Support streaming (#4)

- Host data file (4696d9fc2f7f678cd27880718b9f102cefe74496)
- Update and refactor code (46790a05ffbb5a4277937ec95f08970c396398b7)

data.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5430c7563887db6067e75bcfbad99dfa83b02b38834dbf5cfd355919d84554ec
+size 3627968

sofc_materials_articles.py

@@ -45,7 +45,7 @@ _HOMEPAGE = "https://arxiv.org/abs/2006.03039"
 
 _LICENSE = ""
 
-_URL = "https://github.com/boschresearch/sofc-exp_textmining_resources/archive/master.zip"
+_URL = "data.zip"
 
 
 class SOFCMaterialsArticles(datasets.GeneratorBasedBuilder):
@@ -232,33 +232,14 @@ class SOFCMaterialsArticles(datasets.GeneratorBasedBuilder):
 
     def _split_generators(self, dl_manager):
         """Returns SplitGenerators."""
-
-        my_urls = _URL
-        data_dir = dl_manager.download_and_extract(my_urls)
-
-        data_dir = os.path.join(data_dir, "sofc-exp_textmining_resources-master/sofc-exp-corpus")
-
-        metadata = pd.read_csv(os.path.join(data_dir, "SOFC-Exp-Metadata.csv"), sep="\t")
-
-        text_base_path = os.path.join(data_dir, "texts")
-
-        text_files_available = [
-            os.path.split(i.rstrip(".txt"))[-1] for i in glob.glob(os.path.join(text_base_path, "*.txt"))
-        ]
-
-        metadata = metadata[metadata["name"].map(lambda x: x in text_files_available)]
-
-        names = {}
-        splits = ["train", "test", "dev"]
-        for split in splits:
-            names[split] = metadata[metadata["set"] == split]["name"].tolist()
+        data_dir = dl_manager.download_and_extract(_URL)
+        data_dir = os.path.join(data_dir, "sofc-exp-corpus")
 
         return [
             datasets.SplitGenerator(
                 name=datasets.Split.TRAIN,
                 # These kwargs will be passed to _generate_examples
                 gen_kwargs={
-                    "names": names["train"],
                     "data_dir": data_dir,
                     "split": "train",
                 },
@@ -266,21 +247,26 @@ class SOFCMaterialsArticles(datasets.GeneratorBasedBuilder):
             datasets.SplitGenerator(
                 name=datasets.Split.TEST,
                 # These kwargs will be passed to _generate_examples
-                gen_kwargs={"names": names["test"], "data_dir": data_dir, "split": "test"},
+                gen_kwargs={
+                    "data_dir": data_dir,
+                    "split": "test",
+                },
             ),
             datasets.SplitGenerator(
                 name=datasets.Split.VALIDATION,
                 # These kwargs will be passed to _generate_examples
                 gen_kwargs={
-                    "names": names["dev"],
                     "data_dir": data_dir,
-                    "split": "validation",
+                    "split": "dev",
                 },
             ),
         ]
 
-    def _generate_examples(self, names, data_dir, split):
+    def _generate_examples(self, data_dir, split):
         """Yields examples."""
+        metadata = pd.read_csv(os.path.join(data_dir, "SOFC-Exp-Metadata.csv"), sep="\t")
+        names = metadata[metadata["set"] == split]["name"].tolist()
+
         # The dataset consists of the original article text as well as annotations
         textfile_base_path = os.path.join(data_dir, "texts")
         annotations_base_path = os.path.join(data_dir, "annotations")
@@ -308,7 +294,6 @@ class SOFCMaterialsArticles(datasets.GeneratorBasedBuilder):
         # For each text file, we'll load all of the
         # associated annotation files
         for id_, name in enumerate(sorted(names)):
-
             # Load the main source text
             textfile_path = os.path.join(textfile_base_path, name + ".txt")
             text = open(textfile_path, encoding="utf-8").read()
@@ -383,7 +368,6 @@ class SOFCMaterialsArticles(datasets.GeneratorBasedBuilder):
             # Iterate through the spans data
             spans = []
             for span in spans_raw:
-
                 # Split out the elements in each tab-delimited line
                 _, span_id, entity_label_or_exp, sentence_id, begin_char_offset, end_char_offset = span.split("\t")