feat: Adding Github Unittests workflow

.github/workflows/unit-tests.yml

@@ -0,0 +1,26 @@
+name: Unit Tests
+
+on: [push, pull_request]
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+
+    steps:
+    - name: Checkout code
+      uses: actions/checkout@v4
+
+    - name: Set up Python
+      id: setup-python
+      uses: actions/setup-python@v5
+      with:
+        python-version: '3.11'
+
+    - name: Install dependencies
+      run: |
+        python -m pip install --upgrade pip
+        pip install pylint poetry
+        poetry install
+
+    - name: Run tests
+      run: poetry run pytest

README.md

@@ -6,6 +6,8 @@
 A command-line tool designed to detect and purge any and all macros and dynamic content from commonly used office document formats (including MS Office Files, PDFs, etc.).
 
 [![Super-Linter](https://github.com/Anti-Malware-Alliance/ms-office-macro-bleach/actions/workflows/pre-commit-actions.yml/badge.svg)](https://github.com/marketplace/actions/super-linter)
+![Unit Tests](https://github.com/Anti-Malware-Alliance/ms-office-macro-bleach/actions/workflows/unit-tests.yml/badge.svg)
+
 
 ## Supported formats
 

tests/test_file_validation.py

@@ -14,57 +14,75 @@
 
 All tests are written for and conducted using pytest.
 """
-
-from subprocess import check_output
-from os import remove, rename
+import unittest
+from subprocess import check_output, CalledProcessError
+import os
 from shutil import copyfile
 
-program_dir = "../docubleach/"
-
-
-def test_valid_file_with_macros():
-    copyfile("test_files/valid_file_with_macros.docm", "test_files/valid_file_with_macros.bak")
-
-    output = check_output(f"python {program_dir}bleach.py test_files/valid_file_with_macros.docm", encoding='utf-8')
-
-    remove("test_files/valid_file_with_macros.docm")
-    rename("test_files/valid_file_with_macros.bak", "test_files/valid_file_with_macros.docm")
-
-    assert output == ""
-
-
-def test_valid_file_with_macros_with_check():
-    copyfile("test_files/valid_file_with_macros_check.docm", "test_files/valid_file_with_macros_check.bak")
-
-    output = check_output(f"python {program_dir}bleach.py test_files/valid_file_with_macros_check.docm -c",
-                          encoding='utf-8')
-
-    remove("test_files/valid_file_with_macros_check.docm")
-    rename("test_files/valid_file_with_macros_check.bak", "test_files/valid_file_with_macros_check.docm")
-
-    assert output == "Macros detected and removed.\n"
-
-
-def test_valid_file_without_macros():
-    output = check_output(f"python {program_dir}bleach.py test_files/valid_file_without_macros.docx", encoding='utf-8')
-
-    assert output == ""
-
-
-def test_invalid_file_type():
-    output = check_output(f"python {program_dir}bleach.py test_files/invalid_file_type.txt", encoding='utf-8')
-
-    assert output == "Unsupported file format.\n"
-
-
-def test_invalid_file_size():
-
-    # Create temporary file exceeding 200MB limit
-    with open("test_files/invalid_file_size.docx", "wb") as out:
-        out.truncate(262144000)
-
-    output = check_output(f"python {program_dir}bleach.py test_files/invalid_file_size.docx", encoding='utf-8')
-
-    remove("test_files/invalid_file_size.docx")
 
-    assert output == "File exceeds size limit.\n"
+class TestFileValidation(unittest.TestCase):
+    @classmethod
+    def setUpClass(cls):
+        cls.cur_dir = os.getcwd()
+        cls.test_dir = os.path.join(cls.cur_dir, 'tests')
+        cls.test_resources_dir = os.path.join(cls.test_dir, 'test_files')
+        cls.program_dir = os.path.join(cls.cur_dir, "docubleach", "bleach.py")
+
+        # Verify the existence of the paths
+        if not os.path.exists(cls.program_dir):
+            raise FileNotFoundError(f"Program file not found: {cls.program_dir}")
+
+    def run_test(self, filename, expected_output, run_check=False):
+        try:
+            if run_check:
+                output = check_output(
+                    ["python", self.program_dir, os.path.join(self.test_resources_dir, filename), "-c"],
+                    encoding='utf-8'
+                )
+            else:
+                output = check_output(
+                    ["python", self.program_dir, os.path.join(self.test_resources_dir, filename)],
+                    encoding='utf-8'
+                )
+
+            self.assertEqual(output, expected_output)
+        except CalledProcessError as e:
+            self.fail(f"Subprocess failed with error: {e}")
+
+    def run_macro_test(self, filename, expected_output, run_check=False):
+        backup_file = os.path.join(self.test_resources_dir, f"{filename}.bak")
+        test_file = os.path.join(self.test_resources_dir, filename)
+
+        copyfile(test_file, backup_file)
+        try:
+            self.run_test(filename, expected_output, run_check)
+        finally:
+            os.remove(test_file)
+            os.rename(backup_file, test_file)
+
+    def test_valid_file_with_macros(self):
+        self.run_macro_test("valid_file_with_macros.docm", "")
+
+    def test_valid_file_with_macros_with_check(self):
+        self.run_macro_test("valid_file_with_macros_check.docm",
+                            "Macros detected and removed.\n", True)
+
+    def test_valid_file_without_macros(self):
+        self.run_macro_test("valid_file_without_macros.docx", "")
+
+    def test_invalid_file_type(self):
+        self.run_macro_test("invalid_file_type.txt", "Unsupported file format.\n")
+
+    def test_invalid_file_size(self):
+        try:
+            # Create temporary file exceeding 200MB limit
+            with open(os.path.join(self.test_resources_dir, "invalid_file_size.docx"), "wb") as out:
+                out.truncate(262144000)
+
+            self.run_macro_test("invalid_file_size.docx", "File exceeds size limit.\n")
+        finally:
+            os.remove(os.path.join(self.test_resources_dir, "invalid_file_size.docx"))
+
+
+if __name__ == '__main__':
+    unittest.main()

