Merge branch 'issue17' into 'main'

Add tests for wrapper & merge_vocab

Closes #17

See merge request nfdi4cat/ta1-ontologies/voc4cat-tool!43

.gitlab-ci.yml

@@ -26,8 +26,8 @@ before_script:
 tests:
   stage: test
   script:
-    - pip install -U pytest
-    - pip install .
+    # install project inlcuding optional test dependencies
+    - pip install -U .[tests]
     - pytest
 
 # https://docs.gitlab.com/ee/ci/testing/test_coverage_visualization.html#python-example

pyproject.toml

@@ -34,11 +34,11 @@ dependencies = [
   # We install vocexcel directly from a specific branch form github (or gitea in LIKAT)
   "vocexcel @ git+https://github.com/nfdi4cat/VocExcel.git@nfdi4cat-dev",
   #vocexcel @ git+https://gogs.catalysis.de/LIKAT_RDM/VocExcel.git@nfdi4cat-dev
-  "django>=4.0.0",
-  "networkx>=2.8",
-  "openpyxl>=3.0.9",
-  "pillow>=9.1.0",
-  "ontospy[Full] >= 1.9.9.4",
+  "django >= 4.1.3",
+  "networkx >= 2.8",
+  "openpyxl >= 3.0.9",
+  "pillow >= 9.1.0",
+  "ontospy >= 2.1.0",
 ]
 
 #dynamic = ["version", "readme"]
@@ -58,7 +58,12 @@ tests = [
 lint = [
   "black",
   "isort",
-  "wemake-python-styleguide",
+]
+dev = [
+    # Recursively including the project's own optional dependencies requires pip>=21.2
+    "voc4cat[tests,lint]",
+    "ruff",
+    "wemake-python-styleguide",
 ]
 
 [project.scripts]
@@ -155,6 +160,7 @@ exclude_lines = [
     "raise AssertionError",
     "raise NotImplementedError",
     "return NotImplemented",
+    "if __name__ == .__main__.:",
 ]
 
 [tool.coverage.html]

src/voc4cat/__init__.py

@@ -0,0 +1,11 @@
+# -*- coding: utf-8 -*-
+from importlib.metadata import PackageNotFoundError, version
+
+try:
+    __version__ = version("voc4cat")
+except PackageNotFoundError:
+    # package is not installed
+    try:
+        from ._version import version as __version__
+    except ImportError:
+        __version__ = "0.0.0"

src/voc4cat/merge_vocab.py

@@ -1,52 +1,55 @@
+# -*- coding: utf-8 -*-
 # This script is mainly useful for CI.
 import os
 import shutil
 import subprocess
 import sys
 from pathlib import Path
 
-VOCAB_DIR = Path("vocabularies")
-OUTBOX = Path("outbox")
 
-
-def main(outbox, vocab):
+def main(ttl_inbox, vocab):
     """
-    Sync ttl-files in outbox and vocab folder
+    Sync ttl-files from ttl_inbox into vocab folder
 
     New files are copied, existing files are synced by git merge-file.
     """
     retcode = 0
-    for p in os.listdir(outbox):
-        new = outbox / Path(p)
+    for p in os.listdir(ttl_inbox):
+        new = ttl_inbox / Path(p)
         if not new.suffix == ".ttl" or new.is_dir():
-            print(f'skipping "{new}"')
+            print(f'Skipping "{new}"')
             continue
         if os.path.exists(vocab / Path(new).name):
             exists = vocab / Path(new).name
             cmd = ["git", "merge-file", "--theirs", str(exists), str(exists), str(new)]
-            print(" ".join(cmd))
+            print("Running cmd: {0}".format(" ".join(cmd)))
             outp = subprocess.run(cmd, capture_output=True)
             print(outp.stdout)
             if retcode := outp.returncode != 0:
                 break
         else:
-            print(f'copying "{new}" to "{vocab}"')
+            print(f'Copying "{new}" to "{vocab}"')
             shutil.copy(new, vocab)
     return retcode
 
 
-def run():
-    if len(sys.argv) != 3:
-        print('Usage: "python merge_vocab.py outbox_dir vocab_dir')
-        sys.exit(1)
-    outbox, vocab = sys.argv[1:3]
+def main_cli(args=None):
+
+    if args is None:  # script run via entrypoint
+        args = sys.argv[1:]
+
+    if len(args) != 2:
+        print("Usage: python merge_vocab.py <outbox_dir> <vocab_dir>")
+        return 1
+    outbox, vocab = args
     if os.path.exists(outbox) and os.path.exists(vocab):
         retcode = main(Path(outbox), Path(vocab))
-        sys.exit(retcode)
-    else:
-        print(f'This script requires both folders to exist: "{outbox}" and "{vocab}"')
-        sys.exit(1)
+        return retcode
+
+    print(f'This script requires both folders to exist: "{outbox}" and "{vocab}"')
+    return 1
 
 
 if __name__ == "__main__":
-    run()
+    err = main_cli(sys.argv[1:])
+    sys.exit(err)

src/voc4cat/util.py

@@ -115,7 +115,7 @@ def dag_to_node_levels(termgraph, baselevel=0):
         for sgc in nx.connected_components(termgraph.to_undirected())
     ]
     tg_copy = termgraph.copy()
-    # We must find & eliminate cycles no matter which direction for the tree representation.
+    # We must find & eliminate cycles no matter if directed or undirected.
     cycles = nx.cycle_basis(tg_copy.to_undirected())
     broken_edges = []
     while cycles:
@@ -125,8 +125,8 @@ def dag_to_node_levels(termgraph, baselevel=0):
         tg_copy.remove_edge(*edge_to_break)
         # nx.cycle_basis does only report cycles >= 3 members.
         cycles = nx.cycle_basis(tg_copy.to_undirected())
-        # It is not clearly documented in networkx if the direction of edges is preserved
-        # when converting to an undirected graph. So we better check...
+        # It is not clearly documented in networkx if the direction of edges is
+        # preserved when converting to an undirected graph. So we better check...
         if termgraph.has_edge(*edge_to_break):
             broken_edges.append(edge_to_break)
         else:  # pragma: no cover