Remove deprecated functions, and warnings - Text (#773)

* Remove deprecated functions, and warnings
* Update links for docstring
* chlog

Co-authored-by: Daniel Stancl <46073029+stancld@users.noreply.github.com>
Co-authored-by: Jirka Borovec <Borda@users.noreply.github.com>

CHANGELOG.md

@@ -20,6 +20,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Removed
 
+- Removed deprecated functions, and warnings in Text ([#773](https://github.com/PyTorchLightning/metrics/pull/773))
+  * `functional.wer`
+  * `WER`
+
 
 ### Fixed
 
@@ -58,8 +62,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 - Renamed IoU -> Jaccard Index ([#662](https://github.com/PyTorchLightning/metrics/pull/662))
 - Renamed text WER metric ([#714](https://github.com/PyTorchLightning/metrics/pull/714))
- * `functional.wer` -> `functional.word_error_rate`
- * `WER` -> `WordErrorRate`
+  * `functional.wer` -> `functional.word_error_rate`
+  * `WER` -> `WordErrorRate`
 - Renamed correlation coefficient classes: ([#710](https://github.com/PyTorchLightning/metrics/pull/710))
   * `MatthewsCorrcoef` -> `MatthewsCorrCoef`
   * `PearsonCorrcoef` -> `PearsonCorrCoef`
@@ -81,27 +85,27 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   * `SNR` -> `SignalNoiseRatio`
   * `SI_SNR` -> `ScaleInvariantSignalNoiseRatio`
 - Renamed F-score metrics: ([#731](https://github.com/PyTorchLightning/metrics/pull/731), [#740](https://github.com/PyTorchLightning/metrics/pull/740))
-  *  `functional.f1` ->  `functional.f1_score`
-  *  `F1` ->  `F1Score`
-  *  `functional.fbeta` ->  `functional.fbeta_score`
-  *  `FBeta` ->  `FBetaScore`
+  * `functional.f1` ->  `functional.f1_score`
+  * `F1` ->  `F1Score`
+  * `functional.fbeta` ->  `functional.fbeta_score`
+  * `FBeta` ->  `FBetaScore`
 - Renamed Hinge metric: ([#734](https://github.com/PyTorchLightning/metrics/pull/734))
-  *  `functional.hinge` ->  `functional.hinge_loss`
-  *  `Hinge` ->  `HingeLoss`
+  * `functional.hinge` ->  `functional.hinge_loss`
+  * `Hinge` ->  `HingeLoss`
 - Renamed image PSNR metrics ([#732](https://github.com/PyTorchLightning/metrics/pull/732))
   * `functional.psnr` -> `functional.peak_signal_noise_ratio`
   * `PSNR` -> `PeakSignalNoiseRatio`
 - Renamed image PIT metric: ([#737](https://github.com/PyTorchLightning/metrics/pull/737))
-  *  `functional.pit` ->  `functional.permutation_invariant_training`
-  *  `PIT` ->  `PermutationInvariantTraining`
+  * `functional.pit` ->  `functional.permutation_invariant_training`
+  * `PIT` ->  `PermutationInvariantTraining`
 - Renamed image SSIM metric: ([#747](https://github.com/PyTorchLightning/metrics/pull/747))
-  *  `functional.ssim` ->  `functional.scale_invariant_signal_noise_ratio`
-  *  `SSIM` ->  `StructuralSimilarityIndexMeasure`
+  * `functional.ssim` ->  `functional.scale_invariant_signal_noise_ratio`
+  * `SSIM` ->  `StructuralSimilarityIndexMeasure`
 - Renamed detection `MAP` to `MeanAveragePrecision` metric ([#754](https://github.com/PyTorchLightning/metrics/pull/754))
 - Renamed Fidelity & LPIPS image metric: ([#752](https://github.com/PyTorchLightning/metrics/pull/752))
-  *  `image.FID` ->  `image.FrechetInceptionDistance`
-  *  `image.KID` ->  `image.KernelInceptionDistance`
-  *  `image.LPIPS` ->  `image.LearnedPerceptualImagePatchSimilarity`
+  * `image.FID` ->  `image.FrechetInceptionDistance`
+  * `image.KID` ->  `image.KernelInceptionDistance`
+  * `image.LPIPS` ->  `image.LearnedPerceptualImagePatchSimilarity`
 