tests/test_ooxml_files.py

@@ -8,144 +8,87 @@
 
 All tests are written for and conducted using pytest.
 """
-from subprocess import check_output
-from os import remove, rename
+import unittest
+import os
+from subprocess import check_output, CalledProcessError
 from shutil import copyfile
 
 
-program_dir = "../docubleach/"
+class OOXMLFilesTest(unittest.TestCase):
+    @classmethod
+    def setUpClass(cls):
+        cls.cur_dir = os.getcwd()
+        cls.test_dir = os.path.join(cls.cur_dir, 'tests')
+        cls.test_resources_dir = os.path.join(cls.test_dir, 'test_files')
+        cls.program_dir = os.path.join(cls.cur_dir, "docubleach", "bleach.py")
 
+        # Verify the existence of the paths
+        if not os.path.exists(cls.program_dir):
+            raise FileNotFoundError(f"Program file not found: {cls.program_dir}")
 
-def test_word_document():
-    output = check_output(f"python {program_dir}bleach.py test_files/word_document.docx -c", encoding='utf-8')
+    def run_test(self, filename, expected_output):
+        try:
+            output = check_output(
+                ["python", self.program_dir, os.path.join(self.test_resources_dir, filename), "-c"],
+                encoding='utf-8'
+            )
+            self.assertEqual(output, expected_output)
+        except CalledProcessError as e:
+            self.fail(f"Subprocess failed with error: {e}")
 
-    assert output == ""
+    def run_macro_test(self, filename, expected_output):
+        backup_file = os.path.join(self.test_resources_dir, f"{filename}.bak")
+        test_file = os.path.join(self.test_resources_dir, filename)
 
+        copyfile(test_file, backup_file)
+        try:
+            self.run_test(filename, expected_output)
+        finally:
+            os.remove(test_file)
+            os.rename(backup_file, test_file)
 
-def test_word_document_with_macros():
-    copyfile("test_files/word_document_with_macros.docm", "test_files/word_document_with_macros.docm.bak")
+    def test_word_document(self):
+        self.run_test('word_document.docx', "")
 
-    output = check_output(f"python {program_dir}bleach.py test_files/word_document_with_macros.docm -c",
-                          encoding='utf-8')
+    def test_word_document_with_macros(self):
+        self.run_macro_test('word_document_with_macros.docm', "Macros detected and removed.\n")
 
-    remove("test_files/word_document_with_macros.docm")
-    rename("test_files/word_document_with_macros.docm.bak", "test_files/word_document_with_macros.docm")
+    def test_word_template(self):
+        self.run_test('word_template.dotx', "")
 
-    assert output == "Macros detected and removed.\n"
+    def test_word_template_with_macros(self):
+        self.run_macro_test('word_template_with_macros.dotm', "Macros detected and removed.\n")
 
+    def test_powerpoint_presentation(self):
+        self.run_test('powerpoint_presentation.pptx', "")
 
-def test_word_template():
-    output = check_output(f"python {program_dir}bleach.py test_files/word_template.dotx -c", encoding='utf-8')
+    def test_powerpoint_presentation_with_macros(self):
+        self.run_macro_test('powerpoint_presentation_with_macros.pptm', "Macros detected and removed.\n")
 
-    assert output == ""
+    def test_powerpoint_template(self):
+        self.run_test('powerpoint_template.potx', "")
 
+    def test_powerpoint_template_with_macros(self):
+        self.run_macro_test('powerpoint_template_with_macros.potm', "Macros detected and removed.\n")
 
-def test_word_template_with_macros():
-    copyfile("test_files/word_template_with_macros.dotm", "test_files/word_template_with_macros.dotm.bak")
+    def test_powerpoint_show(self):
+        self.run_test('powerpoint_show.ppsx', "")
 
-    output = check_output(f"python {program_dir}bleach.py test_files/word_template_with_macros.dotm -c",
-                          encoding='utf-8')
+    def test_powerpoint_show_with_macros(self):
+        self.run_macro_test('powerpoint_show_with_macros.ppsm', "Macros detected and removed.\n")
 
