Repository navigation
Expand file tree
/
Copy pathtransform.py
More file actions
537 lines (442 loc) · 18.2 KB
/
Copy pathtransform.py
File metadata and controls
537 lines (442 loc) · 18.2 KB
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
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
"""
Transformer nodes convert many inputs into a single output, or vice-versa.
"""
from __future__ import annotations
import itertools
from abc import ABC, abstractmethod
from collections.abc import Callable
from dataclasses import MISSING
from dataclasses import dataclass as as_dataclass
from typing import Any, ClassVar
from pandas import DataFrame
from pyiron_snippets.colors import SeabornColors
from pyiron_snippets.factory import classfactory
from pyiron_workflow.channels import NOT_DATA, NotData
from pyiron_workflow.nodes.static_io import StaticNode
class Transformer(StaticNode, ABC):
"""
Transformers are a special case of :class:`StaticNode` nodes that turn many inputs
into a single output or vice-versa.
"""
@property
def color(self) -> str:
"""For drawing the graph"""
return SeabornColors.blue
class FromManyInputs(Transformer, ABC):
_output_name: ClassVar[str] # Mandatory attribute for non-abstract subclasses
_output_type_hint: ClassVar[Any] = None
# _build_inputs_preview required from parent class
# Inputs convert to `run_args` as a value dictionary
# This must be commensurate with the internal expectations of _on_run
@property
def _run_args(self) -> tuple[tuple, dict]:
return (), self.inputs.to_value_dict()
@classmethod
def _build_outputs_preview(cls) -> dict[str, Any]:
return {cls._output_name: cls._output_type_hint}
def process_run_result(self, run_output: Any | tuple) -> Any | tuple:
self.outputs[self._output_name].value = run_output
return run_output
class ToManyOutputs(Transformer, ABC):
_input_name: ClassVar[str] # Mandatory attribute for non-abstract subclasses
_input_type_hint: ClassVar[Any] = None
_input_default: ClassVar[Any | NotData] = NOT_DATA
# _build_outputs_preview still required from parent class
# Must be commensurate with the dictionary returned by transform_to_output
@abstractmethod
def _on_run(self, input_object) -> Callable[..., Any | tuple]:
"""Must take the single object to be transformed"""
@property
def _run_args(self) -> tuple[tuple, dict]:
return (self.inputs[self._input_name].value,), {}
@classmethod
def _build_inputs_preview(cls) -> dict[str, tuple[Any, Any]]:
return {cls._input_name: (cls._input_type_hint, cls._input_default)}
def process_run_result(self, run_output: dict[str, Any]) -> dict[str, Any]:
for k, v in run_output.items():
self.outputs[k].value = v
return run_output
class _HasLength(Transformer, ABC):
_length: ClassVar[int] # Mandatory attribute for non-abstract subclasses
class InputsToList(_HasLength, FromManyInputs, ABC):
_output_name: ClassVar[str] = "list"
_output_type_hint: ClassVar[Any] = list
def _on_run(self, **inputs_to_value_dict):
return list(inputs_to_value_dict.values())
@classmethod
def _build_inputs_preview(cls) -> dict[str, tuple[Any, Any]]:
return {f"item_{i}": (None, NOT_DATA) for i in range(cls._length)}
class ListToOutputs(_HasLength, ToManyOutputs, ABC):
_input_name: ClassVar[str] = "list"
_input_type_hint: ClassVar[Any] = list
def _on_run(self, input_object: list):
return {f"item_{i}": v for i, v in enumerate(input_object)}
@classmethod
def _build_outputs_preview(cls) -> dict[str, Any]:
return {f"item_{i}": None for i in range(cls._length)}
@classfactory
def inputs_to_list_factory(n: int, use_cache: bool = True, /) -> type[InputsToList]:
return ( # type: ignore[return-value]
f"{InputsToList.__name__}{n}",
(InputsToList,),
{
"_length": n,
"use_cache": use_cache,
},
{},
)
def inputs_to_list(n: int, /, *node_args, use_cache: bool = True, **node_kwargs):
"""
Creates and returns an instance of a dynamically generated :class:`InputsToList`
subclass with a specified number of inputs.
Args:
n (int): Number of input channels.
use_cache (bool): Whether this node should default to caching its values.
(Default is True.)