 ### Removed
 

docs/source/links.rst

@@ -28,7 +28,7 @@
 .. _sklearn averaging methods: https://scikit-learn.org/stable/modules/model_evaluation.html#multiclass-and-multilabel-classification
 .. _Cosine Similarity: https://en.wikipedia.org/wiki/Cosine_similarity
 .. _spearmans rank correlation coefficient: https://en.wikipedia.org/wiki/Spearman%27s_rank_correlation_coefficient
-.. _WER: https://en.wikipedia.org/wiki/Word_error_rate
+.. _WordErrorRate: https://en.wikipedia.org/wiki/Word_error_rate
 .. _FID: https://en.wikipedia.org/wiki/Fr%C3%A9chet_inception_distance
 .. _mean-squared-error: https://en.wikipedia.org/wiki/Mean_squared_error
 .. _SSIM: https://en.wikipedia.org/wiki/Structural_similarity

docs/source/references/functional.rst

@@ -507,10 +507,10 @@ translation_edit_rate [func]
 .. autofunction:: torchmetrics.functional.translation_edit_rate
     :noindex:
 
-wer [func]
-~~~~~~~~~~
+word_error_rate [func]
+~~~~~~~~~~~~~~~~~~~~~~
 
-.. autofunction:: torchmetrics.functional.wer
+.. autofunction:: torchmetrics.functional.word_error_rate
     :noindex:
 
 word_information_lost [func]

docs/source/references/modules.rst

@@ -678,10 +678,10 @@ TranslationEditRate
 .. autoclass:: torchmetrics.TranslationEditRate
     :noindex:
 
-WER
-~~~
+WordErrorRate
+~~~~~~~~~~~~~
 
-.. autoclass:: torchmetrics.WER
+.. autoclass:: torchmetrics.WordErrorRate
     :noindex:
 
 WordInfoLost

torchmetrics/__init__.py

@@ -90,7 +90,6 @@
     RetrievalRPrecision,
 )
 from torchmetrics.text import (  # noqa: E402
-    WER,
     BLEUScore,
     CharErrorRate,
     CHRFScore,
@@ -187,7 +186,6 @@
     "SumMetric",
     "SymmetricMeanAbsolutePercentageError",
     "TranslationEditRate",
-    "WER",
     "WordErrorRate",
     "CharErrorRate",
     "MatchErrorRate",

torchmetrics/functional/__init__.py

@@ -76,7 +76,7 @@
 from torchmetrics.functional.text.sacre_bleu import sacre_bleu_score
 from torchmetrics.functional.text.squad import squad
 from torchmetrics.functional.text.ter import translation_edit_rate
-from torchmetrics.functional.text.wer import wer, word_error_rate
+from torchmetrics.functional.text.wer import word_error_rate
 from torchmetrics.functional.text.wil import word_information_lost
 from torchmetrics.functional.text.wip import word_information_preserved
 from torchmetrics.utilities.imports import _TRANSFORMERS_AUTO_AVAILABLE
@@ -158,7 +158,6 @@
     "stat_scores",
     "symmetric_mean_absolute_percentage_error",
     "translation_edit_rate",
-    "wer",
     "word_error_rate",
     "char_error_rate",
     "match_error_rate",

