search

tensorflow / addons

Useful extra functionality for TensorFlow 2.x maintained by SIG-addons
Apache License 2.0
1.69k stars 611 forks source link

tfa.text.parse_time fails with has no attribute 'addons_parse_time' #2805

Open sshcherbakov opened 1 year ago

sshcherbakov commented 1 year ago

System information

Describe the bug

The following code:

import tensorflow as tf
import tensorflow_addons as tfa
import numpy as np
from importlib import reload

file_path = 'input.csv'
dataset = tf.data.TextLineDataset(file_path).skip(1)

def parse_csv_line(line):
  fields = tf.io.decode_csv(line, record_defaults=[[""]] * 4)
  timestamp   = tfa.text.parse_time(fields[0], '%Y-%m-%dT%H:%M:%SZ', 'MILLISECOND')
  f1        = tf.strings.to_number(fields[1], tf.float32)  
  return {"timestamp": timestamp, "f1": f1}

dataset = dataset.map(parse_csv_line).batch(5)

Sample input.csv:

2013-09-11T00:33:54Z,99.9
2013-09-16T21:36:36Z,90.0
2013-09-17T15:09:14Z,95.0
2013-09-24T10:20:49Z,90.25
2013-10-01T19:59:40Z,96.0
2013-10-02T02:03:07Z,96.0
2013-10-02T13:41:21Z,93.5

always fails with 

WARNING:tensorflow:AutoGraph could not transform <function parse_csv_line at 0x7fe12f5deb90> and will run it as-is.
Cause: Unable to locate the source code of <function parse_csv_line at 0x7fe12f5deb90>. Note that functions defined in certain environments, like the interactive Python shell, do not expose their source code. If that is the case, you should define them in a .py source file. If you are certain the code is graph-compatible, wrap the call using @tf.autograph.experimental.do_not_convert. Original error: could not get source code
To silence this warning, decorate the function with @tf.autograph.experimental.do_not_convert
WARNING: AutoGraph could not transform <function parse_csv_line at 0x7fe12f5deb90> and will run it as-is.
Cause: Unable to locate the source code of <function parse_csv_line at 0x7fe12f5deb90>. Note that functions defined in certain environments, like the interactive Python shell, do not expose their source code. If that is the case, you should define them in a .py source file. If you are certain the code is graph-compatible, wrap the call using @tf.autograph.experimental.do_not_convert. Original error: could not get source code
To silence this warning, decorate the function with @tf.autograph.experimental.do_not_convert
Traceback (most recent call last):
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/autograph/pyct/parser.py", line 148, in parse_entity
    original_source = inspect_utils.getimmediatesource(entity)
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/autograph/pyct/inspect_utils.py", line 124, in getimmediatesource
    lines, lnum = inspect.findsource(obj)
  File "/Users/sergey/miniconda3/lib/python3.10/inspect.py", line 958, in findsource
    raise OSError('could not get source code')
OSError: could not get source code

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/autograph/impl/api.py", line 427, in converted_call
    converted_f = _convert_actual(target_entity, program_ctx)
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/autograph/impl/api.py", line 269, in _convert_actual
    transformed, module, source_map = _TRANSPILER.transform(entity, program_ctx)
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/autograph/pyct/transpiler.py", line 282, in transform
    return self.transform_function(obj, user_context)
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/autograph/pyct/transpiler.py", line 466, in transform_function
    nodes, ctx = super(PyToPy, self).transform_function(fn, user_context)
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/autograph/pyct/transpiler.py", line 342, in transform_function
    node, source = parser.parse_entity(fn, future_features=future_features)
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/autograph/pyct/parser.py", line 150, in parse_entity
    raise errors.InaccessibleSourceCodeError(
tensorflow.python.autograph.pyct.errors.InaccessibleSourceCodeError: Unable to locate the source code of <function parse_csv_line at 0x7fe12f5deb90>. Note that functions defined in certain environments, like the interactive Python shell, do not expose their source code. If that is the case, you should define them in a .py source file. If you are certain the code is graph-compatible, wrap the call using @tf.autograph.experimental.do_not_convert. Original error: could not get source code

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/data/ops/dataset_ops.py", line 2294, in map
    return MapDataset(self, map_func, preserve_cardinality=True, name=name)
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/data/ops/dataset_ops.py", line 5499, in __init__
    self._map_func = structured_function.StructuredFunctionWrapper(
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/data/ops/structured_function.py", line 263, in __init__
    self._function = fn_factory()
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/eager/polymorphic_function/tracing_compiler.py", line 226, in get_concrete_function
    concrete_function = self._get_concrete_function_garbage_collected(
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/eager/polymorphic_function/tracing_compiler.py", line 192, in _get_concrete_function_garbage_collected
    concrete_function, _ = self._maybe_define_concrete_function(args, kwargs)
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/eager/polymorphic_function/tracing_compiler.py", line 157, in _maybe_define_concrete_function
    return self._maybe_define_function(args, kwargs)
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/eager/polymorphic_function/tracing_compiler.py", line 360, in _maybe_define_function
    concrete_function = self._create_concrete_function(args, kwargs)
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/eager/polymorphic_function/tracing_compiler.py", line 284, in _create_concrete_function
    func_graph_module.func_graph_from_py_func(
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/framework/func_graph.py", line 1283, in func_graph_from_py_func
    func_outputs = python_func(*func_args, **func_kwargs)
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/data/ops/structured_function.py", line 240, in wrapped_fn
    ret = wrapper_helper(*args)
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/data/ops/structured_function.py", line 171, in wrapper_helper
    ret = autograph.tf_convert(self._func, ag_ctx)(*nested_args)
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/autograph/impl/api.py", line 689, in wrapper
    return converted_call(f, args, kwargs, options=options)
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/autograph/impl/api.py", line 434, in converted_call
    return _fall_back_unconverted(f, args, kwargs, options, e)
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/autograph/impl/api.py", line 484, in _fall_back_unconverted
    return _call_unconverted(f, args, kwargs, options)
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow/python/autograph/impl/api.py", line 458, in _call_unconverted
    return f(*args, **kwargs)
  File "<stdin>", line 5, in parse_csv_line
  File "/Users/sergey/miniconda3/lib/python3.10/site-packages/tensorflow_addons/text/parse_time_op.py", line 86, in parse_time
    return _parse_time_so.ops.addons_parse_time(time_string, time_format, output_unit)
AttributeError: module '012ff3e36e3c24aefc4a3a7b68a03fedd1e7a7e1' has no attribute 'addons_parse_time'

I've tried multiple combinations on multiple OSes and environments, including IDE, command line, interactive and non-interactive execution. The result is always the same.

The combination of the Python, TensorFlow and TensorFlow Addons seems to be fully supported.