black library internals
This page is all about the black python library, the contents below is auto-generated by inspecting a python package’s code base.
Package Files
- black/__init__.py
- black/__main__.py
- black/brackets.py
- black/cache.py
- black/comments.py
- black/concurrency.py
- black/const.py
- black/debug.py
- black/files.py
- black/handle_ipynb_magics.py
- black/linegen.py
- black/lines.py
- black/mode.py
- black/nodes.py
- black/numerics.py
- black/output.py
- black/parsing.py
- black/py.typed
- black/report.py
- black/rusty.py
- black/strings.py
- black/trans.py
Package Classes
- <enum ‘Changed’>
- <class ‘black.lines.EmptyLineTracker’>
- <class ‘str’>
- <enum ‘Enum’>
- <class ‘concurrent.futures._base.Executor’>
- <enum ‘Feature’>
- <class ‘str’>
- <class ‘black.mode.Mode’>
- <class ‘pathspec.patterns.gitwildmatch.GitWildMatchPatternError’>
- <class ‘black.parsing.InvalidInput’>
- <class ‘json.decoder.JSONDecodeError’>
- <class ‘blib2to3.pytree.Leaf’>
- <class ‘black.lines.Line’>
- <class ‘black.linegen.LineGenerator’>
- <class ‘black.mode.Mode’>
- <class ‘str’>
- <class ‘blib2to3.pytree.Node’>
- <class ‘black.report.NothingChanged’>
- <class ‘pathlib.Path’>
- <class ‘concurrent.futures.process.ProcessPoolExecutor’>
- <class ‘black.report.Report’>
- <enum ‘TargetVersion’>
- <class ‘concurrent.futures.thread.ThreadPoolExecutor’>
- <enum ‘WriteBack’>
- <class ‘datetime.datetime’>
Package Methods
black.assert_equivalent()
black.assert_equivalent(src: str, dst: str, *, pass_num: int = 1) -> None
black.assert_equivalent(
src,
dst,
pass_num,
)
black.assert_stable()
black.assert_stable(src: str, dst: str, mode: black.mode.Mode) -> None
black.assert_stable(
src,
dst,
mode,
)
black.cancel()
black.cancel(tasks: Iterable[ForwardRef('asyncio.Task[Any]')]) -> None
black.cancel(
tasks,
)
black.check_stability_and_equivalence()
Raise AssertionError if source and destination contents are not equivalent, or if a second pass of the formatter would format the content differently.
black.check_stability_and_equivalence(src_contents: str, dst_contents: str, *, mode: black.mode.Mode) -> None
black.check_stability_and_equivalence(
src_contents,
dst_contents,
mode,
)
black.color_diff()
black.color_diff(contents: str) -> str
black.color_diff(
contents,
)
black.contextmanager()
black.contextmanager(func)
black.contextmanager(
func,
)
black.decode_bytes()
newline
is either CRLF or LF but decoded_contents
is decoded with
universal newlines (i.e. only contains LF).
black.decode_bytes(src: bytes) -> Tuple[str, str, str]
black.decode_bytes(
src,
)
black.detect_target_versions()
black.detect_target_versions(node: blib2to3.pytree.Node) -> Set[black.mode.TargetVersion]
black.detect_target_versions(
node,
)
black.diff()
black.diff(a: str, b: str, a_name: str, b_name: str) -> str
black.diff(
a,
b,
a_name,
b_name,
)
black.dump_to_file()
black.dump_to_file(*output: str, ensure_final_newline: bool = True) -> str
black.dump_to_file(
output,
ensure_final_newline,
)
black.err()
black.err(message: Optional[str] = None, nl: bool = True, **styles: Any) -> None
black.err(
message,
nl,
styles,
)
black.filter_cached()
The first contains paths of files that modified on disk or are not in the cache. The other contains paths to non-modified files.
black.filter_cached(cache: Dict[str, Tuple[float, int]], sources: Iterable[pathlib.Path]) -> Tuple[Set[pathlib.Path], Set[pathlib.Pa
th]]
black.filter_cached(
cache,
sources,
)
black.find_pyproject_toml()
black.find_pyproject_toml(path_search_start: Tuple[str, ...]) -> Optional[str]
black.find_pyproject_toml(
path_search_start,
)
black.format_cell()
General idea is:
- if cell has trailing semicolon, remove it;
- if cell has IPython magics, mask them;
- format cell;
- reinstate IPython magics;
- reinstate trailing semicolon (if originally present);
- strip trailing newlines.