torchmetrics/functional/text/__init__.py

@@ -20,7 +20,7 @@
 from torchmetrics.functional.text.sacre_bleu import sacre_bleu_score  # noqa: F401
 from torchmetrics.functional.text.squad import squad  # noqa: F401
 from torchmetrics.functional.text.ter import translation_edit_rate  # noqa: F401
-from torchmetrics.functional.text.wer import wer, word_error_rate  # noqa: F401
+from torchmetrics.functional.text.wer import word_error_rate  # noqa: F401
 from torchmetrics.functional.text.wil import word_information_lost  # noqa: F401
 from torchmetrics.functional.text.wip import word_information_preserved  # noqa: F401
 from torchmetrics.utilities.imports import _NLTK_AVAILABLE, _TRANSFORMERS_AUTO_AVAILABLE

torchmetrics/functional/text/bert.py

@@ -19,11 +19,9 @@
 from warnings import warn
 
 import torch
-from deprecate import deprecated
 from torch import Tensor
 from torch.utils.data import DataLoader, Dataset
 
-from torchmetrics.utilities import _future_warning
 from torchmetrics.utilities.imports import _TQDM_AVAILABLE, _TRANSFORMERS_AUTO_AVAILABLE
 
 if _TRANSFORMERS_AUTO_AVAILABLE:
@@ -457,13 +455,6 @@ def _rescale_metrics_with_baseline(
     return all_metrics[..., 0], all_metrics[..., 1], all_metrics[..., 2]
 
 
-@deprecated(
-    args_mapping={"predictions": "preds", "references": "target"},
-    target=True,
-    deprecated_in="0.7",
-    remove_in="0.8",
-    stream=_future_warning,
-)
 def bert_score(
     preds: Union[List[str], Dict[str, Tensor]],
     target: Union[List[str], Dict[str, Tensor]],
@@ -549,13 +540,6 @@ def bert_score(
     Returns:
         Python dictionary containing the keys `precision`, `recall` and `f1` with corresponding values.
 
-    .. deprecated:: v0.7
-        Args:
-            predictions:
-                This argument is deprecated in favor of  `preds` and will be removed in v0.8.
-            references:
-                This argument is deprecated in favor of  `target` and will be removed in v0.8.
-
     Raises:
         ValueError:
             If `len(preds) != len(target)`.

torchmetrics/functional/text/bleu.py

@@ -18,14 +18,10 @@
 # Link: https://pytorch.org/text/_modules/torchtext/data/metrics.html#bleu_score
 from collections import Counter
 from typing import Callable, Sequence, Tuple, Union
-from warnings import warn
 
 import torch
-from deprecate import deprecated
 from torch import Tensor, tensor
 
-from torchmetrics.utilities import _future_warning
-
 
 def _count_ngram(ngram_input_list: Sequence[str], n_gram: int) -> Counter:
     """Counting how many times each word appears in a given text with ngram.
@@ -146,13 +142,6 @@ def _bleu_score_compute(
     return bleu
 
 
-@deprecated(
-    args_mapping={"translate_corpus": "preds", "reference_corpus": "target"},
-    target=True,
-    deprecated_in="0.7",
-    remove_in="0.8",
-    stream=_future_warning,
-)
 def bleu_score(
     preds: Union[str, Sequence[str]],
     target: Sequence[Union[str, Sequence[str]]],
@@ -174,13 +163,6 @@ def bleu_score(
     Return:
         Tensor with BLEU Score
 
-    .. deprecated:: v0.7
-        Args:
-            translate_corpus:
-                This argument is deprecated in favor of  `preds` and will be removed in v0.8.
-            reference_corpus:
-                This argument is deprecated in favor of  `target` and will be removed in v0.8.
-
     Example:
         >>> from torchmetrics.functional import bleu_score
         >>> preds = ['the cat is on the mat']
@@ -195,11 +177,6 @@ def bleu_score(
         [2] Automatic Evaluation of Machine Translation Quality Using Longest Common Subsequence
         and Skip-Bigram Statistics by Chin-Yew Lin and Franz Josef Och `Machine Translation Evolution`_
     """
-    warn(
-        "Input order of targets and preds were changed to predictions firsts and targets second in v0.7."
-        " Warning will be removed in v0.8."
-    )
-
     preds_ = [preds] if isinstance(preds, str) else preds
     target_ = [[tgt] if isinstance(tgt, str) else tgt for tgt in target]
 

