Algebraic Data Types 2018-06-04
In this post:
- I start thinking about music theory, composition, synthesis, then veer off into some hobby work from months ago.
I'm going to try to speed through setting up intervals and such a little, because I want to try scraping together a prototype for writing MIDI files. Long term, I'd like to try doing my own synthesis because that seems cool, but it's easier to synthesize a note when you've decided just how a note is represented.
Ugh, thinking about how to do this, I feel like I'm getting bitten by the fact that I never did get my ADT code into a workable state. I think I'll switch gears a little, to getting that working. I've come to the conclusion that I want to try using dataclasses-style class decorators, because attempting to use metaclasses was really unpleasant. I'd like to paste some old code to help make sense of all the stuff I'm saying, but the old code I have doesn't really make sense. Instead, I'll sketch out how I'd like to use the new thing.
1 2 3 4 5 6 7 8 9 10 11 12 | from typing import Generic, TypeVar from wherever import algebraic R = typing.TypeVar('R') E = typing.TypeVar('E') @algebraic class Result(Generic[R, E]): Ok: R Err: E |
The decorator should do most of the same work as the dataclass decorator, but with some differences:
- The instances are always immutable, and backed by tuples.
- The generated __new__ method is a factory, which creates instances of generated classes that are a subclass of the decorated class.
- No field customization; the types are the types.
- No subclassing.
Let's see what methods I want to generate:
- __new__
- __repr__ (if not present)
- comparison methods, with similar logic to dataclasses
- hashing, with a cut-down version of the logic
- probably an __init_subclass__
- something with docstrings
Let's get started on this.
Okay, I've got code that resembles a stripped-down version of dataclasses, but because the layout is more constrained, there's less generated code. I think for now, I'm going to avoid flexibility until I discover I need it, so all algebraic classes will be comparable, orderable, and hashable with no control. Consumers will be able to tweak the repr logic, because that's not a big deal.
Actually, I really should put the customization in, but I want to consider carefully how this is going to work. If the algebraic code defines equality, then it "might as well" define hashing. If the algebraic code does not define equality, then it should not define hashing. I'm currently using equality in some pattern-matching code, but I could just use the functions directly... Except actually, it should be fine. What I'm going to do differently is, I'm going to say that the methods get added in bundles.
Except I just realized that my implementation of ordering is very wrong. When I get this into a repository, I'm going to strip it out.
Regardless, aside from that hiccup, I'm sure the code I wrote over the course of two days is perfect in every way.
... Okay, so I need to use the descriptor protocol to get at the name of wrapped methods. Fine.
... Okay, it coughs up bizarre errors if I attempt to actually instantiate generic classes. It looks like this is a result of some kind of optimization attempt that bypasses the normal MRO. The problem appears to be that I'm creating subclasses of a generic class, that add a built-in class as a superclass. I believe I can hack around this by looking for an attribute, and changing it if it exists.
I was incorrect. The problem was that I wasn't explicitly delegating __new__ calls through a base class, which I think confused things more than it should have? Anyway, next I removed the __init_subclass__ stuff, because I forgot that specializing a generic class in Python subclasses it. ... Actually, I can keep that in, I just need to restrict the scope.
