HPI/my/core/util.py

235 lines
7 KiB
Python

from pathlib import Path
from itertools import chain
from importlib import import_module
import os
import pkgutil
import re
import sys
from typing import List, Iterable, NamedTuple, Optional
class HPIModule(NamedTuple):
name: str
skip_reason: Optional[str]
doc: Optional[str] = None
file: Optional[Path] = None
def modules() -> Iterable[HPIModule]:
import my
for m in _iter_all_importables(my):
yield m
def ignored(m: str) -> bool:
excluded = [
# legacy stuff left for backwards compatibility
'core.*',
'config.*',
]
exs = '|'.join(excluded)
return re.match(f'^my.({exs})$', m) is not None
def get_stats(module: str):
# todo detect via ast?
try:
mod = import_module(module)
except Exception as e:
return None
return getattr(mod, 'stats', None)
__NOT_HPI_MODULE__ = 'Import this to mark a python file as a helper, not an actual HPI module'
def has_not_module_flag(module: str) -> bool:
# if module == 'my.books.kobo':
# breakpoint()
# pass
try:
mod = import_module(module)
except Exception as e:
return False
return any(x is __NOT_HPI_MODULE__ for x in vars(mod).values())
def is_not_hpi_module(module: str) -> Optional[str]:
# None if a module, otherwise returns reason
if has_not_module_flag(module):
return "marked explicitly (via __NOT_HPI_MODULE__)"
stats = get_stats(module)
if stats is None:
return "has no 'stats()' function"
return None
# todo reuse in readme/blog post
# borrowed from https://github.com/sanitizers/octomachinery/blob/24288774d6dcf977c5033ae11311dbff89394c89/tests/circular_imports_test.py#L22-L55
def _iter_all_importables(pkg) -> Iterable[HPIModule]:
# todo crap. why does it include some stuff three times??
yield from chain.from_iterable(
_discover_path_importables(Path(p), pkg.__name__)
# todo might need to handle __path__ for individual modules too?
# not sure why __path__ was duplicated, but it did happen..
for p in set(pkg.__path__)
)
def _discover_path_importables(pkg_pth, pkg_name) -> Iterable[HPIModule]:
from .core_config import config
"""Yield all importables under a given path and package."""
for dir_path, dirs, file_names in os.walk(pkg_pth):
file_names.sort()
# NOTE: sorting dirs in place is intended, it's the way you're supposed to do it with os.walk
dirs.sort()
pkg_dir_path = Path(dir_path)
if pkg_dir_path.parts[-1] == '__pycache__':
continue
if all(Path(_).suffix != '.py' for _ in file_names):
continue
rel_pt = pkg_dir_path.relative_to(pkg_pth)
pkg_pref = '.'.join((pkg_name, ) + rel_pt.parts)
yield from _walk_packages(
(str(pkg_dir_path), ), prefix=f'{pkg_pref}.',
)
# TODO might need to make it defensive and yield Exception (otherwise hpi doctor might fail for no good reason)
# use onerror=?
# ignored explicitly -> not HPI
# if enabled in config -> HPI
# if disabled in config -> HPI
# otherwise, check for stats
# recursion is relied upon using .*
# TODO when do we need to recurse?
def _walk_packages(path=None, prefix='', onerror=None) -> Iterable[HPIModule]:
'''
Modified version of https://github.com/python/cpython/blob/d50a0700265536a20bcce3fb108c954746d97625/Lib/pkgutil.py#L53,
to alvoid importing modules that are skipped
'''
from .core_config import config
def seen(p, m={}):
if p in m:
return True
m[p] = True
for info in pkgutil.iter_modules(path, prefix):
mname = info.name
if ignored(mname):
# not sure if need to yield?
continue
active = config._is_module_active(mname)
skip_reason = None
if active is False:
skip_reason = 'suppressed in the user config'
elif active is None:
# unspecified by the user, rely on other means
is_not_module = is_not_hpi_module(mname)
if is_not_module is not None:
skip_reason = is_not_module
else: # active is True
# nothing to do, enabled explicitly
pass
yield HPIModule(
name=mname,
skip_reason=skip_reason,
)
if not info.ispkg:
continue
recurse = config._is_module_active(mname + '.')
if not recurse:
continue
try:
__import__(mname)
except ImportError:
if onerror is not None:
onerror(mname)
except Exception:
if onerror is not None:
onerror(mname)
else:
raise
else:
path = getattr(sys.modules[mname], '__path__', None) or []
# don't traverse path items we've seen before
path = [p for p in path if not seen(p)]
yield from _walk_packages(path, mname+'.', onerror)
def modules_via_ast() -> Iterable[HPIModule]:
'''
Experimental version, which isn't importing the modules, making it more robust and safe.
'''
import ast
my_root = Path(__file__).absolute().parent.parent
for f in sorted(my_root.rglob('*.py')):
if f.is_symlink():
continue # meh
mp = f.relative_to(my_root.parent)
if mp.name == '__init__.py':
mp = mp.parent
m = str(mp.with_suffix('')).replace('/', '.')
if ignored(m):
continue
a = ast.parse(f.read_text())
NM = '__NOT_HPI_MODULE__'
is_not_module = any(
getattr(node, 'name', None) == NM # direct definition
or
any(getattr(n, 'name', None) == NM for n in getattr(node, 'names', [])) # import from
for node in a.body)
if is_not_module:
continue
doc = ast.get_docstring(a, clean=False)
yield HPIModule(
name=m,
skip_reason=None,
doc=doc,
file=f.relative_to(my_root.parent), # todo not sure if should be relative
)
# deprecate?
def get_modules() -> List[HPIModule]:
return list(modules())
### tests start
## FIXME: add test when there is an import error -- should be defensive and yield exception
def test_module_detection() -> None:
from .core_config import _reset_config as reset
with reset() as cc:
cc.disabled_modules = ['my.location.*', 'my.body.*', 'my.workouts.*', 'my.private.*']
mods = {m.name: m for m in modules()}
assert mods['my.demo'] .skip_reason == "has no 'stats()' function"
with reset() as cc:
cc.disabled_modules = ['my.location.*', 'my.body.*', 'my.workouts.*', 'my.private.*', 'my.lastfm']
cc.enabled_modules = ['my.demo']
mods = {m.name: m for m in modules()}
assert mods['my.demo'] .skip_reason is None # not skipped
assert mods['my.lastfm'].skip_reason == "suppressed in the user config"
### tests end