Merge pull request #13 from aarhusstadsarkiv/tests

Versions 0.3.1: Tests & Fixes

.github/workflows/on-push.yml

@@ -1,8 +1,32 @@
 name: Linting & Test
 
 on:
-    push
+  push
+
+env:
+  PYTHON_VERSION: 3.9.18
+  POETRY_VERSION: 1.6.1
 
 jobs:
-    call-linting-from-utils:
-        uses: aarhusstadsarkiv/acautils/.github/workflows/linting_ruff-mypy.yml@main
+  linting:
+    uses: aarhusstadsarkiv/acautils/.github/workflows/linting_ruff-mypy.yml@main
+  pytest:
+    name: pytest
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v4
+        with:
+          python-version: ${{ env.PYTHON_VERSION }}
+      - uses: abatilo/actions-poetry@v2
+        with:
+          poetry-version: ${{ env.POETRY_VERSION }}
+      - uses: actions/setup-go@v4
+      - run: poetry install
+      - run: go install github.com/richardlehane/siegfried/cmd/sf@latest
+      - name: Unit test
+        env:
+          GOPATH: /home/runner/go
+        run: |
+          poetry run coverage run -m pytest
+          poetry run coverage report -m --fail-under=80 --skip-empty --skip-covered

.gitignore

@@ -162,3 +162,5 @@ cython_debug/
 /*.db
 
 .ruff_cache/
+
+/tests/tmp/

acacore/__version__.py

@@ -1 +1 @@
-__version__ = "0.3.0"
+__version__ = "0.3.1"

acacore/database/column.py

@@ -56,7 +56,9 @@ def _schema_to_column(name: str, schema: dict, defs: Optional[dict[str, dict]] =
         sql_type = _sql_schema_types.get(schema_type, None)
         type_name: str = schema.get("format", schema_type)
 
-        if type_name in _sql_schema_type_converters:
+        if schema.get("enum"):
+            to_entry, from_entry = lambda e: e.value, str
+        elif type_name in _sql_schema_type_converters:
             to_entry, from_entry = _sql_schema_type_converters[type_name]
         else:
             raise TypeError(f"Cannot recognize type from schema {schema!r}")

acacore/database/files_db.py

@@ -126,6 +126,7 @@ def init(self):
         self.files.create(True)
         self.metadata.create(True)
         self.converted_files.create(True)
+        self.history.create(True)
         self.not_converted.create(True)
         self.identification_warnings.create(True)
         self.signature_count.create(True)

acacore/models/base.py

@@ -1,4 +1,3 @@
-import json
 from pathlib import Path
 from typing import Any
 
@@ -12,4 +11,4 @@ def dump(self, to_file: Path) -> None:
         to_file.write_text(super().model_dump_json(), encoding="utf-8")
 
     def encode(self) -> Any:  # noqa: ANN401
-        return json.loads(super().model_dump_json())
+        return super().model_dump(mode="json")

acacore/models/file.py

@@ -20,10 +20,10 @@
 
 
 class Action(Enum):
-    CONVERT = "Convertool: To convert."
-    REPLACE = "Convertool: Replace with template. File is not preservable."
-    MANUAL = "Manual: File should be converted manually. [info about the manual conversion from reference_files]."
-    RENAME = "Renamer: File has extension mismatch. Should be renamed"
+    CONVERT = "CONVERT"  # To convert.
+    REPLACE = "REPLACE"  # Replace with template. File is not preservable.
+    MANUAL = "MANUAL"  # File should be converted manually. [info about the manual conversion from reference_files].
+    RENAME = "RENAME"  # File has extension mismatch. Should be renamed
 
 
 # -----------------------------------------------------------------------------
@@ -150,14 +150,11 @@ def size_fmt(self) -> str:
         str
             File size in human-readable format.
         """
-        return str(size_fmt(self.get_absolute_path().stat().st_size))
+        return size_fmt(self.get_absolute_path().stat().st_size)
 
     def get_bof_and_eof(self) -> Tuple[str, str]:
         """Get the first and last kilobyte of the file.
 
-        Args:
-            file (Path): Path to file
-
         Returns:
             Tuple[str,str]: BOF and then EOF as `str`.
         """
@@ -172,7 +169,7 @@ def get_bof_and_eof(self) -> Tuple[str, str]:
                 # File too small :)
                 file_bytes.seek(-file_bytes.tell(), 2)
             eof = file_bytes.read(1024).hex()
-        return (bof, eof)
+        return bof, eof
 
 
 class ArchiveFile(Identification, File):

acacore/utils/functions.py

@@ -1,3 +1,4 @@
+from pathlib import Path
 from typing import Callable
 from typing import Optional
 from typing import TypeVar
@@ -18,3 +19,14 @@ def or_none(func: Callable[[T], R]) -> Callable[[T], Optional[R]]:
         object: A function of type (T) -> R | None.
     """
     return lambda x: None if x is None else func(x)
+
+
+def rm_tree(path: Path):
+    if not path.is_dir():
+        path.unlink(missing_ok=True)
+        return
+
+    for item in path.iterdir():
+        rm_tree(item) if item.is_dir() else item.unlink(missing_ok=True)
+
+    path.rmdir()

acacore/utils/io.py

@@ -9,13 +9,12 @@ def size_fmt(size: float) -> str:
 
     Parameters
     ----------
-    size: float
+    size:
         The file size in bytes.
 
     Returns:
     -------
-    str
-        Human readable string representing size in binary multiples.
+    Human readable string representing size in binary multiples.
     """
     unit: int = int(log2(size) // 10)
     unit = unit if unit < len(binary_units) else len(binary_units) - 1

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "acacore"
-version = "0.3.0"
+version = "0.3.1"
 description = ""
 authors = ["Matteo Campinoti <[email protected]>"]
 license = "GPL-3.0"
@@ -10,11 +10,13 @@ readme = "README.md"
 python = "^3.9"
 pydantic = "^2.4.2"
 tqdm = "^4.66.1"
+coverage = "^7.3.2"
 
 
 [tool.poetry.group.dev.dependencies]
 ruff = "^0.0.286"
 black = "^23.7.0"
+pytest = "^7.4.3"
 
 [build-system]
 requires = ["poetry-core"]

tests/conftest.py

@@ -0,0 +1,32 @@
+from json import loads
+from pathlib import Path
+
+import pytest
+
+from acacore.utils.functions import rm_tree
+
+
+@pytest.fixture(scope="session")
+def test_folder() -> Path:
+    return Path(__file__).parent
+
+
+@pytest.fixture(scope="session")
+def temp_folder(test_folder: Path) -> Path:
+    return test_folder / "tmp"
+
+
+@pytest.fixture(scope="session")
+def test_files(test_folder: Path) -> Path:
+    return test_folder / "files"
+
+
+@pytest.fixture(scope="session")
+def test_files_data(test_files: Path) -> dict[str, dict]:
+    return loads(test_files.joinpath("files.json").read_text())
+
+
+@pytest.fixture(autouse=True, scope="session")
+def _pre_test(temp_folder: Path):
+    rm_tree(temp_folder)
+    temp_folder.mkdir(parents=True, exist_ok=True)