All right, here's where things ended up. My main issue right now is, I made this in part to work well with typing information, and mypy reacts to my carefully crafted annotations like I threw garbage in its face. I can't actually use this in my projects until I work out how to alter mypy's interpretation of it.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 | """Class decorator for defining abstract data types.""" import collections import keyword import typing def _as_tuple(value) -> tuple: """Convert non-tuple values to a 1-tuple, return tuples unchanged.""" if not isinstance(value, tuple): value = (value,) return value def _name(cls, function) -> str: """Return the name of a function accessed through a descriptor.""" return function.__get__(None, cls).__name__ def _set_new_functions(cls, *functions) -> typing.Optional[str]: """Attempt to set the attributes corresponding to the functions on cls. If any attributes are already defined, fail before setting, and return the already-defined name. """ for function in functions: if _name(cls, function) in cls.__dict__: return _name(cls, function) for function in functions: setattr(cls, _name(cls, function), function) return None class MatchFailure(BaseException): """An exception that signals a failure in ADT matching.""" def desugar(constructor: type, instance: tuple) -> tuple: """Return the inside of an ADT instance, given its constructor.""" # I really do want to match exactly, I swear. if type(instance) is not constructor: raise MatchFailure return tuple.__getitem__(instance, slice(None)) def _unpack(instance: tuple) -> tuple: """Return the inside of any ADT instance. This function is not meant for general use. """ return desugar(type(instance), instance) def _algebraic_base(obj): return getattr(obj.__class__, '__algebraic_base__', None) class AlgebraicConstructor: """Base class for ADT Constructor classes.""" __slots__ = () def __new__(cls, *args, **kwargs): """Explicitly forward to base class.""" return super().__new__(cls, *args, **kwargs) def __repr__(self): return self.__class__.__qualname__ + ( f'({", ".join(repr(item) for item in _unpack(self))})') def __eq__(self, other): if other.__class__ is self.__class__: return _unpack(self) == _unpack(other) if _algebraic_base(other) is _algebraic_base(self): return False return NotImplemented def __ne__(self, other): if other.__class__ is self.__class__: return _unpack(self) != _unpack(other) if _algebraic_base(other) is _algebraic_base(self): return True return NotImplemented def __lt__(self, other): if other.__class__ is self.__class__: return _unpack(self) < _unpack(other) if _algebraic_base(other) is _algebraic_base(self): order = _algebraic_base(self).__subclass_order__ self_index = order.index(self.__class__) other_index = order.index(other.__class__) return self_index < other_index return NotImplemented def __le__(self, other): if other.__class__ is self.__class__: return _unpack(self) <= _unpack(other) if _algebraic_base(other) is _algebraic_base(self): order = _algebraic_base(self).__subclass_order__ self_index = order.index(self.__class__) other_index = order.index(other.__class__) return self_index <= other_index return NotImplemented def __gt__(self, other): if other.__class__ is self.__class__: return _unpack(self) > _unpack(other) if _algebraic_base(other) is _algebraic_base(self): order = _algebraic_base(self).__subclass_order__ self_index = order.index(self.__class__) other_index = order.index(other.__class__) return self_index > other_index return NotImplemented def __ge__(self, other): if other.__class__ is self.__class__: return _unpack(self) >= _unpack(other) if _algebraic_base(other) is _algebraic_base(self): order = _algebraic_base(self).__subclass_order__ self_index = order.index(self.__class__) other_index = order.index(other.__class__) return self_index >= other_index return NotImplemented def __hash__(self): return hash(_unpack(self)) def _make_constructor(_cls, name, length, subclasses, subclass_order): class Constructor(_cls, AlgebraicConstructor, tuple): """Auto-generated subclass of an ADT.""" __slots__ = () __algebraic_base__ = _cls def __new__(cls, *args): if len(args) != length: raise ValueError return super().__new__(cls, args) Constructor.__name__ = name Constructor.__qualname__ = f'{_cls.__qualname__}.{name}' subclasses.add(Constructor) setattr(_cls, name, Constructor) subclass_order.append(Constructor) def _process_class(_cls, _repr, eq, order): if order and not eq: raise ValueError('eq must be true if order is true') lengths = {} subclasses = set() subclass_order = [] for cls in reversed(_cls.