-    remove("test_files/word_template_with_macros.dotm")
-    rename("test_files/word_template_with_macros.dotm.bak", "test_files/word_template_with_macros.dotm")
+    def test_excel_spreadsheet(self):
+        self.run_test('excel_spreadsheet.xlsx', "")
 
-    assert output == "Macros detected and removed.\n"
+    def test_excel_spreadsheet_with_macros(self):
+        self.run_macro_test('excel_spreadsheet_with_macros.xlsm', "Macros detected and removed.\n")
 
+    def test_excel_template(self):
+        self.run_test('excel_template.xltx', "")
 
-def test_powerpoint_presentation():
-    output = check_output(f"python {program_dir}bleach.py test_files/powerpoint_presentation.pptx -c",
-                          encoding='utf-8')
+    def test_excel_template_with_macros(self):
+        self.run_macro_test('excel_template_with_macros.xltm', "Macros detected and removed.\n")
 
-    assert output == ""
 
-
-def test_powerpoint_presentation_with_macros():
-    copyfile("test_files/powerpoint_presentation_with_macros.pptm",
-             "test_files/powerpoint_presentation_with_macros.pptm.bak")
-
-    output = check_output(f"python {program_dir}bleach.py test_files/powerpoint_presentation_with_macros.pptm -c",
-                          encoding='utf-8')
-
-    remove("test_files/powerpoint_presentation_with_macros.pptm")
-    rename("test_files/powerpoint_presentation_with_macros.pptm.bak",
-           "test_files/powerpoint_presentation_with_macros.pptm")
-
-    assert output == "Macros detected and removed.\n"
-
-
-def test_powerpoint_template():
-    output = check_output(f"python {program_dir}bleach.py test_files/powerpoint_template.potx -c",
-                          encoding='utf-8')
-
-    assert output == ""
-
-
-def test_powerpoint_template_with_macros():
-    copyfile("test_files/powerpoint_template_with_macros.potm",
-             "test_files/powerpoint_template_with_macros.potm.bak")
-
-    output = check_output(f"python {program_dir}bleach.py test_files/powerpoint_template_with_macros.potm -c",
-                          encoding='utf-8')
-
-    remove("test_files/powerpoint_template_with_macros.potm")
-    rename("test_files/powerpoint_template_with_macros.potm.bak",
-           "test_files/powerpoint_template_with_macros.potm")
-
-    assert output == "Macros detected and removed.\n"
-
-
-def test_powerpoint_show():
-    output = check_output(f"python {program_dir}bleach.py test_files/powerpoint_show.ppsx -c",
-                          encoding='utf-8')
-
-    assert output == ""
-
-
-def test_powerpoint_show_with_macros():
-    copyfile("test_files/powerpoint_show_with_macros.ppsm", "test_files/powerpoint_show_with_macros.ppsm.bak")
-
-    output = check_output(f"python {program_dir}bleach.py test_files/powerpoint_show_with_macros.ppsm -c",
-                          encoding='utf-8')
-
-    remove("test_files/powerpoint_show_with_macros.ppsm")
-    rename("test_files/powerpoint_show_with_macros.ppsm.bak", "test_files/powerpoint_show_with_macros.ppsm")
-
-    assert output == "Macros detected and removed.\n"
-
-
-def test_excel_spreadsheet():
-    output = check_output(f"python {program_dir}bleach.py test_files/excel_spreadsheet.xlsx -c",
-                          encoding='utf-8')
-
-    assert output == ""
-
-
-def test_excel_spreadsheet_with_macros():
-    copyfile("test_files/excel_spreadsheet_with_macros.xlsm",
-             "test_files/excel_spreadsheet_with_macros.xlsm.bak")
-
-    output = check_output(f"python {program_dir}bleach.py test_files/excel_spreadsheet_with_macros.xlsm -c",
-                          encoding='utf-8')
-
-    remove("test_files/excel_spreadsheet_with_macros.xlsm")
-    rename("test_files/excel_spreadsheet_with_macros.xlsm.bak", "test_files/excel_spreadsheet_with_macros.xlsm")
-
-    assert output == "Macros detected and removed.\n"
-
-
-def test_excel_template():
-    output = check_output(f"python {program_dir}bleach.py test_files/excel_template.xltx -c", encoding='utf-8')
-
-    assert output == ""
-
-
-def test_excel_template_with_macros():
-    copyfile("test_files/excel_template_with_macros.xltm", "test_files/excel_template_with_macros.xltm.bak")
-
-    output = check_output(f"python {program_dir}bleach.py test_files/excel_template_with_macros.xltm -c",
-                          encoding='utf-8')
-
-    remove("test_files/excel_template_with_macros.xltm")
-    rename("test_files/excel_template_with_macros.xltm.bak", "test_files/excel_template_with_macros.xltm")
-
-    assert output == "Macros detected and removed.\n"
+if __name__ == '__main__':
+    unittest.main()