Update sd-nlp-non-tokenized.py

Added disease and filtered data

sd-nlp-non-tokenized.py

@@ -42,6 +42,8 @@ class SourceDataNLP(datasets.GeneratorBasedBuilder):
         "I-ORGANISM",
         "B-EXP_ASSAY",
         "I-EXP_ASSAY",
+        "B-DISEASE",
+        "I-DISEASE",
     ]
     _SEMANTIC_GENEPROD_ROLES_LABEL_NAMES =  ["O", "B-CONTROLLED_VAR", "I-CONTROLLED_VAR", "B-MEASURED_VAR", "I-MEASURED_VAR"]
     _SEMANTIC_SMALL_MOL_ROLES_LABEL_NAMES = ["O", "B-CONTROLLED_VAR", "I-CONTROLLED_VAR", "B-MEASURED_VAR", "I-MEASURED_VAR"]
@@ -68,8 +70,8 @@ class SourceDataNLP(datasets.GeneratorBasedBuilder):
     VERSION = datasets.Version("1.0.0")
 
     _URLS = {
-        "NER": f"{_BASE_URL}sd_panels.zip",
-        "PANELIZATION": f"{_BASE_URL}sd_fig_general_tokenization.zip",
+        "NER": f"{_BASE_URL}sd_panels_filtered.zip",
+        "PANELIZATION": f"{_BASE_URL}sd_panelization_filtered.zip",
     }
     BUILDER_CONFIGS = [
         datasets.BuilderConfig(name="NER", version=VERSION, description="Dataset for entity recognition"),
@@ -95,6 +97,7 @@ class SourceDataNLP(datasets.GeneratorBasedBuilder):
                                                     names=self._NER_LABEL_NAMES)
                     ),
                     "tag_mask": datasets.Sequence(feature=datasets.Value("int8")),
+                    "panel_id": datasets.Sequence(feature=datasets.Value("string")),
                 }
             )
         elif self.config.name == "GENEPROD_ROLES":
@@ -108,6 +111,7 @@ class SourceDataNLP(datasets.GeneratorBasedBuilder):
                         )
                     ),
                     "tag_mask": datasets.Sequence(feature=datasets.Value("int8")),
+                    "panel_id": datasets.Sequence(feature=datasets.Value("string")),
                 }
             )
         elif self.config.name == "SMALL_MOL_ROLES":
@@ -121,6 +125,7 @@ class SourceDataNLP(datasets.GeneratorBasedBuilder):
                         )
                     ),
                     "tag_mask": datasets.Sequence(feature=datasets.Value("int8")),
+                    "panel_id": datasets.Sequence(feature=datasets.Value("string")),
                 }
             )
         elif self.config.name == "BORING":
@@ -131,6 +136,7 @@ class SourceDataNLP(datasets.GeneratorBasedBuilder):
                         feature=datasets.ClassLabel(num_classes=len(self._BORING_LABEL_NAMES),
                                                     names=self._BORING_LABEL_NAMES)
                     ),
+                    "panel_id": datasets.Sequence(feature=datasets.Value("string")),
                 }
             )
         elif self.config.name == "PANELIZATION":
@@ -142,6 +148,7 @@ class SourceDataNLP(datasets.GeneratorBasedBuilder):
                                                     names=self._PANEL_START_NAMES)
                     ),
                     "tag_mask": datasets.Sequence(feature=datasets.Value("int8")),
+                    "panel_id": datasets.Sequence(feature=datasets.Value("string")),
                 }
             )
 
@@ -203,7 +210,8 @@ class SourceDataNLP(datasets.GeneratorBasedBuilder):
                     yield id_, {
                         "words": data["words"],
                         "labels": labels,
-                        "tag_mask": tag_mask
+                        "tag_mask": tag_mask,
+                        "panel_id": data["panel_id"]
                     }
                 elif self.config.name == "GENEPROD_ROLES":
                     labels = data["label_ids"]["geneprod_roles"]
@@ -213,6 +221,7 @@ class SourceDataNLP(datasets.GeneratorBasedBuilder):
                         "words": data["words"],
                         "labels": labels,
                         "tag_mask": tag_mask,
+                        "panel_id": data["panel_id"]
                     }
                 elif self.config.name == "SMALL_MOL_ROLES":
                     labels = data["label_ids"]["small_mol_roles"]
@@ -222,6 +231,7 @@ class SourceDataNLP(datasets.GeneratorBasedBuilder):
                         "words": data["words"],
                         "labels": labels,
                         "tag_mask": tag_mask,
+                        "panel_id": data["panel_id"]
                     }
                 elif self.config.name == "BORING":
                     yield id_, {"words": data["words"],