dev

.github/workflows/tox_matrix.yml

@@ -0,0 +1,47 @@
+on: [push, pull_request]
+name: Tox Multi-platform Compatibility Test
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+
+
+jobs:
+  setup:
+    name: Setup
+    runs-on: ubuntu-latest
+    outputs:
+      tox_matrix: ${{ steps.dataStep.outputs.myoutput }}
+    steps:
+      - uses: actions/checkout@v6
+      - uses: actions/setup-python@v6
+        name: Get tox environments
+        with:
+          python-version: '3.13'
+          cache: 'pip' # caching pip dependencies
+      - id: dataStep
+        run: |
+          pip install uv
+          echo "myoutput=$(uv run --quiet --only-group=tox tox --listenvs -a | jq  --raw-input . | jq -c --slurp )" >> $GITHUB_OUTPUT
+
+  build:
+    needs: setup
+    runs-on: ${{ matrix.os }}
+    strategy:
+      matrix:
+        os: [ubuntu-latest, macos-latest, windows-latest]
+        tox-environment: ${{ fromJson(needs.setup.outputs.tox_matrix) }}
+      fail-fast: false
+    name: "${{ matrix.os }}: ${{ matrix.tox-environment }} "
+    steps:
+      - uses: actions/checkout@v6
+      - uses: actions/setup-python@v6
+        with:
+          cache: 'pip' # caching pip dependencies
+      - name: Install requirements
+        run: |
+          pip install uv
+      - name: Run tox
+        env:
+          TOXENV: ${{ matrix.tox-environment }}
+        run: uv run --only-group tox-uv tox run -e ${{matrix.tox-environment}}

pyproject.toml

@@ -6,7 +6,7 @@ build-backend = "setuptools.build_meta"
 name = "HathiValidate"
 version = "0.3.9.dev0"
 dependencies = [
-    "lxml<6.0; sys_platform != 'darwin' or python_version != '3.8' or platform_machine != 'arm64'",
+    "lxml",
     "PyYAML",
     ]
 requires-python = ">=3.10"
@@ -29,24 +29,35 @@ Documentation = "https://www.library.illinois.edu/dccdocs/hathi_validate"
 Download = "https://github.com/UIUCLibrary/HathiValidate"
 
 [dependency-groups]
+audit-dependencies = ["uv-secure"]
 publish = ["twine"]
 tox = ["tox"]
+tox-uv = [
+    {include-group = "tox"},
+    "tox-uv-bare"
+]
 docs = ["sphinx"]
 test =["pytest", "behave"]
+type-checking = ["mypy", "types-PyYAML"]
+code-style = ["pydocstyle", "flake8"]
+lint = [
+    {include-group = "code-style"},
+    "pylint"
+]
+security = ["bandit"]
 dev = [
     "coverage",
-    "flake8",
     "lxml",
-    "mypy",
     "paramiko",
     "pep517",
-    "pydocstyle",
-    "pylint",
     "PyYAML",
     "setuptools>=30.3.0",
-    {include-group = "test"},
+    {include-group = "code-style"},
     {include-group = "docs"},
+    {include-group = "lint"},
+    {include-group = "test"},
     {include-group = "tox"},
+    {include-group = "type-checking"},
     {include-group = "publish"},
     "types-PyYAML"
 ]
@@ -61,7 +72,7 @@ zip-safe = false
 package-dir = {"" = "src"}
 
 [tool.setuptools.package-data]
-hathi_validate = ["xsd/*.xsd", "py.typed"]
+hathi_validate = ["xsd/*.xsd", "py.typed", "catalog.xml"]
 
 [project.scripts]
 hathivalidate = "hathi_validate.cli:main"

src/hathi_validate/catalog.xml

@@ -0,0 +1,7 @@
+<?xml version="1.0"?>
+<!DOCTYPE catalog PUBLIC "-//OASIS//DTD XML Catalogs V1.0//EN"
+        "http://www.oasis-open.org/committees/entity/release/1.0/catalog.dtd">
+<catalog xmlns="urn:oasis:names:tc:entity:xmlns:xml:catalog">
+    <uri name="http://www.loc.gov/standards/xlink/xlink.xsd"
+         uri="xsd/xlink.xsd"/>
+</catalog>

src/hathi_validate/process.py

@@ -12,8 +12,9 @@
 import yaml
 from lxml import etree
 
-from importlib.resources import files
+from importlib.resources import files, as_file
 
+import hathi_validate
 from hathi_validate import result
 from hathi_validate import xsd as hathi_xsd
 from . import validator
@@ -102,7 +103,7 @@
 
 def calculate_md5(filename: str, chunk_size: int = 8192) -> str:
     """Calculate the md5 hash value of a file."""
-    md5 = hashlib.md5()
+    md5 = hashlib.md5(usedforsecurity=False)
 
     with open(filename, "rb") as file_handle:
         while True:
@@ -416,7 +417,7 @@
     return finder.find_errors()
 
 
 def find_errors_ocr(path: str) -> result.ResultSummary:
     """Validate all xml files located in the given path.
 
