we first try to read it as a file list and store all listed file names.
+    If any filename is invalid, then proceed to treat it as a data file.
+    Otherwise if we have a directory, glob all files within it.
+
     Returns
     -------
     args argparse.Namespace
 
     """
+
     if not Path(args.input).exists():
         raise ValueError("Please specify valid input file or directory.")
 
     if not Path(args.input).is_dir():
         input_dir = Path.cwd() / Path(args.input).parent
-        input_file_name = Path(args.input).name
+        file_names = []
+        with open(args.input, "r") as f:
+            for line in f:
+                if not os.path.isfile(line.strip()):
+                    file_names = []
+                    break
+                else:
+                    file_name = line.strip()
+                    file_names.append(file_name)
+
+        if len(file_names) > 0:
+            input_file_name = file_names
+        else:
+            input_file_name = Path(args.input).name
+
     else:
         input_dir = Path(args.input).resolve()
         input_files = [file for file in glob.glob(str(input_dir) + "/*", recursive=True) if os.path.isfile(file)]
         input_file_name = [os.path.basename(input_file_path) for input_file_path in input_files]
+
     setattr(args, "input_directory", input_dir)
     setattr(args, "input_file", input_file_name)
     return args