Improving code quality

.github/workflows/main.yml

@@ -32,13 +32,20 @@ jobs:
           sudo cpan -f -i Email::Outlook::Message
 
       - name: Run tests
+        env:
+          PYTHONPATH: src
         run: |
-          pytest
+          pytest --cov=mailparser --cov-report=xml
           python -m mailparser -v
           python -m mailparser -h
           mail-parser -f tests/mails/mail_malformed_3 -j
           cat tests/mails/mail_malformed_3 | mail-parser -k -j
 
+      - name: Run pre-commit
+        if: matrix.python-version == '3.10'
+        run: |
+          make pre-commit
+
       - name: Report to Coveralls
         if: matrix.python-version == '3.10'
         uses: coverallsapp/github-action@v2.2.3
@@ -57,4 +64,4 @@ jobs:
           name: build-artifacts
           path: |
             dist/mail-parser-*.tar.gz
-            dist/mail_parser-*.whl
+            dist/mail_parser-*.whl

.pre-commit-config.yaml

@@ -0,0 +1,29 @@
+# See https://pre-commit.com for more information
+# See https://pre-commit.com/hooks.html for more hooks
+repos:
+- repo: https://github.com/pre-commit/pre-commit-hooks
+  rev: v5.0.0
+  hooks:
+    - id: trailing-whitespace
+    - id: end-of-file-fixer
+    - id: check-yaml
+    - id: check-added-large-files
+      args: ['--maxkb=5000']
+    - id: check-case-conflict
+    - id: check-json
+    - id: check-merge-conflict
+    - id: detect-aws-credentials
+      args: ["--allow-missing-credentials"]
+    - id: detect-private-key
+    - id: mixed-line-ending
+    - id: check-ast
+
+- repo: https://github.com/astral-sh/ruff-pre-commit
+  # Ruff version.
+  rev: v0.7.0
+  hooks:
+    # Run the linter.
+    - id: ruff
+      args: [ --fix ]
+    # Run the formatter.
+    - id: ruff-format

Makefile

@@ -55,4 +55,4 @@ dist: clean-all ## builds source and wheel package
 	python -m build
 
 release: dist ## package and upload a release
-	twine upload dist/*
+	twine upload dist/*

README.md

@@ -1,4 +1,4 @@
-[![PyPI version](https://badge.fury.io/py/mail-parser.svg)](https://badge.fury.io/py/mail-parser)
+![PyPI - Version](https://img.shields.io/pypi/v/mail-parser)
 [![Coverage Status](https://coveralls.io/repos/github/SpamScope/mail-parser/badge.svg?branch=develop)](https://coveralls.io/github/SpamScope/mail-parser?branch=develop)
 