*node_args: Positional arguments for the node instance.
**node_kwargs: Keyword arguments for the node instance.
Returns:
InputsToList: An instance of the dynamically created :class:`InputsToList`
subclass.
"""
cls = inputs_to_list_factory(n, use_cache)
cls.preview_io()
return cls(*node_args, **node_kwargs)
@classfactory
def list_to_outputs_factory(n: int, use_cache: bool = True, /) -> type[ListToOutputs]:
return ( # type: ignore[return-value]
f"{ListToOutputs.__name__}{n}",
(ListToOutputs,),
{
"_length": n,
"use_cache": use_cache,
},
{},
)
def list_to_outputs(
n: int, /, *node_args, use_cache: bool = True, **node_kwargs
) -> ListToOutputs:
"""
Creates and returns an instance of a dynamically generated :class:`ListToOutputs`
subclass with a specified number of outputs.
Args:
n (int): Number of output channels.
use_cache (bool): Whether this node should default to caching its values.
(Default is True.)
*node_args: Positional arguments for the node instance.
**node_kwargs: Keyword arguments for the node instance.
Returns:
ListToOutputs: An instance of the dynamically created :class:`ListToOutputs`
subclass.
"""
cls = list_to_outputs_factory(n, use_cache)
cls.preview_io()
return cls(*node_args, **node_kwargs)
class InputsToDict(FromManyInputs, ABC):
_output_name: ClassVar[str] = "dict"
_output_type_hint: ClassVar[Any] = dict
_input_specification: ClassVar[
list[str] | dict[str, tuple[Any | None, Any | NotData]]
]
def _on_run(self, **inputs_to_value_dict):
return inputs_to_value_dict
@classmethod
def _build_inputs_preview(cls) -> dict[str, tuple[Any | None, Any | NotData]]:
if isinstance(cls._input_specification, list):
return dict.fromkeys(cls._input_specification, (None, NOT_DATA))
else:
return cls._input_specification
@staticmethod
def hash_specification(
input_specification: list[str] | dict[str, tuple[Any | None, Any | NotData]],
):
"""For generating unique subclass names."""
if isinstance(input_specification, list):
return hash(tuple(input_specification))
else:
flattened_tuple = tuple(
itertools.chain.from_iterable(
(key, *value) for key, value in input_specification.items()
)
)
try:
return hash(flattened_tuple)
except Exception as e:
raise ValueError(
f"To automatically generate a unique name for subclasses of "
f"{InputsToDict.__name__}, the input specification must be fully "
f"hashable, but it was not. Either pass fully hashable hints and "
f"defaults, or explicitly provide a class name suffix. Received "
f"specification: {input_specification}"
) from e
@classfactory
def inputs_to_dict_factory(
input_specification: list[str] | dict[str, tuple[Any | None, Any | NotData]],
class_name_suffix: str | None,
use_cache: bool = True,
/,
) -> type[InputsToDict]:
if class_name_suffix is None:
class_name_suffix = str(
InputsToDict.hash_specification(input_specification)
).replace("-", "m")
return ( # type: ignore[return-value]
f"{InputsToDict.__name__}{class_name_suffix}",
(InputsToDict,),
{
"_input_specification": input_specification,
"use_cache": use_cache,
},
{},
)
def inputs_to_dict(
input_specification: list[str] | dict[str, tuple[Any | None, Any | NotData]],
*node_args,
class_name_suffix: str | None = None,
use_cache: bool = True,
**node_kwargs,
):
"""
Build a new :class:`InputsToDict` subclass and instantiate it.
Tries to automatically generate a subclass name by hashing the
:param:`input_specification`. If such hashing fails, you will instead _need_ to
provide an explicit :param:`class_name_suffix`
Args:
input_specification (list[str] | dict[str, tuple[Any | None, Any | NOT_DATA]]):
The input channel names, or full input specification in the form
`{key: (type_hint, default_value))}`.
*node_args: Other args for the node instance.
class_name_suffix (str | None): The suffix to use in the class name. (Default
is None, try to generate the suffix by hashing :param:`input_specification`.
use_cache (bool): Whether this node should default to caching its values.
(Default is True.)
**node_kwargs: Other kwargs for the node instance.
Returns:
(InputsToDict): A new node for transforming inputs into a dictionary.
