adjusted preview

README.md

@@ -185,7 +185,6 @@ from datasets import load_dataset
 
 dataset = load_dataset(
     "ReadyAi/5000-podcast-conversations-with-metadata-and-embedding-dataset",
-    data_files="data/bittensor-conversational-tags-and-embeddings-part-*.parquet",
     split="train",
     streaming=True
 )

dataset.py

@@ -1,57 +1,95 @@
+import glob
+
 import pandas as pd
-from datasets import DatasetInfo, GeneratorBasedBuilder, Split, SplitGenerator, Value, Features, Sequence
+from datasets import DatasetInfo, Features, GeneratorBasedBuilder, Sequence, Split, SplitGenerator, Value
+
 
 class PodcastConversationsWithMetadataAndEmbedding(GeneratorBasedBuilder):
     def _info(self):
         return DatasetInfo(
-            features=Features({
-                "c_guid": Value("string"),
-                "participants": Sequence(Value("string")),
-                "transcript": Sequence({
-                    "chunk": Value("string"),
-                    "speaker": Value("string"),
-                    "text": Value("string")
-                })
-            })
+            features=Features(
+                {
+                    "c_guid": Value("string"),
+                    "participants": Sequence(Value("string")),
+                    "transcript": Sequence({"chunk": Value("string"), "speaker": Value("string"), "text": Value("string")}),
+                }
+            )
         )
 
-    def _split_generators(self, dl_manager):
-        data_files = self.config.data_files
+    def _info(self):
+        split_name = self.config.name
 
+        if split_name == ["train", "conversations"]:
+            features = Features(
+                {
+                    "c_guid": Value("string"),
+                    "transcript": Sequence({"chunk": Value("string"), "speaker": Value("string"), "text": Value("string")}),
+                    "participants": Sequence(Value("string")),
+                }
+            )
+        elif split_name == "full":
+            features = Features(
+                {
+                    "c_guid": Value("string"),
+                    "tag_id": Value("int64"),
+                    "tag": Value("string"),
+                    "vector": Sequence(Value("float32")),
+                }
+            )
+        elif split_name == "conversations_to_tags":
+            features = Features(
+                {
+                    "c_guid": Value("string"),
+                    "tag_id": Value("int64"),
+                    "tag": Value("string"),
+                }
+            )
+        elif split_name == "tag_to_id":
+            features = Features(
+                {
+                    "tag_id": Value("int64"),
+                    "tag": Value("string"),
+                }
+            )
+        else:
+            raise ValueError(f"Unknown split: {split_name}")
+
+        return DatasetInfo(features=features)
+
+    def _split_generators(self, dl_manager):
         return [
             SplitGenerator(
                 name=Split.TRAIN,
-                gen_kwargs={"filepath": data_files["train"][0]},
+                gen_kwargs={"filepaths": ["conversations_train.parquet"]},
             ),
             SplitGenerator(
                 name="conversations",
-                gen_kwargs={"filepath": data_files["conversations"][0]},
+                gen_kwargs={"filepaths": ["conversations.parquet"]},
             ),
             SplitGenerator(
                 name="conversations_to_tags",
-                gen_kwargs={"filepath": data_files["conversations_to_tags"][0]},
+                gen_kwargs={"filepaths": ["conversations_to_tags.parquet"]},
             ),
             SplitGenerator(
                 name="tag_to_id",
-                gen_kwargs={"filepath": data_files["tag_to_id"][0]},
+                gen_kwargs={"filepaths": ["tag_to_id.parquet"]},
             ),
             SplitGenerator(
                 name="full",
-                gen_kwargs={"filepath": data_files["full"][0]},
+                gen_kwargs={"filepaths": sorted(glob.glob("data/bittensor-conversational-tags-and-embeddings-part-*.parquet"))},
             ),
         ]
 
-    def _generate_examples(self, filepath):
-        df = pd.read_parquet(filepath)
-
-        for idx, row in df.iterrows():
-            record = row.to_dict()
+    def _generate_examples(self, filepaths):
+        for path in filepaths:
+            df = pd.read_parquet(path)
+            for idx, row in df.iterrows():
+                record = row.to_dict()
 
-            # Normalize vector field if it exists
-            if "vector" in record:
-                if isinstance(record["vector"], (list, tuple)):
-                    record["vector"] = list(map(float, record["vector"]))
-                else:
-                    record["vector"] = []
+                if "vector" in record:
+                    if isinstance(record["vector"], (list, tuple)):
+                        record["vector"] = list(map(float, record["vector"]))
+                    else:
+                        record["vector"] = []
 
-            yield idx, record
+                yield f"{path}-{idx}", record