update per-language splits

wiki_lingua.py

@@ -17,7 +17,6 @@
 import os
 import glob
 import pickle
-import re
 import datasets
 
 
@@ -170,7 +169,7 @@ class WikiLingua(datasets.GeneratorBasedBuilder):
                     "target_language": datasets.Value("string"),
                     "source": datasets.Value("string"),
                     "target": datasets.Value("string"),
-                    "references": [datasets.Value("string")]
+                    "references": [datasets.Value("string")],
                 }
             ),
             supervised_keys=None,
@@ -181,59 +180,65 @@ class WikiLingua(datasets.GeneratorBasedBuilder):
 
     def _split_generators(self, dl_manager):
         """Returns SplitGenerators."""
+
         dl_dir = dl_manager.download_and_extract(_URL)
         data_dir = os.path.join(dl_dir, "GEM_V2")
 
         return [
             datasets.SplitGenerator(
                 name=datasets.Split.TRAIN,
-                gen_kwargs={"filepath": data_dir, "split": "train"},
+                gen_kwargs={
+                    "filepaths": glob.glob(
+                        os.path.join(data_dir, f"wikilingua_*.train.pk")
+                    )
+                },
             ),
             datasets.SplitGenerator(
                 name=datasets.Split.VALIDATION,
-                gen_kwargs={"filepath": data_dir, "split": "validation"},
+                gen_kwargs={
+                    "filepaths": glob.glob(
+                        os.path.join(data_dir, f"wikilingua_*lingual.val.pk")
+                    )
+                },
             ),
-        ] + [
             datasets.SplitGenerator(
-                name=re.search("wikilingua_(.+test)", fn).group(1).replace(".", "_"),
-                gen_kwargs={"filepath": fn, "split": "test"},
-            )
-            for fn in glob.glob(os.path.join(data_dir, "wikilingua_*.test.pk"))
+                name=datasets.Split.TEST,
+                gen_kwargs={
+                    "filepaths": glob.glob(
+                        os.path.join(data_dir, f"wikilingua_*lingual.test.pk")
+                    )
+                },
+            ),
+            datasets.SplitGenerator(
+                name=f"sampled_{datasets.Split.VALIDATION}",
+                gen_kwargs={
+                    "filepaths": glob.glob(
+                        os.path.join(data_dir, f"wikilingua_*_sampled.val.pk")
+                    )
+                },
+            ),
+            datasets.SplitGenerator(
+                name=f"sampled_{datasets.Split.TEST}",
+                gen_kwargs={
+                    "filepaths": glob.glob(
+                        os.path.join(data_dir, f"wikilingua_*_sampled.test.pk")
+                    )
+                },
+            ),
         ]
 
-    def _generate_examples(self, filepath, split):
+    def _generate_examples(self, filepaths):
         """Yields examples."""
-
-        if split == "test":
-            with open(filepath, "rb") as f:
-                data = pickle.load(f)
-                for d in data:
-                    idx = d["id"].replace(".", "-")
-                    yield idx, {
-                        "gem_id": idx,
-                        "gem_parent_id": idx,
-                        "source_language": d["source"],
-                        "target_language": d["target"],
-                        "source": d["document"].strip(),
-                        "target": d["summary"].strip(),
-                        "references": [d["summary"].strip()],
-                    }
-        else:
-            # filter data as needed for train & validation sets
-            if split == "validation":
-                filepaths = glob.glob(os.path.join(filepath, "wikilingua_*.val.pk"))
+        for filepath in filepaths:
+            if (
+                self.config.name == WikilinguaModes.MULTILINGUAL
+                and WikilinguaModes.CROSSLINGUAL in filepath
+            ) or (
+                self.config.name == WikilinguaModes.CROSSLINGUAL
+                and WikilinguaModes.MULTILINGUAL in filepath
+            ):
+                yield from []
             else:
-                filepaths = glob.glob(os.path.join(filepath, "wikilingua_*.train.pk"))
-            for filepath in filepaths:
-                # skip files if they are irrelevant to task mode
-                if (
-                    self.config.mode == WikilinguaModes.MULTILINGUAL
-                    and "crosslingual" in filepath
-                ) or (
-                    self.config.mode == WikilinguaModes.CROSSLINGUAL
-                    and "multilingual" in filepath
-                ):
-                    yield from []
 
                 with open(filepath, "rb") as f:
                     data = pickle.load(f)
@@ -265,5 +270,5 @@ class WikiLingua(datasets.GeneratorBasedBuilder):
                             "target_language": tgt_lang,
                             "source": src,
                             "target": tgt,
-                            "references": [tgt]
+                            "references": [tgt],
                         }