"""
cls = inputs_to_dict_factory(input_specification, class_name_suffix, use_cache)
cls.preview_io()
return cls(*node_args, **node_kwargs)
class InputsToDataframe(_HasLength, FromManyInputs, ABC):
"""
Turns inputs of dictionaries (all with the same keys) into a single
:class:`pandas.DataFrame`.
"""
_output_name: ClassVar[str] = "df"
_output_type_hint: ClassVar[Any] = DataFrame
def _on_run(self, *rows: dict[str, Any]) -> Any:
df_dict = {}
for i, row in enumerate(rows):
for key, value in row.items():
if i == 0:
df_dict[key] = [value]
else:
df_dict[key].append(value)
return DataFrame(df_dict)
@property
def _run_args(self) -> tuple[tuple, dict]:
return tuple(self.inputs.to_value_dict().values()), {}
@classmethod
def _build_inputs_preview(cls) -> dict[str, tuple[Any, Any]]:
return {f"row_{i}": (dict, NOT_DATA) for i in range(cls._length)}
@classfactory
def inputs_to_dataframe_factory(
n: int, use_cache: bool = True, /
) -> type[InputsToDataframe]:
return ( # type: ignore[return-value]
f"{InputsToDataframe.__name__}{n}",
(InputsToDataframe,),
{
"_length": n,
"use_cache": use_cache,
},
{},
)
def inputs_to_dataframe(n: int, use_cache: bool = True, *node_args, **node_kwargs):
"""
Creates and returns an instance of a dynamically generated
:class:`InputsToDataframe` subclass with a specified number of inputs, each being a
dictionary to form rows of the dataframe.
Args:
n (int): Number of input channels.
use_cache (bool): Whether this node should default to caching its values.
(Default is True.)
*node_args: Positional arguments for the node instance.
**node_kwargs: Keyword arguments for the node instance.
Returns:
InputsToDataframe: An instance of the dynamically created
:class:`InputsToDataframe` subclass.
"""
cls = inputs_to_dataframe_factory(n, use_cache)
cls.preview_io()
return cls(*node_args, **node_kwargs)
class DataclassNode(FromManyInputs, ABC):
"""
A base class for a node that converts inputs into a dataclass instance.
"""
dataclass: ClassVar[type] # Mandatory in children, must pass `is_dataclass`
_output_name: ClassVar[str] = "dataclass"
@classmethod
def _dataclass_fields(cls):
return cls.dataclass.__dataclass_fields__
def _setup_node(self) -> None:
super()._setup_node()
# Then leverage default factories from the dataclass
for name, channel in self.inputs.items():
if (
channel.value is NOT_DATA
and self._dataclass_fields()[name].default_factory is not MISSING
):
self.inputs[name] = self._dataclass_fields()[name].default_factory()
def _on_run(self, **inputs_to_value_dict):
return self.dataclass(**inputs_to_value_dict)
@property
def _run_args(self) -> tuple[tuple, dict]:
return (), self.inputs.to_value_dict()
@classmethod
def _build_inputs_preview(cls) -> dict[str, tuple[Any, Any]]:
# Make a channel for each field
return {
name: (f.type, NOT_DATA if f.default is MISSING else f.default)
for name, f in cls._dataclass_fields().items()
}
@classmethod
def _extra_info(cls) -> str:
return cls.dataclass.__doc__ or ""
@classfactory
def dataclass_node_factory(
dataclass: type, use_cache: bool = True, /
) -> type[DataclassNode]:
if not isinstance(dataclass, type):
raise TypeError(
f"{DataclassNode} expected to get a dataclass but {dataclass} is not "
f"type `type`."
