Source code for semantic_release.history.parser_angular
"""
Angular commit style parser
https://github.com/angular/angular/blob/master/CONTRIBUTING.md#-commit-message-guidelines
"""
import logging
import re
from ..errors import ImproperConfigurationError, UnknownCommitMessageStyleError
from ..helpers import LoggedFunction
from ..settings import config
from .parser_helpers import ParsedCommit, parse_paragraphs, re_breaking
logger = logging.getLogger(__name__)
# types with long names in changelog
LONG_TYPE_NAMES = {
"feat": "feature",
"docs": "documentation",
"perf": "performance",
}
LEVEL_BUMPS = {"no-release": 0, "patch": 1, "minor": 2, "major": 3}
[docs]@LoggedFunction(logger)
def parse_commit_message(message: str) -> ParsedCommit:
"""
Parse a commit message according to the angular commit guidelines specification.
:param message: A string of a commit message.
:return: A tuple of (level to bump, type of change, scope of change, a tuple with descriptions)
:raises UnknownCommitMessageStyleError: if regular expression matching fails
"""
# loading these are here to make it easier to mock in tests
allowed_types = config.get("parser_angular_allowed_types").split(",")
minor_types = config.get("parser_angular_minor_types").split(",")
patch_types = config.get("parser_angular_patch_types").split(",")
re_parser = re.compile(
r"(?P<type>" + "|".join(allowed_types) + ")"
r"(?:\((?P<scope>[^\n]+)\))?"
r"(?P<break>!)?: "
r"(?P<subject>[^\n]+)"
r"(:?\n\n(?P<text>.+))?",
re.DOTALL,
)
# Attempt to parse the commit message with a regular expression
parsed = re_parser.match(message)
if not parsed:
raise UnknownCommitMessageStyleError(
f"Unable to parse the given commit message: {message}"
)
parsed_break = parsed.group("break")
parsed_scope = parsed.group("scope")
parsed_subject = parsed.group("subject")
parsed_text = parsed.group("text")
parsed_type = parsed.group("type")
if parsed_text:
descriptions = parse_paragraphs(parsed_text)
else:
descriptions = list()
# Insert the subject before the other paragraphs
descriptions.insert(0, parsed_subject)
# Look for descriptions of breaking changes
breaking_descriptions = [
match.group(1)
for match in (re_breaking.match(p) for p in descriptions[1:])
if match
]
default_level_bump = config.get("parser_angular_default_level_bump").lower()
if default_level_bump not in LEVEL_BUMPS.keys():
raise ImproperConfigurationError(
f"{default_level_bump} is not a valid option for "
f"parser_angular_default_level_bump.\n"
f"valid options are: {', '.join(LEVEL_BUMPS.keys())}"
)
level_bump = LEVEL_BUMPS[default_level_bump]
if parsed_break or breaking_descriptions:
level_bump = 3 # Major
elif parsed_type in minor_types:
level_bump = 2 # Minor
elif parsed_type in patch_types:
level_bump = 1 # Patch
parsed_type_long = LONG_TYPE_NAMES.get(parsed_type, parsed_type)
# first param is the key you're getting from the dict,
# second param is the default value
# allows only putting types with a long name in the LONG_TYPE_NAMES dict
return ParsedCommit(
level_bump,
parsed_type_long,
parsed_scope,
descriptions,
breaking_descriptions,
)