ghga-transpiler 1.4.0__tar.gz → 2.0.0__tar.gz

{ghga-transpiler-1.4.0 → ghga_transpiler-2.0.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
-Name: ghga-transpiler
-Version: 1.4.0
+Name: ghga_transpiler
+Version: 2.0.0
 Summary: GHGA-Transpiler - excel to JSON converter
 Author-email: "German Human Genome Phenome Archive (GHGA)" <contact@ghga.de>
 License: Apache 2.0
@@ -19,9 +19,10 @@ Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: typer~=0.9.0
 Requires-Dist: openpyxl==3.*,>=3.1.2
-Requires-Dist: defusedxml
+Requires-Dist: defusedxml==0.*,>=0.7
 Requires-Dist: pydantic<3,>=2
 Requires-Dist: PyYAML~=6.0
+Requires-Dist: semver==3.*
 
 
 [![tests](https://github.com/ghga-de/ghga-transpiler/actions/workflows/tests.yaml/badge.svg)](https://github.com/ghga-de/ghga-transpiler/actions/workflows/unit_and_int_tests.yaml)

{ghga-transpiler-1.4.0 → ghga_transpiler-2.0.0}/pyproject.toml

@@ -4,8 +4,8 @@ build-backend = "setuptools.build_meta"
 
 [project]
 # please adapt to package name
-name = "ghga-transpiler"
-version = "1.4.0"
+name = "ghga_transpiler"
+version = "2.0.0"
 description = "GHGA-Transpiler - excel to JSON converter"
 readme = "README.md"
 authors = [
@@ -27,9 +27,10 @@ classifiers = [
 dependencies = [
     "typer ~= 0.9.0",
     "openpyxl >= 3.1.2, == 3.*",
-    "defusedxml", # whichever version is compatible with openpyxl
+    "defusedxml >= 0.7, == 0.*",
     "pydantic >=2, <3",
-    "PyYAML ~= 6.0"
+    "PyYAML ~= 6.0",
+    "semver == 3.*"
 ]
 
 [project.urls]

{ghga-transpiler-1.4.0 → ghga_transpiler-2.0.0}/src/ghga_transpiler/__init__.py

@@ -17,4 +17,11 @@
 
 from importlib.metadata import version
 
+from openpyxl.xml import DEFUSEDXML
+
 __version__ = version(__package__)
+
+if not DEFUSEDXML:
+    raise RuntimeError(
+        "The 'defusedxml' package must be present to safely run ghga-transpiler."
+    )

{ghga-transpiler-1.4.0 → ghga_transpiler-2.0.0}/src/ghga_transpiler/cli.py

@@ -22,7 +22,7 @@ import typer
 
 from . import __version__, io
 from .config.exceptions import UnknownVersionError
-from .core import convert_workbook
+from .core import InvalidSematicVersion, convert_workbook
 
 cli = typer.Typer()
 
@@ -66,7 +66,7 @@ def transpile(
     """
     try:
         ghga_workbook = io.read_workbook(spread_sheet)
-    except (SyntaxError, UnknownVersionError) as exc:
+    except (SyntaxError, UnknownVersionError, InvalidSematicVersion) as exc:
         sys.exit(f"Unable to parse input file '{spread_sheet}': {exc}")
 
     converted = convert_workbook(ghga_workbook)

{ghga-transpiler-1.4.0 → ghga_transpiler-2.0.0}/src/ghga_transpiler/core.py

@@ -15,16 +15,17 @@
 #
 
 """This module contains functionalities for processing excel sheets into json object."""
-import re
 from importlib import resources
 from typing import Callable, Optional, Union
 
+import semver
 from openpyxl import Workbook
 
 from . import config
 
-# pylint: disable=line-too-long
-SEMVER_REGEX = r"^(0|[1-9]\d*)\.(0|[1-9]\d*)\.(0|[1-9]\d*)(?:-((?:0|[1-9]\d*|\d*[a-zA-Z-][0-9a-zA-Z-]*)(?:\.(?:0|[1-9]\d*|\d*[a-zA-Z-][0-9a-zA-Z-]*))*))?(?:\+([0-9a-zA-Z-]+(?:\.[0-9a-zA-Z-]+)*))?$"
+
+class InvalidSematicVersion(Exception):
+    """Raised when a version string is invalid."""
 
 
 class GHGAWorkbook:
@@ -33,20 +34,28 @@ class GHGAWorkbook:
     def __init__(self, workbook: Workbook, configs_package: resources.Package):
         """Create a new GHGAWorkbook object from an XLSX workbook"""
         self.workbook = workbook
-        self.version = GHGAWorkbook._get_version(workbook)
-        self.config = config.load_config(self.version, configs_package)
+        self.wb_version = GHGAWorkbook._get_version(workbook)
+        self.config = config.load_config(self.major_minor_version, configs_package)
 
