File size: 6,133 Bytes
ad029a3
 
 
 
 
 
 
 
 
 
 
 
e957cc1
 
 
ad029a3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e957cc1
ad029a3
 
 
 
 
 
 
e957cc1
ad029a3
 
 
e957cc1
ad029a3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
# /home/perk/mymodel/categorisation-mt5x/tasks.py

import functools
import seqio
import tensorflow_datasets as tfds
from t5.evaluation import metrics
import my_metrics
from t5.data import preprocessors
import t5
import tensorflow.compat.v1 as tf

tsv_path = {
        "train": "gs://north-t5x/corpus/deuncaser/norwegian/train.tsv",
        "validation": "gs://north-t5x/corpus/deuncaser/norwegian/validation.tsv",
        "test": "gs://north-t5x/corpus/deuncaser/norwegian/validation.tsv"
}

vocabulary = seqio.SentencePieceVocabulary(
                'gs://t5-data/vocabs/mc4.250000.100extra/sentencepiece.model', extra_ids=0)

DEFAULT_OUTPUT_FEATURES = {
    "inputs":
        seqio.Feature(
            vocabulary=vocabulary, add_eos=True),
    "targets":
        seqio.Feature(
            vocabulary=vocabulary, add_eos=True)
}

def categorise_preprocessor(ds):
  def normalize_text(text):
    """Lowercase and remove quotes from a TensorFlow string."""
    text = tf.strings.regex_replace(text,"'(.*)'", r"\1")
    return text

  def to_inputs_and_targets(ex):
    """Map {"source": ..., "source": ...}->{"target": ..., "target": ...}."""
    return {
        "inputs":
             tf.strings.join(
                 [normalize_text(ex["source"])]),
        "targets": 
	    tf.strings.join(
                 [normalize_text(ex["target"])]),
    }

  return ds.map(to_inputs_and_targets, 
                num_parallel_calls=tf.data.experimental.AUTOTUNE)


def categorise_fulltext_preprocessor(ds):
  def normalize_text(text):
    """Lowercase and remove quotes from a TensorFlow string."""
    text = tf.strings.regex_replace(text,"'(.*)'", r"\1")
    return text

  def fulltext(t):
    if t=="0":
        t="il testo è favorevole alla vaccinazione"
    elif t=="1":
        t="il testo è neutro rispetto alla vaccinazione"
    elif t=="2":
        t="is testo è sfavorevole alla vaccinazione"
    return t

  def to_inputs_and_targets(ex):
    """Map {"source": ..., "source": ...}->{"target": ..., "target": ...}."""
    return {
        "inputs":
             tf.strings.join(
                 [normalize_text(ex["source"])]),
        "targets": 
	    tf.strings.join(
                 [fulltext(normalize_text(ex["target"]))]),
    }
  return ds.map(to_inputs_and_targets, 
                num_parallel_calls=tf.data.experimental.AUTOTUNE)


def categorise_fulltext_word_preprocessor(ds):
  def normalize_text(text):
    """Lowercase and remove quotes from a TensorFlow string."""
    text = tf.strings.regex_replace(text,"'(.*)'", r"\1")
    return text

  def fulltext(t):
    if t=="0":
        t="promozionale"
    elif t=="1":
        t="neutro"
    elif t=="2":
        t="scoraggiante"
    return t

  def to_inputs_and_targets(ex):
    """Map {"source": ..., "source": ...}->{"target": ..., "target": ...}."""
    return {
        "inputs":
             tf.strings.join(
                 [normalize_text(ex["source"])]),
        "targets": 
	    tf.strings.join(
                 [fulltext(normalize_text(ex["target"]))]),
    }
  return ds.map(to_inputs_and_targets, 
                num_parallel_calls=tf.data.experimental.AUTOTUNE)



def categorise_binary_preprocessor(ds):
  def normalize_text(text):
    """Lowercase and remove quotes from a TensorFlow string."""
    text = tf.strings.regex_replace(text,"'(.*)'", r"\1")
    return text

  def fulltext(t):
    if t=="0":
        t="1"
    elif t=="1":
        t="1"
    elif t=="2":
        t="2"
    return t

  def to_inputs_and_targets(ex):
    """Map {"source": ..., "source": ...}->{"target": ..., "target": ...}."""
    return {
        "inputs":
             tf.strings.join(
                 [normalize_text(ex["source"])]),
        "targets": 
	    tf.strings.join(
                 [fulltext(normalize_text(ex["target"]))]),
    }
  return ds.map(to_inputs_and_targets, 
                num_parallel_calls=tf.data.experimental.AUTOTUNE)



seqio.TaskRegistry.add(
    "deuncaser",
    source=seqio.TextLineDataSource(
        split_to_filepattern=tsv_path,
        #num_input_examples=num_nq_examples
        ),
    preprocessors=[
      functools.partial(
          t5.data.preprocessors.parse_tsv,
          field_names=["id","methods","source","target"]),
      categorise_preprocessor,
      seqio.preprocessors.tokenize_and_append_eos,
    ],
    metric_fns=[metrics.accuracy,metrics.bleu],
    output_features=DEFAULT_OUTPUT_FEATURES,
)

seqio.TaskRegistry.add(
    "classify_tweets_fulltext",
    source=seqio.TextLineDataSource(
        split_to_filepattern=tsv_path,
        #num_input_examples=num_nq_examples
        ),
    preprocessors=[
      functools.partial(
          t5.data.preprocessors.parse_tsv,
          field_names=["annotator1","annotator2","annotator3","target","source","id"]),
      categorise_fulltext_preprocessor,
      seqio.preprocessors.tokenize_and_append_eos,
    ],
    metric_fns=[metrics.accuracy,my_metrics.f1_macro],
    output_features=DEFAULT_OUTPUT_FEATURES,
)

seqio.TaskRegistry.add(
    "classify_tweets_binary",
    source=seqio.TextLineDataSource(
        split_to_filepattern=tsv_path,
        #num_input_examples=num_nq_examples
        ),
    preprocessors=[
      functools.partial(
          t5.data.preprocessors.parse_tsv,
          field_names=["annotator1","annotator2","annotator3","target","source","id"]),
      categorise_binary_preprocessor,
      seqio.preprocessors.tokenize_and_append_eos,
    ],
    metric_fns=[metrics.accuracy,my_metrics.f1_macro],
    output_features=DEFAULT_OUTPUT_FEATURES,
)

seqio.TaskRegistry.add(
    "classify_tweets_fulltext_word",
    source=seqio.TextLineDataSource(
        split_to_filepattern=tsv_path,
        #num_input_examples=num_nq_examples
        ),
    preprocessors=[
      functools.partial(
          t5.data.preprocessors.parse_tsv,
          field_names=["annotator1","annotator2","annotator3","target","source","id"]),
      categorise_fulltext_word_preprocessor,
      seqio.preprocessors.tokenize_and_append_eos,
    ],
    metric_fns=[metrics.accuracy,my_metrics.f1_macro],
    output_features=DEFAULT_OUTPUT_FEATURES,
)