         Make sure they are valid to the alto scheme
@@ -442,11 +443,25 @@
         return True
 
     logger = logging.getLogger(__name__)
-    alto_scheme = etree.XMLSchema(
-        etree.XML(
-            files(hathi_xsd).joinpath("alto.xsd").read_bytes()
-        )
-    )
+    existing_xml_catalog_file: Optional[str] = None
+
+    try:
+        # This is because libxml2 no longer downloads xsd files automatically.
+        # This redirects it to use the xlink document included
+        with as_file(
+            files(hathi_validate).joinpath('catalog.xml')
+        ) as catalog_file:
+            os.environ['XML_CATALOG_FILES'] = str(catalog_file)
+            alto_scheme = etree.XMLSchema(
+                etree.XML(
+                    files(hathi_xsd).joinpath("alto.xsd").read_bytes()
+                )
+            )
+    finally:
+        if existing_xml_catalog_file is None:
+            os.environ.pop('XML_CATALOG_FILES', None)
+        else:
+            os.environ['XML_CATALOG_FILES'] = existing_xml_catalog_file
 
     summary_builder = result.SummaryDirector(source=path)
     for xml_file in filter(ocr_filter, os.scandir(path)):
@@ -455,9 +470,12 @@
                 doc = etree.fromstring(file_handle.read().encode("utf-8"))
 
             if not alto_scheme.validate(doc):
-                summary_builder.add_error(
-                    "{} does not validate to ALTO scheme".format(xml_file.name)
-                )
+                for error in alto_scheme.error_log:
+                    summary_builder.add_error(
+                        f"{xml_file.name} does not validate to ALTO scheme. "
+                        f"Line: {error.line}, Column: {error.column}, "
+                        f"Reason: {error.message}"
+                    )
             else:
                 logger.info(
                     "%s validates to the ALTO XML scheme", xml_file.name

tox.ini

@@ -3,13 +3,14 @@
 ; pytest_args = "--junitxml=reports/junit-{env:OS:UNKNOWN_OS}-{envname}.xml --junit-prefix={env:OS:UNKNOWN_OS}  --basetemp={envtmpdir}"
 ; sphinx_args = "-W -b html -d {envtmpdir}/doctrees source  {envtmpdir}/html"
 [tox]
-envlist = py310, py311, py312, py313
+envlist = py310, py311, py312, py313, py314, py314t
 isolated_build = true
 min_version = 4.20
 
 [tool:pytest]
 testpath = tests
 addopts = --verbose
+
 [testenv]
 dependency_groups = test
 commands=
@@ -19,44 +20,40 @@ commands=
 
 [testenv:docs]
 changedir=docs
-deps=
-    sphinx
+dependency_groups = docs
 commands=
-    sphinx-build {env:sphinx_args:-W -b html -d {envtmpdir}/doctrees source  {distdir}/html}
-
+    sphinx-build {env:sphinx_args:-W -b html -d {envtmpdir}/doctrees source  {temp_dir}/docs/html}
 
 [testenv:mypy]
-deps =
-    mypy
+dependency_groups = type-checking
 skip_install=True
+ignore_outcome = true
 setenv = MYPY_CACHE_DIR = {temp_dir}/.mypy_cache
-commands = mypy {posargs: -p hathi_validate}
+commands = mypy {posargs: src/hathi_validate}
 
 [testenv:flake8]
 description = check the code style
-deps =
-    flake8
+dependency_groups = code-style
 skip_install=True
-commands = flake8 {posargs: hathi_validate}
+commands = flake8 {posargs: src/hathi_validate}
 
 
 [testenv:pylint]
 description = check the code style
-deps =
-    pylint
-skip_install=True
-commands =  pylint {posargs: hathi_validate} --disable import-error
-
+dependency_groups = lint
+skip_install = true
+ignore_outcome = true
+commands =  pylint {posargs: src/hathi_validate} --disable import-error
 
 [testenv:pydocstyle]
 skip_install = true
-deps=
-    pydocstyle
+ignore_outcome = true
+dependency_groups = code-style
 commands =
-    pydocstyle {posargs: {toxinidir}/hathi_validate}
+    pydocstyle {posargs: {toxinidir}/src/hathi_validate}
+
 [testenv:bandit]
 skip_install = true
-deps=
-    bandit
+dependency_groups = security
 commands =
-    bandit {posargs: --recursive {toxinidir}/hathi_validate}
+    bandit {posargs: --recursive {toxinidir}/src/hathi_validate}