     @staticmethod
     def _get_version(workbook):
         """Function to get workbook version from the worksheet _properties"""
         if "__properties" in workbook.sheetnames:
-            version = str(workbook["__properties"].cell(1, 1).value)
-            if re.fullmatch(SEMVER_REGEX, version):
-                return version
+            try:
+                return semver.Version.parse(workbook["__properties"].cell(1, 1).value)
+            except ValueError:
+                raise InvalidSematicVersion(
+                    "Unable to extract metadata model version from the provided workbook (not a valid semantic version)."
+                ) from None
         raise SyntaxError(
-            "Unable to extract metadata version from the provided workbook."
+            "Unable to extract metadata model version from the provided workbook (missing)."
         )
 
+    @property
+    def major_minor_version(self):
+        """Returns only major and minor version numbers"""
+        return f"{self.wb_version.major}.{self.wb_version.minor}"
+
 
 def get_worksheet_rows(
     worksheet,

{ghga-transpiler-1.4.0 → ghga_transpiler-2.0.0}/src/ghga_transpiler/transformations.py

@@ -38,7 +38,28 @@ def to_attributes() -> Callable:
         return dict(zip(("key", "value"), splitted))
 
     def split_mult(value: str) -> list[dict]:
-        """Function to convert string to attributes"""
+        """Converts string to attributes"""
         return [split_one(elem) for elem in split_by_semicolon(value)]
 
     return split_mult
+
+
+def snake_case(cv: str) -> str:
+    """Converts format of a string to SNAKE_CASE"""
+    return cv.replace(" ", "_").upper()
+
+
+def to_snake_case() -> Callable:
+    """Returns a function that converts a string to SNAKE_CASE"""
+    return snake_case
+
+
+def snake_case_list(value: str) -> list[str]:
+    """Combines the functions to split_by_semicolon and convert_to_snake_case"""
+    list_to_convert = split_by_semicolon(value)
+    return [snake_case(elem) for elem in list_to_convert]
+
+
+def to_snake_case_list() -> Callable:
+    """Returns a function that converts a semicolon separated string into a list of snake-cased strings"""
+    return snake_case_list

{ghga-transpiler-1.4.0 → ghga_transpiler-2.0.0}/src/ghga_transpiler.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: ghga-transpiler
-Version: 1.4.0
+Version: 2.0.0
 Summary: GHGA-Transpiler - excel to JSON converter
 Author-email: "German Human Genome Phenome Archive (GHGA)" <contact@ghga.de>
 License: Apache 2.0
@@ -19,9 +19,10 @@ Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: typer~=0.9.0
 Requires-Dist: openpyxl==3.*,>=3.1.2
-Requires-Dist: defusedxml
+Requires-Dist: defusedxml==0.*,>=0.7
 Requires-Dist: pydantic<3,>=2
 Requires-Dist: PyYAML~=6.0
+Requires-Dist: semver==3.*
 
 
 [![tests](https://github.com/ghga-de/ghga-transpiler/actions/workflows/tests.yaml/badge.svg)](https://github.com/ghga-de/ghga-transpiler/actions/workflows/unit_and_int_tests.yaml)

{ghga-transpiler-1.4.0 → ghga_transpiler-2.0.0}/src/ghga_transpiler.egg-info/requires.txt

@@ -1,5 +1,6 @@
 typer~=0.9.0
 openpyxl==3.*,>=3.1.2
-defusedxml
+defusedxml==0.*,>=0.7
 pydantic<3,>=2
 PyYAML~=6.0
+semver==3.*

{ghga-transpiler-1.4.0 → ghga_transpiler-2.0.0}/tests/test_process_workbook.py

@@ -16,8 +16,9 @@
 """Unit tests for core functions"""
 
 import pytest
+import semver
 
-from ghga_transpiler.core import GHGAWorkbook
+from ghga_transpiler.core import GHGAWorkbook, InvalidSematicVersion
 
 from .fixtures.utils import create_workbook
 
@@ -30,7 +31,7 @@ def test_extract_good_version() -> None:
     value = workbook["__properties"].cell(row=1, column=1, value="10.3.1-rc2").value
     # pylint: disable=protected-access
     version = GHGAWorkbook._get_version(workbook)
-    assert version == value
+    assert version == semver.Version.parse(str(value))
 
 
 def test_extract_bad_version() -> None:
@@ -39,6 +40,7 @@ def test_extract_bad_version() -> None:
     """
     workbook = create_workbook("__properties")
     workbook["__properties"].cell(row=1, column=1, value="20.10.3.1")
-    with pytest.raises(SyntaxError):
+
+    with pytest.raises(InvalidSematicVersion):
         # pylint: disable=protected-access
         GHGAWorkbook._get_version(workbook)