Merge pull request #231 from nasa/feature/issue-230-allow-single-file…

…-netcdf-filepath-input

Feature/issue 230 allow single file netcdf filepath input

CHANGELOG.md

@@ -4,6 +4,12 @@ All notable changes to this project will be documented in this file.
 The format is based on [Common Changelog](https://common-changelog.org/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## Unreleased
+
+### Changed
+- Allow single netCDF file input in addition to single text file listings ([#230](https://github.com/nasa/stitchee/issues/230))([**@danielfromearth**](https://github.com/danielfromearth))
+
+
 ## [1.3.0] - 2024-07-11
 
 ### Changed

concatenator/file_ops.py

@@ -8,6 +8,8 @@
 
 logger = logging.getLogger(__name__)
 
+netcdf_extensions = [".nc", ".nc4", ".netcdf"]
+
 
 def add_label_to_path(x: str, label="_flat_intermediate") -> str:
     """Constructs new filepath with label at end."""
@@ -55,7 +57,15 @@ def validate_output_path(filepath: str, overwrite: bool = False) -> str:
 
 
 def validate_input_path(path_or_paths: list[str]) -> list[str]:
-    """Checks whether input is a valid directory, list of files, or a text file containing paths."""
+    """Checks whether input is a list of files, a directory, or a text file containing paths.
+
+    If the input is...
+    - a list of filepaths, then use those filepaths.
+    - a valid directory, then get the paths for all the files in the directory.
+    - a single file:
+        - that is a valid text file, then get the names of the files from each row in the text file.
+        - that is a valid netCDF file, then use that one filepath
+    """
     print(f"parsed_input === {path_or_paths}")
     if len(path_or_paths) > 1:
         input_files = path_or_paths
@@ -64,7 +74,10 @@ def validate_input_path(path_or_paths: list[str]) -> list[str]:
         if directory_or_path.is_dir():
             input_files = _get_list_of_filepaths_from_dir(directory_or_path)
         elif directory_or_path.is_file():
-            input_files = _get_list_of_filepaths_from_file(directory_or_path)
+            if directory_or_path.suffix in netcdf_extensions:
+                input_files = [str(directory_or_path)]
+            else:
+                input_files = _get_list_of_filepaths_from_file(directory_or_path)
         else:
             raise TypeError(
                 "If one path is provided for 'data_dir_or_file_or_filepaths', "

tests/conftest.py

@@ -8,6 +8,11 @@
 import numpy as np
 import pytest
 
+test_path = Path(__file__).parents[0].resolve()
+data_path = test_path.joinpath("data")
+harmony_path = data_path.joinpath("harmony")
+granules_path = harmony_path.joinpath("granules")
+
 
 class DataDirs(typing.NamedTuple):
     test_path: Path
@@ -160,3 +165,26 @@ def ds_3dims_3vars_4coords_1group_part3(temp_toy_data_dir):
     f.close()
 
     return filepath
+
+
+@pytest.fixture(scope="function")
+def text_file_with_three_paths(temp_toy_data_dir) -> Path:
+    filepath = temp_toy_data_dir / "text_file_with_paths.txt"
+
+    paths = [
+        path_str(granules_path, x)
+        for x in [
+            "TEMPO_NO2_L2_V03_20240601T210934Z_S012G01_subsetted.nc4",
+            "TEMPO_NO2_L2_V03_20240601T211614Z_S012G02_subsetted.nc4",
+            "TEMPO_NO2_L2_V03_20240601T212254Z_S012G03_subsetted.nc4",
+        ]
+    ]
+
+    contents = f"""{paths[0]}
+{paths[1]}
+{paths[2]}
+"""
+    with open(filepath, "w") as f:
+        f.write(contents)
+
+    return filepath

tests/unit/test_run_stitchee.py

@@ -1,13 +1,12 @@
 import sys
-from pathlib import Path
 from unittest.mock import patch
 
 import pytest
 
 import concatenator
 from concatenator.run_stitchee import parse_args
 
-from ..conftest import path_str
+from ..conftest import granules_path, path_str
 
 
 def test_parser():
@@ -24,23 +23,13 @@ def test_parser():
 
 @pytest.mark.usefixtures("pass_options")
 class TestBatching:
-    __test_path = Path(__file__).parents[1].resolve()
-    __data_path = __test_path.joinpath("data")
-    __harmony_path = __data_path.joinpath("harmony")
-    __granules_path = __harmony_path.joinpath("granules")
 
     def test_run_stitchee_cli_with_three_filepaths(self, temp_output_dir):
         test_args = [
             concatenator.run_stitchee.__file__,
-            path_str(
-                self.__granules_path, "TEMPO_NO2_L2_V03_20240601T210934Z_S012G01_subsetted.nc4"
-            ),
-            path_str(
-                self.__granules_path, "TEMPO_NO2_L2_V03_20240601T211614Z_S012G02_subsetted.nc4"
-            ),
-            path_str(
-                self.__granules_path, "TEMPO_NO2_L2_V03_20240601T212254Z_S012G03_subsetted.nc4"
-            ),
+            path_str(granules_path, "TEMPO_NO2_L2_V03_20240601T210934Z_S012G01_subsetted.nc4"),
+            path_str(granules_path, "TEMPO_NO2_L2_V03_20240601T211614Z_S012G02_subsetted.nc4"),
+            path_str(granules_path, "TEMPO_NO2_L2_V03_20240601T212254Z_S012G03_subsetted.nc4"),
             "--copy_input_files",
             "--verbose",
             "-o",
@@ -57,7 +46,43 @@ def test_run_stitchee_cli_with_three_filepaths(self, temp_output_dir):
     def test_run_stitchee_cli_with_one_directorypath(self, temp_output_dir):
         test_args = [
             concatenator.run_stitchee.__file__,
-            str(self.__granules_path),
+            str(granules_path),
+            "--copy_input_files",
+            "--verbose",
+            "-o",
+            path_str(temp_output_dir, "test_run_stitchee_output.nc"),
+            "--concat_method",
+            "xarray-concat",
+            "--concat_dim",
+            "mirror_step",
+        ]
+
+        with patch.object(sys, "argv", test_args):
+            concatenator.run_stitchee.main()
+
+    def test_run_stitchee_cli_with_one_netCDFpath(self, temp_output_dir):
+        test_args = [
+            concatenator.run_stitchee.__file__,
+            path_str(granules_path, "TEMPO_NO2_L2_V03_20240601T210934Z_S012G01_subsetted.nc4"),
+            "--copy_input_files",
+            "--verbose",
+            "-o",
+            path_str(temp_output_dir, "test_run_stitchee_output.nc"),
+            "--concat_method",
+            "xarray-concat",
+            "--concat_dim",
+            "mirror_step",
+        ]
+
+        with patch.object(sys, "argv", test_args):
+            concatenator.run_stitchee.main()
+
+    def test_run_stitchee_cli_with_one_path_to_text_listing_of_three_files(
+        self, temp_output_dir, text_file_with_three_paths
+    ):
+        test_args = [
+            concatenator.run_stitchee.__file__,
+            str(text_file_with_three_paths),
             "--copy_input_files",
             "--verbose",
             "-o",