Files
sqlalchemy/tools/generate_tuple_map_overloads.py
T
Mike Bayer ad11c482e2 pep484 ORM / SQL result support
after some experimentation it seems mypy is more amenable
to the generic types being fully integrated rather than
having separate spin-off types.   so key structures
like Result, Row, Select become generic.  For DML
Insert, Update, Delete, these are spun into type-specific
subclasses ReturningInsert, ReturningUpdate, ReturningDelete,
which is fine since the "row-ness" of these constructs
doesn't happen until returning() is called in any case.

a Tuple based model is then integrated so that these
objects can carry along information about their return
types.  Overloads at the .execute() level carry through
the Tuple from the invoked object to the result.

To suit the issue of AliasedClass generating attributes
that are dynamic, experimented with a custom subclass
AsAliased, but then just settled on having aliased()
lie to the type checker and return `Type[_O]`, essentially.
will need some type-related accessors for with_polymorphic()
also.

Additionally, identified an issue in Update when used
"mysql style" against a join(), it basically doesn't work
if asked to UPDATE two tables on the same column name.
added an error message to the specific condition where
it happens with a very non-specific error message that we
hit a thing we can't do right now, suggest multi-table
update as a possible cause.

Change-Id: I5eff7eefe1d6166ee74160b2785c5e6a81fa8b95
2022-04-27 14:46:36 -04:00

175 lines
5.2 KiB
Python

r"""Generate tuple mapping overloads.
the problem solved by this script is that of there's no way in current
pep-484 typing to unpack \*args: _T into Tuple[_T]. pep-646 is the first
pep to provide this, but it doesn't work for the actual Tuple class
and also mypy does not have support for pep-646 as of yet. Better pep-646
support would allow us to use a TypeVarTuple with Unpack, but TypeVarTuple
does not have support for sequence operations like ``__getitem__`` and
iteration; there's also no way for TypeVarTuple to be translated back to a
Tuple which does have those things without a combinatoric hardcoding approach
to each length of tuple.
So here, the script creates a map from `*args` to a Tuple directly using a
combinatoric generated code approach.
.. versionadded:: 2.0
"""
from __future__ import annotations
from argparse import ArgumentParser
import importlib
import os
from pathlib import Path
import re
import shutil
import sys
from tempfile import NamedTemporaryFile
import textwrap
from sqlalchemy.util.langhelpers import console_scripts
is_posix = os.name == "posix"
sys.path.append(str(Path(__file__).parent.parent))
def process_module(modname: str, filename: str) -> str:
# use tempfile in same path as the module, or at least in the
# current working directory, so that black / zimports use
# local pyproject.toml
with NamedTemporaryFile(
mode="w", delete=False, suffix=".py", dir=Path(filename).parent
) as buf, open(filename) as orig_py:
indent = ""
in_block = False
current_fnname = given_fnname = None
for line in orig_py:
m = re.match(
r"^( *)# START OVERLOADED FUNCTIONS ([\.\w_]+) ([\w_]+) (\d+)-(\d+)$", # noqa: E501
line,
)
if m:
indent = m.group(1)
given_fnname = current_fnname = m.group(2)
if current_fnname.startswith("self."):
use_self = True
current_fnname = current_fnname.split(".")[1]
else:
use_self = False
return_type = m.group(3)
start_index = int(m.group(4))
end_index = int(m.group(5))
sys.stderr.write(
f"Generating {start_index}-{end_index} overloads "
f"attributes for "
f"class {'self.' if use_self else ''}{current_fnname} "
f"-> {return_type}\n"
)
in_block = True
buf.write(line)
buf.write(
"\n # code within this block is "
"**programmatically, \n"
" # statically generated** by"
f" tools/{os.path.basename(__file__)}\n\n"
)
for num_args in range(start_index, end_index + 1):
combinations = [
[
f"__ent{arg}: _TCCA[_T{arg}]"
for arg in range(num_args)
]
]
for combination in combinations:
buf.write(
textwrap.indent(
f"""
@overload
def {current_fnname}(
{'self, ' if use_self else ''}{", ".join(combination)}
) -> {return_type}[Tuple[{', '.join(f'_T{i}' for i in range(num_args))}]]:
...
""", # noqa: E501
indent,
)
)
if in_block and line.startswith(
f"{indent}# END OVERLOADED FUNCTIONS {given_fnname}"
):
in_block = False
if not in_block:
buf.write(line)
return buf.name
def run_module(modname, stdout):
sys.stderr.write(f"importing module {modname}\n")
mod = importlib.import_module(modname)
filename = destination_path = mod.__file__
assert filename is not None
tempfile = process_module(modname, filename)
ignore_output = stdout
console_scripts(
str(tempfile),
{"entrypoint": "zimports"},
ignore_output=ignore_output,
)
console_scripts(
str(tempfile),
{"entrypoint": "black"},
ignore_output=ignore_output,
)
if stdout:
with open(tempfile) as tf:
print(tf.read())
os.unlink(tempfile)
else:
sys.stderr.write(f"Writing {destination_path}...\n")
shutil.move(tempfile, destination_path)
def main(args):
for modname in entries:
if args.module in {"all", modname}:
run_module(modname, args.stdout)
entries = [
"sqlalchemy.sql._selectable_constructors",
"sqlalchemy.orm.session",
"sqlalchemy.orm.query",
"sqlalchemy.sql.selectable",
"sqlalchemy.sql.dml",
]
if __name__ == "__main__":
parser = ArgumentParser()
parser.add_argument(
"--module",
choices=entries + ["all"],
default="all",
help="Which file to generate. Default is to regenerate all files",
)
parser.add_argument(
"--stdout",
action="store_true",
help="Write to stdout instead of saving to file",
)
args = parser.parse_args()
main(args)