Cells with syntax errors will not be processed, as they could potentially be automagics or multi-line magics, which are currently not supported.
black.format_cell(src: str, *, fast: bool, mode: black.mode.Mode) -> str
black.format_cell(
src,
fast,
mode,
)
black.format_file_contents()
If fast
is False, additionally confirm that the reformatted code is
valid by calling :func:assert_equivalent
and :func:assert_stable
on it.
mode
is passed to :func:format_str
.
black.format_file_contents(src_contents: str, *, fast: bool, mode: black.mode.Mode) -> str
black.format_file_contents(
src_contents,
fast,
mode,
)
black.format_file_in_place()
If write_back
is DIFF, write a diff to stdout. If it is YES, write reformatted
code to the file.
mode
and fast
options are passed to :func:format_file_contents
.
black.format_file_in_place(src: pathlib.Path, fast: bool, mode: black.mode.Mode, write_back: black.WriteBack = <WriteBack.NO: 0>, lo
ck: Any = None) -> bool
black.format_file_in_place(
src,
fast,
mode,
write_back,
lock,
)
black.format_ipynb_string()
Operate cell-by-cell, only on code cells, only for Python notebooks.
If the .ipynb
originally had a trailing newline, it’ll be preserved.
black.format_ipynb_string(src_contents: str, *, fast: bool, mode: black.mode.Mode) -> str
black.format_ipynb_string(
src_contents,
fast,
mode,
)
black.format_stdin_to_stdout()
If content is None, it’s read from sys.stdin.
If write_back
is YES, write reformatted code back to stdout. If it is DIFF,
write a diff to stdout. The mode
argument is passed to
black.format_stdin_to_stdout(fast: bool, *, content: Optional[str] = None, write_back: black.WriteBack = <WriteBack.NO: 0>, mode: bl
ack.mode.Mode) -> bool
black.format_stdin_to_stdout(
fast,
content,
write_back,
mode,
)
black.format_str()
mode
determines formatting options, such as how many characters per line are
allowed. Example:
>>> import black >>> print(black.format_str(“def f(arg:str=’’)->None:…”, mode=black.Mode())) def f(arg: str = “”) -> None: …
A more complex example:
>>> print( … black.format_str( … “def f(arg:str=’’)->None: hey”, … mode=black.Mode( … target_versions={black.TargetVersion.PY36}, … line_length=10, … string_normalization=False, … is_pyi=False, … ), … ), … ) def f( arg: str = ‘’, ) -> None: hey
black.format_str(src_contents: str, *, mode: black.mode.Mode) -> str
black.format_str(
src_contents,
mode,
)
black.gen_python_files()
exclude_regex
, extend_exclude
, or force_exclude
regexes,
but are included by the include
regex.
Symbolic links pointing outside of the root
directory are ignored.
report
is where output about exclusions goes.
black.gen_python_files(paths: Iterable[pathlib.Path], root: pathlib.Path, include: Pattern[str], exclude: Pattern[str], extend_exclu
de: Optional[Pattern[str]], force_exclude: Optional[Pattern[str]], report: black.report.Report, gitignore: Optional[pathspec.pathspe
c.PathSpec], *, verbose: bool, quiet: bool) -> Iterator[pathlib.Path]
black.gen_python_files(
paths,
root,
include,
exclude,
extend_exclude,
force_exclude,
report,
gitignore,
verbose,
quiet,
)
black.get_cache_info()
black.get_cache_info(path: pathlib.Path) -> Tuple[float, int]
black.get_cache_info(
path,
)
black.get_features_used()
Currently looking for:
- f-strings;
- underscores in numeric literals;
- trailing commas after * or ** in function signatures and calls;
- positional only arguments in function signatures and lambdas;
- assignment expression;
- relaxed decorator syntax;
- print / exec statements;
black.get_features_used(node: blib2to3.pytree.Node) -> Set[black.mode.Feature]
black.get_features_used(
node,
)
black.get_future_imports()
black.get_future_imports(node: blib2to3.pytree.Node) -> Set[str]
black.get_future_imports(
node,
)
black.get_sources()
black.get_sources(*, ctx: click.core.Context, src: Tuple[str, ...], quiet: bool, verbose: bool, include: Pattern[str], exclude: Opti
onal[Pattern[str]], extend_exclude: Optional[Pattern[str]], force_exclude: Optional[Pattern[str]], report: 'Report', stdin_filename:
Optional[str]) -> Set[pathlib.Path]
black.get_sources(
ctx,
src,
quiet,
verbose,
include,
exclude,
extend_exclude,
force_exclude,
report,
stdin_filename,
)
black.ipynb_diff()
black.ipynb_diff(a: str, b: str, a_name: str, b_name: str) -> str
black.ipynb_diff(
a,
b,
a_name,
b_name,
)
black.is_simple_decorator_expression()
This function takes the node of the ’namedexpr_test’ of the new decorator grammar and test if it would be valid under the old decorator grammar.