torchmetrics/functional/text/cer.py

@@ -15,11 +15,9 @@
 from typing import List, Tuple, Union
 
 import torch
-from deprecate import deprecated
 from torch import Tensor, tensor
 
 from torchmetrics.functional.text.helper import _edit_distance
-from torchmetrics.utilities import _future_warning
 
 
 def _cer_update(
@@ -61,30 +59,17 @@ def _cer_compute(errors: Tensor, total: Tensor) -> Tensor:
     return errors / total
 
 
-@deprecated(
-    args_mapping={"predictions": "preds", "references": "target"},
-    target=True,
-    deprecated_in="0.7",
-    remove_in="0.8",
-    stream=_future_warning,
-)
 def char_error_rate(preds: Union[str, List[str]], target: Union[str, List[str]]) -> Tensor:
     """character error rate is a common metric of the performance of an automatic speech recognition system. This
     value indicates the percentage of characters that were incorrectly predicted. The lower the value, the better the
     performance of the ASR system with a CER of 0 being a perfect score.
     Args:
         preds: Transcription(s) to score as a string or list of strings
         target: Reference(s) for each speech input as a string or list of strings
+
     Returns:
         Character error rate score
 
-    .. deprecated:: v0.7
-        Args:
-            predictions:
-                This argument is deprecated in favor of  `preds` and will be removed in v0.8.
-            references:
-                This argument is deprecated in favor of  `target` and will be removed in v0.8.
-
     Examples:
         >>> preds = ["this is the prediction", "there is an other sample"]
         >>> target = ["this is the reference", "there is another one"]

torchmetrics/functional/text/sacre_bleu.py

@@ -40,15 +40,12 @@
 import re
 from functools import partial
 from typing import Sequence
-from warnings import warn
 
 import torch
-from deprecate import deprecated
 from torch import Tensor, tensor
 from typing_extensions import Literal
 
 from torchmetrics.functional.text.bleu import _bleu_score_compute, _bleu_score_update
-from torchmetrics.utilities import _future_warning
 from torchmetrics.utilities.imports import _REGEX_AVAILABLE
 
 AVAILABLE_TOKENIZERS = ("none", "13a", "zh", "intl", "char")
@@ -278,13 +275,6 @@ def _lower(line: str, lowercase: bool) -> str:
         return line
 
 
-@deprecated(
-    args_mapping={"translate_corpus": "preds", "reference_corpus": "target"},
-    target=True,
-    deprecated_in="0.7",
-    remove_in="0.8",
-    stream=_future_warning,
-)
 def sacre_bleu_score(
     preds: Sequence[str],
     target: Sequence[Sequence[str]],
@@ -314,13 +304,6 @@ def sacre_bleu_score(
     Return:
         Tensor with BLEU Score
 
-    .. deprecated:: v0.7
-        Args:
-            translate_corpus:
-                This argument is deprecated in favor of  `preds` and will be removed in v0.8.
-            reference_corpus:
-                This argument is deprecated in favor of  `target` and will be removed in v0.8.
-
     Example:
         >>> from torchmetrics.functional import sacre_bleu_score
         >>> preds = ['the cat is on the mat']
@@ -337,10 +320,6 @@ def sacre_bleu_score(
         [3] Automatic Evaluation of Machine Translation Quality Using Longest Common Subsequence
         and Skip-Bigram Statistics by Chin-Yew Lin and Franz Josef Och `Machine Translation Evolution`_
     """
-    warn(
-        "Input order of targets and preds were changed to predictions firsts and targets second in v0.7."
-        " Warning will be removed in v0.8."
-    )
 
     if tokenize not in AVAILABLE_TOKENIZERS:
         raise ValueError(f"Argument `tokenize` expected to be one of {AVAILABLE_TOKENIZERS} but got {tokenize}.")