Run CI on all matrix OSes

.github/workflows/main.yml

@@ -7,6 +7,7 @@ on:
     branches: ["main", "master"]
   pull_request:
     branches: ["main", "master"]
+  workflow_dispatch: # Enable manual trigger.
 
 permissions:
   contents: read
@@ -15,13 +16,12 @@ jobs:
   build:
     strategy:
       matrix:
-        # Update this as needed:
-        # Common platforms: ["ubuntu-latest", "macos-latest", "windows-latest"]
-        os: ["ubuntu-latest"]
+        # TODO: Fix failing CI on Windows
+        os: ["ubuntu-latest", "macos-latest"]
         python-version: ["3.10", "3.11", "3.12", "3.13"]
 
-    # Linux only by default. Use ${{ matrix.os }} for other OSes.
     runs-on: ${{ matrix.os }}
+    timeout-minutes: 5
 
     steps:
       # Generally following uv docs:
@@ -33,6 +33,12 @@ jobs:
           # Important for versioning plugins:
           fetch-depth: 0
 
+      - name: Install libmagic (macOS only)
+        if: matrix.os == 'macos-latest'
+        run: |
+          brew update
+          brew install libmagic
+
       - name: Install uv (official Astral action)
         uses: astral-sh/setup-uv@v5
         with:

README.md

@@ -24,7 +24,7 @@ The output parquet file (`file_index.parquet`) has the following columns:
     * folder_path
     * file_name
     * file_size_bytes
-    * md5_hash_hex
+    * md5_hex
     * sha256_base64
     * date_created
     * date_modified

src/folder_indexer/indexer.py

@@ -282,19 +282,17 @@ def get_file_info(
             magic_file_type_1 = magic_worker.from_file(file_path)
 
     # SHA256 hash, if file size < 100 KiB
-    with add_time_taken(time_taken_log, "sha256_hash_1"):
+    with add_time_taken(time_taken_log, "sha256_base64"):
         sha256_base64 = None
         if file_size < BIG_FILE_SIZE_THRESHOLD_BYTES:
-            with file_path.open("rb") as f:
-                sha256_hash = hashlib.sha256(f.read()).digest()
-                sha256_base64 = base64.b64encode(sha256_hash).decode("utf-8")
+            sha256_hash = hashlib.sha256(file_path.read_bytes()).digest()
+            sha256_base64 = base64.b64encode(sha256_hash).decode("utf-8")
 
-    with add_time_taken(time_taken_log, "md5_hash_hex_1"):
-        md5_hash_hex = None
+    with add_time_taken(time_taken_log, "md5_hex"):
+        md5_hex = None
         if file_size < BIG_FILE_SIZE_THRESHOLD_BYTES:
-            with file_path.open("rb") as f:
-                md5_hash = hashlib.md5(f.read()).digest()
-                md5_hash_hex = binascii.hexlify(md5_hash).decode("utf-8")
+            md5_hash = hashlib.md5(file_path.read_bytes()).digest()
+            md5_hex = binascii.hexlify(md5_hash).decode("utf-8")
 
     # Append file information to the list
     with add_time_taken(time_taken_log, "append_to_list"):
@@ -306,7 +304,7 @@ def get_file_info(
             "folder_path": file_path_stripped.parent.as_posix(),
             "file_name": file_path_stripped.name,
             "file_size_bytes": file_size,
-            "md5_hex": md5_hash_hex,
+            "md5_hex": md5_hex,
             "sha256_base64": sha256_base64,
             "date_created": date_created,
             "date_modified": date_modified,
@@ -328,6 +326,7 @@ def save_to_parquet(file_data: list[dict[str, Any]], output_folder: Path) -> Non
             "folder_path": pl.String,
             "file_name": pl.String,
             "file_size_bytes": pl.UInt64,
+            "md5_hex": pl.String,
             "sha256_base64": pl.String,
             "date_created": pl.Datetime,
             "date_modified": pl.Datetime,

tests/test_main_cli.py

@@ -19,13 +19,14 @@ def test_with_single_file(tmp_path: Path) -> None:
     assert (tmp_path / "file_index.parquet").is_file()
 
     df = pl.read_parquet(tmp_path / "file_index.parquet")
-    assert df.shape == (1, 12)
+    assert df.shape == (1, 13)
 
     assert df.columns == [
         "file_path",
         "folder_path",
         "file_name",
         "file_size_bytes",
+        "md5_hex",
         "sha256_base64",
         "date_created",
         "date_modified",