 ![SpamScope](https://raw.githubusercontent.com/SpamScope/spamscope/develop/docs/logo/spamscope.png)
@@ -288,4 +288,4 @@ Then you can try to run the command line tool:
 $ mail-parser -f tests/mails/mail_malformed_3 -j
 ```
 
-If all is ok, you can start to develop.
+If all is ok, you can start to develop.

docker/Dockerfile

@@ -5,7 +5,7 @@ RUN apt-get -yqq update; \
     apt-get -yqq --no-install-recommends install libemail-outlook-message-perl; \
     apt-get clean; \
     rm -rf /var/lib/apt/lists/*; \
-    git clone -b $BRANCH --single-branch https://github.com/SpamScope/mail-parser.git $MAIL_PARSER_PATH; \ 
+    git clone -b $BRANCH --single-branch https://github.com/SpamScope/mail-parser.git $MAIL_PARSER_PATH; \
     cd $MAIL_PARSER_PATH && python setup.py install
 ENTRYPOINT ["mailparser"]
 CMD ["-h"]

docker/README.md

@@ -17,7 +17,7 @@ This command runs mail-parser help as default, but you can use all others option
 
 To share the "mails" directory between your host and the container, create a "mails" directory on your host.
 
-There also is an example of `docker-compose` 
+There also is an example of `docker-compose`
 
 From the `docker-compose.yml` directory, run:
 

pyproject.toml

@@ -1,3 +1,3 @@
 [build-system]
 requires = ["setuptools >= 40.6.0", "wheel"]
-build-backend = "setuptools.build_meta"
+build-backend = "setuptools.build_meta"

setup.cfg

@@ -67,4 +67,4 @@ addopts =
 	--cov-report=html
 	--junitxml=junit.xml
 	--verbose
-testpaths = tests
+testpaths = tests

setup.py

@@ -17,4 +17,4 @@
 import setuptools
 
 if __name__ == "__main__":
-    setuptools.setup()
+    setuptools.setup()

src/mailparser/__init__.py

@@ -17,13 +17,23 @@
 limitations under the License.
 """
 
-
-from .mailparser import (
+from mailparser.mailparser import (
     MailParser,
     parse_from_bytes,
     parse_from_file,
     parse_from_file_msg,
     parse_from_file_obj,
-    parse_from_string)
+    parse_from_string,
+)
+
+from mailparser.utils import get_header
 
-from .utils import get_header
+__all__ = [
+    "MailParser",
+    "parse_from_bytes",
+    "parse_from_file",
+    "parse_from_file_msg",
+    "parse_from_file_obj",
+    "parse_from_string",
+    "get_header",
+]

src/mailparser/__main__.py

@@ -35,159 +35,155 @@
 
 current = os.path.realpath(os.path.dirname(__file__))
 
-__version__ = runpy.run_path(
-    os.path.join(current, "version.py"))["__version__"]
+__version__ = runpy.run_path(os.path.join(current, "version.py"))["__version__"]
 
 
 def get_args():
     parser = argparse.ArgumentParser(
         description="Wrapper for email Python Standard Library",
         epilog="It takes as input a raw mail and generates a parsed object.",
-        formatter_class=argparse.ArgumentDefaultsHelpFormatter)
+        formatter_class=argparse.ArgumentDefaultsHelpFormatter,
+    )
 
     parsing_group = parser.add_mutually_exclusive_group(required=True)
-    parsing_group.add_argument(
-        "-f",
-        "--file",
-        dest="file",
-        help="Raw email file")
-    parsing_group.add_argument(
-        "-s",
-        "--string",
-        dest="string",
-        help="Raw email string")
+    parsing_group.add_argument("-f", "--file", dest="file", help="Raw email file")
+    parsing_group.add_argument("-s", "--string", dest="string", help="Raw email string")
     parsing_group.add_argument(
         "-k",
         "--stdin",
         dest="stdin",
         action="store_true",
-        help="Enable parsing from stdin")
+        help="Enable parsing from stdin",
+    )
 
     parser.add_argument(
         "-l",
         "--log-level",
         dest="log_level",
         default="WARNING",
         choices=["CRITICAL", "ERROR", "WARNING", "INFO", "DEBUG", "NOTSET"],
-        help="Set log level")
+        help="Set log level",
+    )
 
     parser.add_argument(
         "-j",
         "--json",
         dest="json",
         action="store_true",
-        help="Show the JSON of parsed mail")
+        help="Show the JSON of parsed mail",
+    )
 
     parser.add_argument(
-        "-b",
-        "--body",
-        dest="body",
-        action="store_true",
-        help="Print the body of mail")
+        "-b", "--body", dest="body", action="store_true", help="Print the body of mail"
+    )
 
     parser.add_argument(
         "-a",
         "--attachments",
         dest="attachments",
         action="store_true",
-        help="Print the attachments of mail")
+        help="Print the attachments of mail",
+    )
 
     parser.add_argument(
         "-r",
         "--headers",
         dest="headers",
         action="store_true",
-        help="Print the headers of mail")
+        help="Print the headers of mail",
+    )
 
     parser.add_argument(
-        "-t",
-        "--to",
-        dest="to",
-        action="store_true",
-        help="Print the to of mail")
+        "-t", "--to", dest="to", action="store_true", help="Print the to of mail"
+    )
 
     parser.add_argument(
         "-dt",
         "--delivered-to",
         dest="delivered_to",
         action="store_true",
-        help="Print the delivered-to of mail")
+        help="Print the delivered-to of mail",
+    )
 
     parser.add_argument(
-        "-m",
-        "--from",
-        dest="from_",
-        action="store_true",
-        help="Print the from of mail")
+        "-m", "--from", dest="from_", action="store_true", help="Print the from of mail"
+    )
 
     parser.add_argument(
         "-u",
         "--subject",
         dest="subject",
         action="store_true",
-        help="Print the subject of mail")
+        help="Print the subject of mail",
+    )
 
     parser.add_argument(
         "-c",
         "--receiveds",
         dest="receiveds",
         action="store_true",
-        help="Print all receiveds of mail")
+        help="Print all receiveds of mail",
+    )
 
     parser.add_argument(
         "-d",
         "--defects",
         dest="defects",
         action="store_true",
-        help="Print the defects of mail")
+        help="Print the defects of mail",
+    )
 
     parser.add_argument(
         "-o",
         "--outlook",
         dest="outlook",
         action="store_true",
-        help="Analyze Outlook msg")
+        help="Analyze Outlook msg",
+    )
 
     parser.add_argument(
         "-i",
         "--senderip",
         dest="senderip",
         metavar="Trust mail server string",
-        help="Extract a reliable sender IP address heuristically")
+        help="Extract a reliable sender IP address heuristically",
+    )
 
     parser.add_argument(
         "-p",
         "--mail-hash",
         dest="mail_hash",
         action="store_true",
-        help="Print mail fingerprints without headers")
+        help="Print mail fingerprints without headers",
+    )
 
     parser.add_argument(
         "-z",
         "--attachments-hash",
         dest="attachments_hash",
         action="store_true",
-        help="Print attachments with fingerprints")
+        help="Print attachments with fingerprints",
+    )
 
     parser.add_argument(
         "-sa",
         "--store-attachments",
         dest="store_attachments",
         action="store_true",
-        help="Store attachments on disk")
+        help="Store attachments on disk",
+    )
 
     parser.add_argument(
         "-ap",
         "--attachments-path",
         dest="attachments_path",
         default="/tmp",
-        help="Path where store attachments")
+        help="Path where store attachments",
+    )
 
     parser.add_argument(
-        '-v',
-        '--version',
-        action='version',
-        version='%(prog)s {}'.format(__version__))
+        "-v", "--version", action="version", version="%(prog)s {}".format(__version__)
+    )
 
     return parser
 
@@ -206,8 +202,7 @@ def main():
         parser = mailparser.parse_from_string(args.string)
     elif args.stdin:
         if args.outlook:
-            raise MailParserOutlookError(
-                "You can't use stdin with msg Outlook")
+            raise MailParserOutlookError("You can't use stdin with msg Outlook")
         parser = mailparser.parse_from_file_obj(sys.stdin)
 
     if args.json:
@@ -260,5 +255,5 @@ def main():
         write_attachments(parser.attachments, args.attachments_path)
 
 
-if __name__ == '__main__':
+if __name__ == "__main__":
     main()