The old grammar was: decorator: @ dotted_name [arguments] NEWLINE The new grammar is : decorator: @ namedexpr_test NEWLINE
black.is_simple_decorator_expression(node: Union[blib2to3.pytree.Leaf, blib2to3.pytree.Node]) -> bool
black.is_simple_decorator_expression(
node,
)
black.lib2to3_parse()
black.lib2to3_parse(src_txt: str, target_versions: Iterable[black.mode.TargetVersion] = ()) -> blib2to3.pytree.Node
black.lib2to3_parse(
src_txt,
target_versions,
)
black.mask_cell()
For example,
%matplotlib inline
'foo'
becomes
"25716f358c32750e"
'foo'
The replacements are returned, along with the transformed code.
black.mask_cell(src: str) -> Tuple[str, List[black.handle_ipynb_magics.Replacement]]
black.mask_cell(
src,
)
black.maybe_install_uvloop()
This is called only from command-line entry points to avoid interfering with the parent process if Black is used as a library.
black.maybe_install_uvloop() -> None
black.maybe_install_uvloop(
)
black.mypyc_attr()
black.mypyc_attr(*attrs, **kwattrs)
black.mypyc_attr(
attrs,
kwattrs,
)
black.normalize_fmt_off()
black.normalize_fmt_off(node: blib2to3.pytree.Node) -> None
black.normalize_fmt_off(
node,
)
black.normalize_path_maybe_ignore()
report
is where “path ignored” output goes.
black.normalize_path_maybe_ignore(path: pathlib.Path, root: pathlib.Path, report: black.report.Report) -> Optional[str]
black.normalize_path_maybe_ignore(
path,
root,
report,
)
black.nullcontext()
To be used like nullcontext
in Python 3.7.
black.nullcontext() -> Iterator[NoneType]
black.nullcontext(
)
black.out()
black.out(message: Optional[str] = None, nl: bool = True, **styles: Any) -> None
black.out(
message,
nl,
styles,
)
black.parse_ast()
black.parse_ast(src: str) -> ast.AST
black.parse_ast(
src,
)
black.parse_pyproject_toml()
If parsing fails, will raise a tomli.TOMLDecodeError
black.parse_pyproject_toml(path_config: str) -> Dict[str, Any]
black.parse_pyproject_toml(
path_config,
)
black.patch_click()
On certain misconfigured environments, Python 3 selects the ASCII encoding as the default which restricts paths that it can access during the lifetime of the application. Click refuses to work in this scenario by raising a RuntimeError.
In case of Black the likelihood that non-ASCII characters are going to be used in file paths is minimal since it’s Python source code. Moreover, this crash was spurious on Python 3.7 thanks to PEP 538 and PEP 540.
black.patch_click() -> None
black.patch_click(
)
black.patched_main()
black.patched_main() -> None
black.patched_main(
)
black.path_empty()
black.path_empty(src: Sized, msg: str, quiet: bool, verbose: bool, ctx: click.core.Context) -> None
black.path_empty(
src,
msg,
quiet,
verbose,
ctx,
)
black.put_trailing_semicolon_back()
Mirrors the logic in quiet
from IPython.core.displayhook
, but uses
tokenize_rt
so that round-tripping works fine.
black.put_trailing_semicolon_back(src: str, has_trailing_semicolon: bool) -> str
black.put_trailing_semicolon_back(
src,
has_trailing_semicolon,
)
black.re_compile_maybe_verbose()
If it contains newlines, use verbose mode.
black.re_compile_maybe_verbose(regex: str) -> Pattern[str]
black.re_compile_maybe_verbose(
regex,
)
black.read_cache()
If it is not well formed, the call to write_cache later should resolve the issue.
black.read_cache(mode: black.mode.Mode) -> Dict[str, Tuple[float, int]]
black.read_cache(
mode,
)
black.read_pyproject_toml()
Returns the path to a successfully found and read configuration file, None otherwise.
black.read_pyproject_toml(ctx: click.core.Context, param: click.core.Parameter, value: Optional[str]) -> Optional[str]
black.read_pyproject_toml(
ctx,
param,
value,
)
black.reformat_code()
Similar to reformat_one
, but for string content.
fast
, write_back
, and mode
options are passed to
black.reformat_code(content: str, fast: bool, write_back: black.WriteBack, mode: black.mode.Mode, report: black.report.Report) -> No
ne
black.reformat_code(
content,
fast,
write_back,
mode,
report,
)
black.reformat_many()
black.reformat_many(sources: Set[pathlib.Path], fast: bool, write_back: black.WriteBack, mode: black.mode.Mode, report: 'Report', wo
rkers: Optional[int]) -> None
black.reformat_many(
sources,
fast,
write_back,
mode,
report,
workers,
)
black.reformat_one()
fast
, write_back
, and mode
options are passed to
black.reformat_one(src: pathlib.Path, fast: bool, write_back: black.WriteBack, mode: black.mode.Mode, report: 'Report') -> None
black.reformat_one(
src,
fast,
write_back,
mode,
report,
)
black.remove_trailing_semicolon()
For example,
fig, ax = plt.subplots()
ax.plot(x_data, y_data); # plot data
would become
fig, ax = plt.subplots()
ax.plot(x_data, y_data) # plot data
Mirrors the logic in quiet
from IPython.core.displayhook
, but uses
tokenize_rt
so that round-tripping works fine.
