start

README.md

@@ -0,0 +1,16 @@
+# WMT14 English-German Translation Data w/ further preprocessing
+
+The original pre-processing script is [here](https://github.com/pytorch/fairseq/blob/master/examples/translation/prepare-wmt14en2de.sh).
+
+This pre-processed dataset was created by running:
+
+```
+git clone https://github.com/pytorch/fairseq
+cd fairseq
+cd examples/translation/
+./prepare-wmt14en2de.sh
+```
+
+It was originally used by  `transformers` [`finetune_trainer.py`](https://github.com/huggingface/transformers/blob/641f418e102218c4bf16fcd3124bfebed6217ef6/examples/seq2seq/finetune_trainer.py)
+
+The data itself resides at https://cdn-datasets.huggingface.co/translation/wmt_en_de.tgz

process.txt

@@ -0,0 +1,21 @@
+
+# how this build script and dataset_infos.json were generated
+
+# this is translation, so let's adapt flores - it has an almost identical input format, just files are named differently:
+cp https://github.com/huggingface/datasets/blob/master/datasets/flores/flores.py wmt14-en-de-pre-processed.py
+perl -pi -e 's|Flores|wmt14-en-de-pre-processed|g' wmt14-en-de-pre-processed.py
+
+(good models for other tasks can be found here: https://huggingface.co/docs/datasets/add_dataset.html#dataset-scripts-of-reference)
+
+# then edit to change the language pairs, file template and the data url
+git add wmt14-en-de-pre-processed.py
+git commit -m "build script" wmt14-en-de-pre-processed.py
+git push
+
+# finally test
+datasets-cli test stas/wmt14-en-de-pre-processed --save_infos --all_configs
+
+# add push the generated config
+git add dataset_infos.json
+git commit -m "add dataset_infos.json" dataset_infos.json
+git push

wmt14-en-de-pre-processed.py

@@ -0,0 +1,133 @@
+# coding=utf-8
+# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+""" WMT16 English-Romanian Translation Data with further preprocessing """
+
+from __future__ import absolute_import, division, print_function
+
+import csv
+import json
+import os
+
+import datasets
+
+_CITATION = """\
+@InProceedings{huggingface:dataset,
+title = {WMT14 English-German Translation Data with further preprocessing},
+authors={},
+year={2016}
+}
+"""
+
+_DESCRIPTION = "WMT14 English-German Translation Data with further preprocessing"
+_HOMEPAGE = "http://www.statmt.org/wmt16/"
+_LICENSE = ""
+
+
+_DATA_URL = "https://cdn-datasets.huggingface.co/translation/wmt_en_de.tgz"
+
+
+class Wmt14EnDePreProcessedConfig(datasets.BuilderConfig):
+    """BuilderConfig for wmt16."""
+
+    def __init__(self, language_pair=(None, None), **kwargs):
+        """BuilderConfig for wmt16
+
+        Args:
+            for the `datasets.features.text.TextEncoder` used for the features feature.
+          language_pair: pair of languages that will be used for translation. Should
+            contain 2-letter coded strings. First will be used at source and second
+            as target in supervised mode. For example: ("se", "en").
+          **kwargs: keyword arguments forwarded to super.
+        """
+        name = "%s%s" % (language_pair[0], language_pair[1])
+
+        description = ("Translation dataset from %s to %s") % (language_pair[0], language_pair[1])
+        super(Wmt14EnDePreProcessedConfig, self).__init__(
+            name=name,
+            description=description,
+            version=datasets.Version("1.1.0", ""),
+            **kwargs,
+        )
+
+        # Validate language pair.
+        assert "en" in language_pair, ("Config language pair must contain `en`, got: %s", language_pair)
+        source, target = language_pair
+        non_en = source if target == "en" else target
+        assert non_en in ["de"], ("Invalid non-en language in pair: %s", non_en)
+
+        self.language_pair = language_pair
+
+
+# TODO: Name of the dataset usually match the script name with CamelCase instead of snake_case
+class Wmt14EnDePreProcessed(datasets.GeneratorBasedBuilder):
+
+    BUILDER_CONFIGS = [
+        Wmt14EnDePreProcessedConfig(
+            language_pair=("en", "ro"),
+        ),
+    ]
+
+    def _info(self):
+        source, target = self.config.language_pair
+        return datasets.DatasetInfo(
+            description=_DESCRIPTION,
+            features=datasets.Features(
+                {"translation": datasets.features.Translation(languages=self.config.language_pair)}
+            ),
+            supervised_keys=(source, target),
+            homepage=_HOMEPAGE,
+            citation=_CITATION,
+        )
+
+    def _split_generators(self, dl_manager):
+        dl_dir = dl_manager.download_and_extract(_DATA_URL)
+
+        source, target = self.config.language_pair
+        non_en = source if target == "en" else target
+        path_tmpl = "{dl_dir}/wmt_en_ro/{split}.{type}"
+
+        files = {}
+        for split in ("train", "val", "test"):
+            files[split] = {
+                "source_file": path_tmpl.format(dl_dir=dl_dir, split=split, type="source"),
+                "target_file": path_tmpl.format(dl_dir=dl_dir, split=split, type="target"),
+            }
+
+        return [
+            datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs=files["train"]),
+            datasets.SplitGenerator(name=datasets.Split.VALIDATION, gen_kwargs=files["val"]),
+            datasets.SplitGenerator(name=datasets.Split.TEST, gen_kwargs=files["test"]),
+        ]
+
+    def _generate_examples(self, source_file, target_file):
+        """This function returns the examples in the raw (text) form."""
+        with open(source_file, encoding="utf-8") as f:
+            source_sentences = f.read().split("\n")
+        with open(target_file, encoding="utf-8") as f:
+            target_sentences = f.read().split("\n")
+
+        assert len(target_sentences) == len(source_sentences), "Sizes do not match: %d vs %d for %s vs %s." % (
+            len(source_sentences),
+            len(target_sentences),
+            source_file,
+            target_file,
+        )
+
+        source, target = self.config.language_pair
+        for idx, (l1, l2) in enumerate(zip(source_sentences, target_sentences)):
+            result = {"translation": {source: l1, target: l2}}
+            # Make sure that both translations are non-empty.
+            if all(result.values()):
+                yield idx, result