)
dataclass = as_dataclass(dataclass)
# Classes inheriting from a dataclass will pass the `dataclasses.is_dataclass` test
# BUT they won't themselves _act_ as dataclass definitions! I.e. if you introduce
# new fields in a sub-dataclass, or update defaults, this won't register _unless_
# that new class is _also_ wrapped as a @dataclasses.dataclass!
# This is not our fault, it's just a python thing. But is our _problem_.
# To make sure dataclass nodes inheriting from other dataclass nodes still act as
# dataclasses, just cast everything as a dataclass -- re-casting an
# already-dataclass is not harmful
# Composition is preferable over inheritance, but we want inheritance to be possible
module, qualname = dataclass.__module__, dataclass.__qualname__
dataclass.__qualname__ += ".dataclass" # So output type hints know where to find it
return ( # type: ignore[return-value]
dataclass.__name__,
(DataclassNode,),
{
"dataclass": dataclass,
"__module__": module,
"__qualname__": qualname,
"_output_type_hint": dataclass,
"__doc__": dataclass.__doc__,
"use_cache": use_cache,
},
{},
)
def as_dataclass_node(dataclass: type):
"""
Decorates a dataclass as a dataclass node -- i.e. a node whose inputs correspond
to dataclass fields and whose output is an instance of the dataclass.
The underlying dataclass can be accessed on the :attr:`.dataclass` class attribute
of the resulting node class.
Leverages defaults (default factories) on dataclass fields to populate input
channel values at class defintion (instantiation).
Args:
dataclass (type): A dataclass, i.e. class passing `dataclasses.is_dataclass`,
or class definition that will be automatically wrapped with
`dataclasses.dataclass`.
use_cache (bool): Whether nodes of this type should default to caching their
values. (Default is True.)
Returns:
(type[DataclassNode]): A :class:`DataclassNode` subclass whose instances
transform inputs to an instance of that dataclass.
Examples:
>>> from dataclasses import dataclass, field
>>>
>>> from pyiron_workflow import Workflow
>>>
>>> def some_list():
... return [1, 2, 3]
>>>
>>> @Workflow.wrap.as_dataclass_node
... class Foo:
... necessary: str
... bar: str = "bar"
... answer: int = 42
... complex_: list = field(default_factory=some_list)
>>>
>>> f = Foo()
>>> print(f.readiness_report)
Foo readiness report:
ready: False
running: False
failed: False
inputs.necessary: False
inputs.bar: True
inputs.answer: True
inputs.complex_: True
<BLANKLINE>
>>> f(necessary="input as a node kwarg")
Foo.dataclass(necessary='input as a node kwarg', bar='bar', answer=42, complex_=[1, 2, 3])
"""
dataclass_node_factory.clear(dataclass.__name__) # Force a fresh class
module, qualname = dataclass.__module__, dataclass.__qualname__
cls = dataclass_node_factory(dataclass)
cls._reduce_imports_as = (module, qualname)
cls.preview_io()
return cls
def dataclass_node(dataclass: type, use_cache: bool = True, *node_args, **node_kwargs):
"""
Builds a dataclass node from a dataclass -- i.e. a node whose inputs correspond
to dataclass fields and whose output is an instance of the dataclass.
The underlying dataclass can be accessed on the :attr:`.dataclass` class attribute
of the resulting node.
Leverages defaults (default factories) on dataclass fields to populate input
channel values at class defintion (instantiation).
Args:
dataclass (type): A dataclass, i.e. class passing `dataclasses.is_dataclass`,
or class variable that will be automatically passed to
`dataclasses.dataclass`.
use_cache (bool): Whether this node should default to caching its values.
(Default is True.)
*node_args: Other :class:`Node` positional arguments.
**node_kwargs: Other :class:`Node` keyword arguments.
Returns:
(DataclassNode): An instance of the dynamically created :class:`DataclassNode`
subclass.
Examples:
>>> from dataclasses import dataclass, field
>>>
>>> from pyiron_workflow import Workflow
>>>
>>> def some_list():
... return [1, 2, 3]
>>>
>>> #@dataclass # Works on actual dataclasses as well as dataclass-like classes
>>> class Foo:
... necessary: str
... bar: str = "bar"
... answer: int = 42
... complex_: list = field(default_factory=some_list)
>>>
>>> f = Workflow.create.transformer.dataclass_node(Foo, label="my_dc")
>>> print(f.readiness_report)
my_dc readiness report:
ready: False
running: False
failed: False
inputs.necessary: False
inputs.bar: True
inputs.answer: True
inputs.complex_: True
<BLANKLINE>
>>> f(necessary="input as a node kwarg")
Foo.dataclass(necessary='input as a node kwarg', bar='bar', answer=42, complex_=[1, 2, 3])
"""
cls = dataclass_node_factory(dataclass)
cls.preview_io()
return cls(*node_args, **node_kwargs)