black.remove_trailing_semicolon(src: str) -> Tuple[str, bool]
black.remove_trailing_semicolon(
src,
)
black.replace()
This is especially useful for frozen classes. Example usage:
@dataclass(frozen=True) class C: x: int y: int
c = C(1, 2) c1 = replace(c, x=3) assert c1.x == 3 and c1.y == 2
black.replace(obj, /, **changes)
black.replace(
obj,
changes,
)
black.schedule_formatting()
(Use ProcessPoolExecutors for actual parallelism.)
write_back
, fast
, and mode
options are passed to
black.schedule_formatting(sources: Set[pathlib.Path], fast: bool, write_back: black.WriteBack, mode: black.mode.Mode, report: 'Repor
t', loop: asyncio.events.AbstractEventLoop, executor: concurrent.futures._base.Executor) -> None
black.schedule_formatting(
sources,
fast,
write_back,
mode,
report,
loop,
executor,
)
black.shutdown()
black.shutdown(loop: asyncio.events.AbstractEventLoop) -> None
black.shutdown(
loop,
)
black.stringify_ast()
black.stringify_ast(node: ast.AST, depth: int = 0) -> Iterator[str]
black.stringify_ast(
node,
depth,
)
black.supports_feature()
black.supports_feature(target_versions: Set[black.mode.TargetVersion], feature: black.mode.Feature) -> bool
black.supports_feature(
target_versions,
feature,
)
black.target_version_option_callback()
This is its own function because mypy couldn’t infer the type correctly when it was a lambda, causing mypyc trouble.
black.target_version_option_callback(c: click.core.Context, p: Union[click.core.Option, click.core.Parameter], v: Tuple[str, ...]) -
> List[black.mode.TargetVersion]
black.target_version_option_callback(
c,
p,
v,
)
black.transform_line()
They should fit in the allotted line_length
but might not be able to.
features
are syntactical features that may be used in the output.
black.transform_line(line: black.lines.Line, mode: black.mode.Mode, features: Collection[black.mode.Feature] = ()) -> Iterator[black
.lines.Line]
black.transform_line(
line,
mode,
features,
)
black.unmask_cell()
For example
"9b20"
foo = bar
becomes
%%time
foo = bar
black.unmask_cell(src: str, replacements: List[black.handle_ipynb_magics.Replacement]) -> str
black.unmask_cell(
src,
replacements,
)
black.validate_cell()
or non-Python cell magics, which might cause tokenizer_rt to break because of indentations.
If a cell contains ``!ls``, then it'll be transformed to
``get_ipython().system('ls')``. However, if the cell originally contained
``get_ipython().system('ls')``, then it would get transformed in the same way:
\>\>\> TransformerManager().transform_cell("get_ipython().system('ls')")
"get_ipython().system('ls')
" >>> TransformerManager().transform_cell("!ls") “get_ipython().system(’ls’) "
Due to the impossibility of safely roundtripping in such situations, cells
containing transformed magics will be ignored.
black.validate_cell(src: str) -> None
black.validate_cell(
src,
)
black.validate_metadata()
All notebook metadata fields are optional, see https://nbformat.readthedocs.io/en/latest/format_description.html. So if a notebook has empty metadata, we will try to parse it anyway.
black.validate_metadata(nb: MutableMapping[str, Any]) -> None
black.validate_metadata(
nb,
)
black.validate_regex()
black.validate_regex(ctx: click.core.Context, param: click.core.Parameter, value: Optional[str]) -> Optional[Pattern[str]]
black.validate_regex(
ctx,
param,
value,
)
black.wrap_stream_for_windows()
If colorama
is unavailable, the original stream is returned unmodified.
Otherwise, the wrap_stream()
function determines whether the stream needs
to be wrapped for a Windows environment and will accordingly either return
an AnsiToWin32
wrapper or the original stream.
black.wrap_stream_for_windows(f: _io.TextIOWrapper) -> Union[_io.TextIOWrapper, ForwardRef('colorama.AnsiToWin32')]
black.wrap_stream_for_windows(
f,
)
black.write_cache()
black.write_cache(cache: Dict[str, Tuple[float, int]], sources: Iterable[pathlib.Path], mode: black.mode.Mode) -> None
black.write_cache(
cache,
sources,
mode,
)