Add initial version of SloIE parsing script

dataset_infos.json

@@ -0,0 +1 @@
+{"default": {"description": "SloIE is a manually labelled dataset of Slovene idiomatic expressions. \nIt contains 29,400 sentences with 75 different expressions that can occur with either a literal or an idiomatic meaning, \nwith appropriate manual annotations for each token. The idiomatic expressions were selected from the Slovene Lexical \nDatabase (http://hdl.handle.net/11356/1030). Only expressions that can occur with both a literal and an idiomatic \nmeaning were selected. The sentences were extracted from the Gigafida corpus.\n", "citation": "@article{skvorc2022mice,\ntitle = {MICE: Mining Idioms with Contextual Embeddings},\njournal = {Knowledge-Based Systems},\nvolume = {235},\npages = {107606},\nyear = {2022},\nissn = {0950-7051},\ndoi = {https://doi.org/10.1016/j.knosys.2021.107606},\nurl = {https://www.sciencedirect.com/science/article/pii/S0950705121008686},\nauthor = {{\u000b S}kvorc, Tadej and Gantar, Polona and Robnik-{\u000b S}ikonja, Marko},\n}\n", "homepage": "http://hdl.handle.net/11356/1030", "license": "Creative Commons - Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0)", "features": {"sentence": {"dtype": "string", "id": null, "_type": "Value"}, "expression": {"dtype": "string", "id": null, "_type": "Value"}, "word_order": {"feature": {"dtype": "int32", "id": null, "_type": "Value"}, "length": -1, "id": null, "_type": "Sequence"}, "sentence_words": {"feature": {"dtype": "string", "id": null, "_type": "Value"}, "length": -1, "id": null, "_type": "Sequence"}, "is_idiom": {"feature": {"dtype": "string", "id": null, "_type": "Value"}, "length": -1, "id": null, "_type": "Sequence"}}, "post_processed": null, "supervised_keys": null, "task_templates": null, "builder_name": "sloie", "config_name": "default", "version": {"version_str": "1.0.0", "description": null, "major": 1, "minor": 0, "patch": 0}, "splits": {"train": {"name": "train", "num_bytes": 15801505, "num_examples": 29399, "dataset_name": "sloie"}}, "download_checksums": {"https://www.clarin.si/repository/xmlui/bitstream/handle/11356/1335/SloIE.zip": {"num_bytes": 4425132, "checksum": "26602f475742c4717c7bfaa32e6d66b98220dfb5c7dc010d1b1cc8e7eb9a33b1"}}, "download_size": 4425132, "post_processing_size": null, "dataset_size": 15801505, "size_in_bytes": 20226637}}

sloie.py

@@ -0,0 +1,102 @@
+""" SloIE is a manually labelled dataset of Slovene idiomatic expressions. """
+
+
+import os
+
+import datasets
+
+_CITATION = """\
+@article{skvorc2022mice,
+title = {MICE: Mining Idioms with Contextual Embeddings},
+journal = {Knowledge-Based Systems},
+volume = {235},
+pages = {107606},
+year = {2022},
+issn = {0950-7051},
+doi = {https://doi.org/10.1016/j.knosys.2021.107606},
+url = {https://www.sciencedirect.com/science/article/pii/S0950705121008686},
+author = {{\v S}kvorc, Tadej and Gantar, Polona and Robnik-{\v S}ikonja, Marko},
+}
+"""
+
+_DESCRIPTION = """\
+SloIE is a manually labelled dataset of Slovene idiomatic expressions. 
+It contains 29,400 sentences with 75 different expressions that can occur with either a literal or an idiomatic meaning, 
+with appropriate manual annotations for each token. The idiomatic expressions were selected from the Slovene Lexical 
+Database (http://hdl.handle.net/11356/1030). Only expressions that can occur with both a literal and an idiomatic 
+meaning were selected. The sentences were extracted from the Gigafida corpus.
+"""
+
+_HOMEPAGE = "http://hdl.handle.net/11356/1030"
+
+_LICENSE = "Creative Commons - Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0)"
+
+_URLS = {
+    "sloie": "https://www.clarin.si/repository/xmlui/bitstream/handle/11356/1335/SloIE.zip"
+}
+
+
+class SloIE(datasets.GeneratorBasedBuilder):
+    """ SloIE is a manually labelled dataset of Slovene idiomatic expressions. """
+
+    VERSION = datasets.Version("1.0.0")
+
+    def _info(self):
+        features = datasets.Features(
+            {
+                "sentence": datasets.Value("string"),
+                "expression": datasets.Value("string"),
+                "word_order": datasets.Sequence(datasets.Value("int32")),
+                "sentence_words": datasets.Sequence(datasets.Value("string")),
+                "is_idiom": datasets.Sequence(datasets.Value("string"))
+            }
+        )
+        return datasets.DatasetInfo(
+            description=_DESCRIPTION,
+            features=features,
+            homepage=_HOMEPAGE,
+            license=_LICENSE,
+            citation=_CITATION,
+        )
+
+    def _split_generators(self, dl_manager):
+        urls = _URLS["sloie"]
+        data_dir = dl_manager.download_and_extract(urls)
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs={"file_path": os.path.join(data_dir, "SloIE.txt")}
+            )
+        ]
+
+    def _generate_examples(self, file_path):
+        idx_instance = 0
+        with open(file_path, "r", encoding="utf-8") as f:
+            line = f.readline().strip()
+            while line:
+                assert line.startswith("#")
+                sent = line[1:]  # Remove initial "#"
+                word_order = list(map(int, f.readline().strip().split(" ")))
+                expression = ""
+                sentence_words, idiomaticity = [], []
+
+                line = f.readline().strip()
+                while line:
+                    token_info = line.split("\t")
+                    word, is_idiomatic_str, expression = token_info
+                    sentence_words.append(word)
+                    idiomaticity.append(is_idiomatic_str)
+
+                    line = f.readline().strip()
+                    # Encountered start of the next sentence - Note that "#" may also be an annotated word, hence the second condition
+                    if line.startswith("#") and len(line.split("\t")) == 1:
+                        break
+
+                yield idx_instance, {
+                    "sentence": sent,
+                    "expression": expression,
+                    "word_order": word_order,
+                    "sentence_words": sentence_words,
+                    "is_idiom": idiomaticity
+                }
+                idx_instance += 1