Merge pull request #52 from emmo-repo/dic2owl

Update `dic2owl` to new annotated ontology schema.

This PR includes several updates to CI and linting as well.

The `dic2owl` package is updated to v0.2.0.

.github/static/get_dic2owl_deps.py

@@ -0,0 +1,53 @@
+"""
+# Retrieve dependencies
+
+Retrieve dependencies from a `requirements.txt`-style file.
+"""
+import argparse
+from pathlib import Path
+import re
+from typing import Set
+
+
+def main(argv_input: list = None) -> Set[str]:
+    """Retrieve dependencies"""
+    parser = argparse.ArgumentParser(
+        description=main.__doc__,
+        formatter_class=argparse.ArgumentDefaultsHelpFormatter,
+    )
+    parser.add_argument(
+        "requirements_files",
+        metavar="FILE",
+        type=Path,
+        nargs="+",
+        help="The path(s) to one or several requirements.txt-style file(s).",
+    )
+    args = parser.parse_args(argv_input)
+
+    requirements_regex = re.compile(
+        r"^(?P<name>[A-Za-z0-9_-]+)(~=|>=|==).*\n$"
+    )
+    dependencies = set()
+    for file in args.requirements_files:
+        if not file.exists():
+            raise FileNotFoundError(f"Could not find {file} !")
+        with open(file.resolve(), "r") as handle:
+            for line in handle.readlines():
+                match = requirements_regex.fullmatch(line)
+                if match is None:
+                    raise ValueError(
+                        (
+                            "Couldn't determine package name from line:\n\n  "
+                            f"{line}"
+                        )
+                    )
+                dependencies.add(match.group("name"))
+    return dependencies
+
+
+if __name__ == "__main__":
+    from sys import argv
+
+    parsed_dependencies = main(argv[1:])
+    grep_extended_regex = f"({'|'.join(parsed_dependencies)})"
+    print(grep_extended_regex)

.github/workflows/ci.yml

@@ -1,6 +1,6 @@
 name: CI
 
-on: [push, pull_request]
+on: [push]
 
 jobs:
 
@@ -22,4 +22,4 @@ jobs:
           pip install pre-commit
 
       - name: Run pre-commit
-        run: pre-commit run --all-files || ( git status --short ; git diff ; exit 1 )
+        run: pre-commit run --all-files || ( git status --short ; git diff ; exit 1 )

.github/workflows/dic2owl_ci.yml

@@ -1,6 +1,9 @@
 name: CI - dic2owl
 
-on: [push, pull_request]
+on:
+  push:
+    paths:
+      - 'dic2owl/**'
 
 jobs:
 
@@ -21,19 +24,19 @@ jobs:
       - name: Install dependencies
         run: |
           python -m pip install --upgrade pip setuptools wheel
-          pip install -U -r dic2owl/requirements.txt
+          pip install -U -e dic2owl
           pip install bandit pylint safety mypy pytest pytest-cov
 
       - name: Run bandit
         run: bandit -r dic2owl/dic2owl
 
       - name: Run PyLint
-        run: pylint dic2owl/dic2owl
+        run: pylint --rcfile=dic2owl/pyproject.toml dic2owl/dic2owl
 
       - name: Run safety
         run: |
-          safety check -r dic2owl/requirements.txt --bare
-          safety check -r dic2owl/requirements_dev.txt --bare
+          DEPENDENCIES=$( python .github/static/get_dic2owl_deps.py dic2owl/requirements.txt dic2owl/requirements_dev.txt )
+          pip freeze | grep -E "${DEPENDENCIES}" | safety check --stdin
 
       - name: Lint with MyPy
         run: mypy --ignore-missing-imports --scripts-are-modules dic2owl/dic2owl

.gitignore

@@ -4,6 +4,8 @@ old
 __pycache__
 *.pyc
 
+.mypy*
+
 *.dic
 *.cif
 

.pre-commit-config.yaml

@@ -10,17 +10,13 @@ repos:
   rev: 21.7b0
   hooks:
   - id: black
-    name: Blacken
+    args:
+      - --config=dic2owl/pyproject.toml
 
