Fixed configs

ik-nlp-22_pestyle.py

@@ -2,7 +2,6 @@ import os
 
 import datasets
 import pandas as pd
-import numpy as np
 
 _CITATION = """No citation information available."""
 
@@ -17,11 +16,11 @@ _HOMEPAGE = "https://www.rug.nl/masters/information-science/?lang=en"
 
 _LICENSE = "Sharing and publishing of the data is not allowed at the moment."
 
-_SPLITS = {
-    "train": os.path.join("IK_NLP_22_PESTYLE", "train.tsv"),
-    "test_mask_subject": os.path.join("IK_NLP_22_PESTYLE", "test.tsv"),
-    "test_mask_modality": os.path.join("IK_NLP_22_PESTYLE", "test.tsv"),
-    "test_mask_time": os.path.join("IK_NLP_22_PESTYLE", "test.tsv")
+_PATHS = {
+    "full": os.path.join("IK_NLP_22_PESTYLE", "train.tsv"),
+    "mask_subject": os.path.join("IK_NLP_22_PESTYLE", "test.tsv"),
+    "mask_modality": os.path.join("IK_NLP_22_PESTYLE", "test.tsv"),
+    "mask_time": os.path.join("IK_NLP_22_PESTYLE", "test.tsv")
 }
 
 _ALL_FIELDS = [
@@ -33,12 +32,6 @@ _ALL_FIELDS = [
     "n_insert", "n_delete", "n_substitute", "n_shift", "bleu", "chrf", "ter", "aligned_edit"
 ]
 
-_INT_FIELDS = [
-    "k_total", "k_letter", "k_digit", "k_white", "k_symbol", "k_nav", "k_erase", 
-    "k_copy", "k_cut", "k_paste", "n_pause_geq_300", "len_pause_geq_300", 
-    "n_pause_geq_1000", "len_pause_geq_1000", "num_annotations"
-]
-
 _FIELDS_MASK_SUBJECT = [f for f in _ALL_FIELDS if f not in ["subject_id"]]
 _FIELDS_MASK_MODALITY = [f for f in _ALL_FIELDS if f not in [
     "modality", "mt_text", "n_insert", "n_delete", "n_substitute", 
@@ -50,14 +43,14 @@ _FIELDS_MASK_TIME = [f for f in _ALL_FIELDS if f not in [
 ]]
 
 _DICT_FIELDS = {
-    "train": _ALL_FIELDS,
-    "test_mask_subject": _FIELDS_MASK_SUBJECT,
-    "test_mask_modality": _FIELDS_MASK_MODALITY,
-    "test_mask_time": _FIELDS_MASK_TIME
+    "full": _ALL_FIELDS,
+    "mask_subject": _FIELDS_MASK_SUBJECT,
+    "mask_modality": _FIELDS_MASK_MODALITY,
+    "mask_time": _FIELDS_MASK_TIME
 }
 
 class IkNlp22PEStyleConfig(datasets.BuilderConfig):
-    """BuilderConfig for the IK NLP '22 HT-Style Dataset."""
+    """BuilderConfig for the IK NLP '22 Post-editing Stylometry Dataset."""
 
     def __init__(
         self,
@@ -79,12 +72,13 @@ class IkNlp22PEStyle(datasets.GeneratorBasedBuilder):
 
     BUILDER_CONFIGS = [
         IkNlp22PEStyleConfig(
-            name="main",
-            features=_ALL_FIELDS,
-        ),
+            name=name,
+            features=fields,
+        )
+        for name, fields in _DICT_FIELDS.items()
     ]
 
-    DEFAULT_CONFIG_NAME = "main"
+    DEFAULT_CONFIG_NAME = "full"
 
     @property
 
@@ -92,20 +86,18 @@ class IkNlp22PEStyle(datasets.GeneratorBasedBuilder):
         return (
             "The access to the data is restricted to students of the IK MSc NLP 2022 course working on a related project."
             "To load the data using this dataset, download and extract the IK_NLP_22_PESTYLE folder you were provided upon selecting the final project."
-            "After extracting it, the folder (referred to as root) must contain a IK_NLP_22_PESTYLE subfolder, containing train.csv and test.csv files."
-            "Then, load the dataset with: `datasets.load_dataset('GroNLP/ik-nlp-22_pestyle', 'main', data_dir='path/to/root/folder')`"
+            "After extracting it, the folder (referred to as root) must contain a IK_NLP_22_PESTYLE subfolder, containing train.tsv and test.tsv files."
+            f"Then, load the dataset with: `datasets.load_dataset('GroNLP/ik-nlp-22_pestyle', '{self.config.name}', data_dir='path/to/root/folder')`"
         )
 
     def _info(self):
-        features = {feature: datasets.Value("float32") for feature in self.config.features}
-        for feat in _INT_FIELDS:
-            features[feat] = datasets.Value("int32")
-        features["subject_id"] = datasets.Value("string")
-        features["modality"] = datasets.Value("string")
-        features["src_text"] = datasets.Value("string")
-        features["mt_text"] = datasets.Value("string")
-        features["tgt_text"] = datasets.Value("string")
-        features["aligned_edit"] = datasets.Value("string")
+        features = {feature: datasets.Value("int32") for feature in self.config.features}
+        for field in ["subject_id", "modality", "src_text", "mt_text", "tgt_text", "aligned_edit"]:
+            if field in self.config.features:
+                features[field] = datasets.Value("string")
+        for field in ["aligned_edit", "bleu", "chrf", "ter"]:
+            if field in self.config.features:
+                features[field] = datasets.Value("float32")
         return datasets.DatasetInfo(
             description=_DESCRIPTION,
             features=datasets.Features(features),
@@ -127,23 +119,17 @@ class IkNlp22PEStyle(datasets.GeneratorBasedBuilder):
             )
         return [
             datasets.SplitGenerator(
-                name=name,
+                name=datasets.Split.TRAIN if self.config.name == "full" else datasets.Split.TEST,
                 gen_kwargs={
-                    "filepath": os.path.join(data_dir, path),
-                    "name": name,
+                    "filepath": os.path.join(data_dir, _PATHS[self.config.name]),
+                    "features": self.config.features,
                 },
             )
-            for name, path in _SPLITS.items()
         ]
     
-    def _generate_examples(self, filepath: str, name):
+    def _generate_examples(self, filepath: str, features: list(str)):
         """Yields examples as (key, example) tuples."""
         data = pd.read_csv(filepath, sep="\t")
-        for f in set(_ALL_FIELDS).difference(set(_DICT_FIELDS[name])):
-            if isinstance(data[f][0], float):
-                data[f] = np.nan
-            else:
-                data[f] = None
-        print(data.shape)
+        data = data[features]
         for id_, row in data.iterrows():
             yield id_, row.to_dict()