Support streaming

data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:405c9bf2a8fb08ba7d26d34e7aa32d53830da0eec1e6a9c82b2aa7f19b11104a
+size 3185028

hate_speech18.py

@@ -47,7 +47,7 @@ forums posts have been sampled from several subforums and split into sentences.
 have been manually labelled as containing hate speech or not, according to certain annotation guidelines.
 """
 
-_DATA_URL = "https://github.com/Vicomtech/hate-speech-dataset/archive/master.zip"
+_DATA_URL = "data.zip"
 
 
 class HateSpeech18(datasets.GeneratorBasedBuilder):
@@ -75,35 +75,21 @@ class HateSpeech18(datasets.GeneratorBasedBuilder):
 
         return [
             datasets.SplitGenerator(
-                name=datasets.Split.TRAIN, gen_kwargs={"filepath": os.path.join(dl_dir, "hate-speech-dataset-master")}
+                name=datasets.Split.TRAIN, gen_kwargs={"data_dir": os.path.join(dl_dir, "data")}
             ),
         ]
 
-    def _generate_examples(self, filepath):
-
-        with open(os.path.join(filepath, "annotations_metadata.csv"), encoding="utf-8") as csv_file:
-
-            csv_reader = csv.reader(
+    def _generate_examples(self, data_dir):
+        all_files_path = os.path.join(data_dir, "all_files")
+        with open(os.path.join(data_dir, "annotations_metadata.csv"), encoding="utf-8") as csv_file:
+            csv_reader = csv.DictReader(
                 csv_file, quotechar='"', delimiter=",", quoting=csv.QUOTE_ALL, skipinitialspace=True
             )
-
-            next(csv_reader)
-
             for idx, row in enumerate(csv_reader):
-
-                file_id, user_id, subforum_id, num_contexts, label = row
-
-                all_files_path = os.path.join(filepath, "all_files")
-
-                path = os.path.join(all_files_path, file_id + ".txt")
-
-                with open(path, encoding="utf-8") as file:
-                    text = file.read()
-
+                text_path = os.path.join(all_files_path, row.pop("file_id") + ".txt")
+                with open(text_path, encoding="utf-8") as text_file:
+                    text = text_file.read()
                 yield idx, {
                     "text": text,
-                    "user_id": user_id,
-                    "subforum_id": subforum_id,
-                    "num_contexts": num_contexts,
-                    "label": label,
+                    **row,
                 }