-- repo: https://gitlab.com/pycqa/flake8
-  rev: '3.9.2'
+- repo: https://github.com/pycqa/pylint
+  rev: 'v2.10.1'
   hooks:
-  - id: flake8
+  - id: pylint
     args:
-    - --count
-    - --show-source
-    - --statistics
-    # E501: Line to long. Handled by black.
-    # W503: Line break before binary operator. This is preferred formatting for black.
-    # E203: Whitespace before ':'.
-    - --ignore=E501,W503,E203
+      - --rcfile=dic2owl/pyproject.toml
+      - --disable=import-error

dic2owl/dic2owl/.gitignore

@@ -0,0 +1,2 @@
+cif_core.ttl
+catalog-v001.xml

dic2owl/dic2owl/__init__.py

@@ -1,3 +1,16 @@
-__version__ = "0.1.0"
-__author__ = "Jesper Friis"
-__author_email__ = "[email protected]"
+"""
+# `dic2owl` Python package
+
+This is a tool to generate ontologies from a CIF `.dic`-file.
+
+It can be used either by importing and using the `dic2owl.dic2owl.main`
+function or running it via the `dic2owl` CLI, which will be automatically
+installed when `pip install`-ing this package.
+For more information on how to run the CLI, run `dic2owl --help` in your
+terminal.
+"""
+# pylint: disable=line-too-long
+
+__version__ = "0.2.0"
+__author__ = "Jesper Friis <[email protected]>, Casper Welzel Andersen <[email protected]>, Francesca Lønstad Bleken <[email protected]>"
+__author_email__ = "[email protected]"

dic2owl/dic2owl/cli.py

@@ -1,18 +1,27 @@
+"""
+# `dic2owl` CLI
+
+The `dic2owl` command line interface (CLI) is an easy way of running the
+ontology-generation tool for CIF `.dic`-files.
+"""
 import argparse
 import logging
 from pathlib import Path
 
 
-LOGGING_LEVELS = [logging.getLevelName(level).lower() for level in range(0, 51, 10)]
+LOGGING_LEVELS = [
+    logging.getLevelName(level).lower() for level in range(0, 51, 10)
+]
 
 
-def main(args: list = None):
+def main(argv: list = None) -> None:
     """Ontologize CIF dictionaries (`.dic`) using OWL.
 
     Produce an OWL Turtle (`.ttl`) file from a CIF dictionary (`.dic`) file.
     """
+    # pylint: disable=import-outside-toplevel
     from dic2owl import __version__
-    from dic2owl.generate_cif import main as dic2owl_run
+    from dic2owl.dic2owl import main as dic2owl_run
 
     parser = argparse.ArgumentParser(
         description=main.__doc__,
@@ -43,24 +52,31 @@ def main(args: list = None):
         dest="ttlfile",
         type=Path,
         help=(
-            'The generated output file. Example "--output cif_core.ttl". If output is not '
-            "provided, the filename will be taken to be similar to the CIF_DIC file."
+            'The generated output file. Example "--output cif_core.ttl". If '
+            "output is not provided, the filename will be taken to be similar "
+            "to the CIF_DIC file."
         ),
     )
     parser.add_argument(
         "dicfile",
         metavar="CIF_DIC",
         type=Path,
-        help="The CIF dictionary file from which to generate an OWL ontologized Turtle file.",
+        help=(
+            "The CIF dictionary file from which to generate an OWL ontologized"
+            " Turtle file."
+        ),
     )
 
-    args = parser.parse_args(args)
+    args = parser.parse_args(argv)
 
     if args.ttlfile is None:
-        args.ttlfile = args.dicfile.resolve().name[: -len(args.dicfile.suffix)] + ".ttl"
+        args.ttlfile = (
+            args.dicfile.resolve().name[: -len(args.dicfile.suffix)] + ".ttl"
+        )
 
     if not args.dicfile.resolve().exists():
-        # The dic-file does not exist, use it as a string instead so it can be downloaded.
+        # The dic-file does not exist, use it as a string instead so it can be
+        # downloaded.
         args.dicfile = str(args.dicfile)
 
     dic2owl_run(dicfile=args.dicfile, ttlfile=args.ttlfile)