__mro__): for key, value in getattr(cls, '__annotations__', {}).items(): lengths[key] = len(_as_tuple(value)) for name, length in lengths.items(): _make_constructor(_cls, name, length, subclasses, subclass_order) @classmethod def __init_subclass__(cls, **kwargs): if issubclass(cls, tuple(subclasses)): raise TypeError # Allow it to go through otherwise, because Generic. return super(_cls, cls).__init_subclass__(**kwargs) _cls.__init_subclass__ = __init_subclass__ @staticmethod def __new__(cls, args): if cls not in subclasses: raise TypeError return super(_cls, cls).__new__(cls, args) if _set_new_functions(_cls, __new__): base__new__ = _cls.__new__ @staticmethod def __new__(cls, args): if cls not in subclasses: raise TypeError return base__new__(cls, args) _cls.__new__ = __new__ if _repr: _set_new_functions(_cls, __repr__) equality_methods_were_set = False if eq: equality_methods_were_set = not _set_new_functions( _cls, __eq__, __ne__) if equality_methods_were_set: _cls.__hash__ = __hash__ if order: if not equality_methods_were_set: raise ValueError( "Can't add ordering methods if equality methods are provided.") collision = _set_new_functions(_cls, __lt__, __le__, __gt__, __ge__) if collision: raise TypeError(f'Cannot overwrite attribute {collision} ' f'in class {_cls.__name__}. Consider using ' 'functools.total_ordering') _cls.__subclass_order__ = tuple(subclass_order) return _cls def algebraic(_cls=None, *, repr=True, eq=True, order=False): """Decorate a class to be an algebraic data type.""" def wrap(cls): """Return the processed class.""" return _process_class(cls, repr, eq, order) if _cls is None: return wrap return wrap(_cls) DISCARD = object() class Matcher(tuple): """A matcher that binds a value to a name.""" __slots__ = () def __new__(cls, name: str): if name == '_': return DISCARD if not name.isidentifier(): raise ValueError if keyword.iskeyword(name): raise ValueError return super().__new__(cls, (name,)) @property def name(self): """Return the name of the matcher.""" return self[0] def __matmul__(self, other): return AsMatcher(self, other) class AsMatcher(tuple): """A matcher that contains further bindings.""" __slots__ = () def __new__(cls, matcher: Matcher, match): if matcher is DISCARD: return match return super().__new__(cls, (matcher, match)) @property def matcher(self): """Return the left-hand-side of the as-match.""" return self[0] @property def match(self): """Return the right-hand-side of the as-match.""" return self[1] def names(target): """Return every name bound by a target.""" name_list = [] names_seen = set() to_process = [target] while to_process: item = to_process.pop() if isinstance(item, Matcher): if item.name in names_seen: raise ValueError names_seen.add(item.name) name_list.append(item.name) elif isinstance(item, AsMatcher): to_process.append(item.match) to_process.append(item.matcher) elif isinstance(item, AlgebraicConstructor): to_process.extend(reversed(_unpack(item))) elif isinstance(item, tuple): to_process.extend(reversed(item)) yield from name_list def _match(target, value): match_dict = collections.OrderedDict() to_process = [(target, value)] while to_process: target, value = to_process.pop() if target is DISCARD: pass elif isinstance(target, Matcher): if target.name in match_dict: raise ValueError match_dict[target.name] = value elif isinstance(target, AsMatcher): to_process.append((target.match, value)) to_process.append((target.matcher, value)) elif isinstance(target, AlgebraicConstructor): to_process.extend(zip(reversed(_unpack(target)), reversed(desugar(type(target), value)))) elif (isinstance(target, tuple) and target.__class__ is value.__class__ and len(target) == len(value)): to_process.extend(zip(reversed(target), reversed(value))) elif target != value: raise MatchFailure return match_dict def get_values(dct, keys): """Unpack a dict, in order.""" for key in keys: yield dct[key] class ValueMatcher: """Given a value, attempt to match against a target.""" def __init__(self, value): self.value = value self.matches = None def match(self, target): """Match against target, generating a set of bindings.""" try: self.matches = _match(target, self.value) except MatchFailure: self.matches = None return self.matches is not None if __name__ == '__main__': R = typing.TypeVar('R') E = typing.TypeVar('E') @algebraic class Result(typing.Generic[R, E]): """Experimental generic ADT.""" Ok: R Err: E my_result: Result[int, str] = Result.Ok(10) |
Next week, or possibly sooner, I try to figure out how to write a mypy plugin.