Parse a mini-language for selecting objects by tag or name

[Home | GitLab | PyPI | ReadTheDocs]

This library is mostly useful for command-line parsing by other tools like tox-stages and nox-stages. It may be used to parse e.g. a command-line specification like @check and not pylint or @tests or ruff and then match it against a list of objects that have names and lists of tags.

Parse stage specifications

The parse_spec() function parses a string specification into a BoolExpr object that may later be used to select matching objects (e.g. test environments).

The specification mini-language may roughly be described as:

expr ::= and_expr ["or" and_expr...]
and_expr ::= not_expr ["and" not_expr...]
not_expr ::= ["not"] atom
atom ::= tag | keyword
tag ::= "@" characters
keyword ::= characters
characters ::= [A-Za-z0-9_-]+

Thus, all of the following:

• @check
• @check and @quick
• @tests and not examples
• not @tests
• pep8 or not @quick and @check

...are valid expressions, with the "not", "and", and "or" keywords having their usual precedence (pep8 or not @quick and @check is parsed as pep8 or ((@not quick) and @check)).

Check whether an object matches a parsed specification

The parse-stages library provides two base dataclasses for objects that may be matched against parsed expressions: TaggedFrozen and Tagged. Both classes have the same members:

• name: a string
• tags: a list of strings
• get_keyword_haystacks(): a method that returns a list of strings, [self.name] unless overridden

When a BoolExpr object's evaluate() method is called for a specific TaggedFrozen or Tagged object, it checks whether the specification matches the tags and keywords defined for this object. Tags are matched exactly, while a keyword is considered to match if it is contained in the checked string; e.g. pep would match both pep8 and exp_pep563, while @black would not match a black-reformat tag.

The get_keyword_haystacks() method returns the strings to look in for matching keywords. By default, it only returns the name field; however, it may be extended, e.g. for Nox sessions it may also return the name of the Python function that implements the session, for test classes with methods it may return the class name and the method name, etc.

Examples

Parse a list of stage specifications into expressions that may later be matched against test environment definitions:

e_check = parse_stages.parse_spec("@check")
e_check_quick = parse_stages.parse_spec("@check and @quick")
e_check_no_ruff = parse_stages.parse_spec("@check and not ruff")

specs = [(spec, parse_stages.parse_spec(spec)) for spec in args.stage_specs]

Select the test environments that match the specification:

# Obtain a list (okay, a dictionary) of test environments in some way
tox_envs = get_tox_environments()  # {"ruff": {"tags": ["check", "quick"]}, ...}

# Convert them to objects that the parsed expressions can match
all_envs = [
    parse_stages.TaggedFrozen(name, env["tags"])
    for name, env in tox_envs.items()
]

# Or define our own class that may hold additional information
@dataclasses.dataclass(frozen=True)
class TestEnv(parse_stages.TaggedFrozen):
    """A single test environment: name, tags, etc."""
    ...

all_envs = [TestEnv(name, env["tags"], ...) for name, env in tox_envs.items()]

# Select the ones that match the "@check" expression
matched = [env for env in all_envs if e_check.evaluate(env)]

# Or if we only care about the names...
quick_names = [env.name for env in all_envs if e_check_quick.evaluate(env)]

Contact

The parse-stages library was written by Peter Pentchev. It is developed in a GitLab repository. This documentation is hosted at Ringlet with a copy at ReadTheDocs.