Merge pull request #4 from Samk13/update-pakcage-setup

Include CSV and YAML files to manifest

.github/workflows/tests.yml

@@ -0,0 +1,40 @@
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2022-2024 KTH Royal Institute of Technology.
+#
+# invenio-subjects-nasa is free software, you can redistribute it and/or
+# modify it under the terms of the MIT License; see LICENSE file details.
+
+name: CI
+
+on:
+  workflow_dispatch:
+    inputs:
+      reason:
+        description: 'Reason'
+        required: false
+        default: 'Manual trigger'
+
+jobs:
+  Tests:
+    runs-on: ubuntu-20.04
+    strategy:
+      matrix:
+          python-version: [3.9]
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+
+      - name: Generate dependencies
+        run: |
+          pip install -e ".[tests]"
+          pip freeze
+
+      - name: Run tests
+        run: |
+          ./run-tests.sh

MANIFEST.in

@@ -7,6 +7,8 @@ include .pylintrc
 include .vscode/settings.json
 include Makefile
 recursive-include tests
+recursive-include tests *.csv
+recursive-include tests *.yaml
 recursive-include invenio_subjects_nasa *.csv
 recursive-include invenio_subjects_nasa *.py
 recursive-include invenio_subjects_nasa *.yaml

invenio_subjects_nasa/config.py

@@ -0,0 +1,14 @@
+# -*- coding: utf-8 -*-
+#
+# Copyright (C) 2022-2024 KTH Royal Institute of Technology.
+#
+# invenio-subjects-nasa is free software, you can redistribute it and/or
+# modify it under the terms of the MIT License; see LICENSE file details.
+from pathlib import Path
+
+nasa_subjects_csv_input_path = (
+    Path.cwd() / "invenio_subjects_nasa" / "downloads" / "thesaurus-CSV-2024-02-05.csv"
+)
+nasa_subjects_yaml_output_path = (
+    Path.cwd() / "invenio_subjects_nasa" / "vocabularies" / "nasa_thesaurus.yaml"
+)

invenio_subjects_nasa/convert_data.py

@@ -99,15 +99,9 @@ def write_to_yaml(data: list, filepath: str):
         yaml.dump(data, file)
 
 
-def main(csv_filepath: str, yaml_filepath: str):
+def main(csv_filepath_input: str, yaml_filepath_output: str):
     """Main function to process CSV data and write it to a YAML file."""
-    data = load_csv_data(csv_filepath)
+    data = load_csv_data(csv_filepath_input)
     transformed_data = [transform_row_to_schema(row) for row in data]
     unique_data = remove_duplicates(transformed_data)
-    write_to_yaml(unique_data, yaml_filepath)
-
-
-if __name__ == "__main__":
-    csv_file_path = "invenio_subjects_nasa/downloads/thesaurus-CSV-2024-02-05.csv"
-    yaml_file_path = "invenio_subjects_nasa/vocabularies/nasa_thesaurus.yaml"
-    main(csv_file_path, yaml_file_path)
+    write_to_yaml(unique_data, yaml_filepath_output)

main.py

@@ -6,23 +6,15 @@
 # modify it under the terms of the MIT License; see LICENSE file details.
 """Nasa subject terms for InvenioRDM"""
 
-from invenio_subjects_nasa.converter import csv_to_yaml, write_to_disk
-from invenio_subjects_nasa.load_data import load_data
+from invenio_subjects_nasa.config import (
+    nasa_subjects_csv_input_path,
+    nasa_subjects_yaml_output_path,
+)
+from invenio_subjects_nasa.convert_data import main
 from invenio_subjects_nasa.utils import logger
 
-row_data = "invenio_subjects_nasa/downloads/thesaurus-CSV.csv"
-output_yaml = "invenio_subjects_nasa/vocabularies/nasa_thesaurus.yaml"
-
-
-def main():
-    """initiate the app"""
-    logger("Converting ...")
-    res = []
-    for row in load_data(row_data):
-        res.append(csv_to_yaml(row))
-    write_to_disk(res, output_yaml)
-    logger("Subjects is been converted successfully!")
-
-
 if __name__ == "__main__":
-    main()
+
+    logger.debug("Converting ...")
+    main(nasa_subjects_csv_input_path, nasa_subjects_yaml_output_path)
+    logger.debug("Subjects is been converted successfully!")

tests/test_config.py

@@ -0,0 +1,20 @@
+from pathlib import Path
+
+from invenio_subjects_nasa import config
+
+
+def test_nasa_subjects_csv_input_path():
+    """Test the default value of the NASA subjects CSV input path."""
+    assert config.nasa_subjects_csv_input_path == (
+        Path.cwd()
+        / "invenio_subjects_nasa"
+        / "downloads"
+        / "thesaurus-CSV-2024-02-05.csv"
+    )
+
+
+def test_nasa_subjects_yaml_output_path():
+    """Test the default value of the NASA subjects YAML output path."""
+    assert config.nasa_subjects_yaml_output_path == (
+        Path.cwd() / "invenio_subjects_nasa" / "vocabularies" / "nasa_thesaurus.yaml"
+    )