update on dataset

dataset_infos.json

@@ -0,0 +1,40 @@
+{
+    "default": {
+        "description": "LibriS2S: A German-English Speech-to-Speech Translation Corpus",
+        "citation": "@inproceedings{jeuris-niehues-2022-libris2s,\n   title = \"{L}ibri{S}2{S}: A {G}erman-{E}nglish Speech-to-Speech Translation Corpus\",\n   author = \"Jeuris, Pedro and Niehues, Jan\",\n   booktitle = \"Proceedings of the Thirteenth Language Resources and Evaluation Conference\",\n   year = \"2022\",\n   url = \"https://aclanthology.org/2022.lrec-1.98\",\n   pages = \"928--935\"\n}",
+        "homepage": "https://huggingface.co/datasets/PedroDKE/LibriS2S",
+        "license": "cc-by-nc-sa-4.0",
+        "features": {
+            "book_id": {
+                "dtype": "int64",
+                "_type": "Value"
+            },
+            "DE_audio": {
+                "_type": "Audio"
+            },
+            "EN_audio": {
+                "_type": "Audio"
+            },
+            "score": {
+                "dtype": "float32",
+                "_type": "Value"
+            },
+            "DE_transcript": {
+                "dtype": "string",
+                "_type": "Value"
+            },
+            "EN_transcript": {
+                "dtype": "string",
+                "_type": "Value"
+            }
+        },
+        "splits": {
+            "train": {
+                "name": "train",
+                "num_bytes": 10069085,
+                "num_examples": 25604,
+                "dataset_name": "LibriS2S"
+            }
+        }
+    }
+} 

libris2s/libris2s.py → libris2s.py

@@ -3,9 +3,21 @@ import os
 from datasets import DatasetInfo, GeneratorBasedBuilder, SplitGenerator, Split, Features, Value, Audio
 
 
-class Libris2s(GeneratorBasedBuilder):
+class LibriS2SDataset(GeneratorBasedBuilder):
+    """LibriS2S: A German-English Speech-to-Speech Translation Corpus"""
+
+    VERSION = "1.0.0"
+    BUILDER_CONFIGS = [
+        {
+            "name": "default",
+            "version": VERSION,
+            "description": "Default configuration for LibriS2S dataset",
+        }
+    ]
+
     def _info(self):
         return DatasetInfo(
+            description="LibriS2S: A German-English Speech-to-Speech Translation Corpus",
             features=Features({
                 "book_id": Value("int64"),
                 "DE_audio": Audio(),
@@ -14,41 +26,40 @@ class Libris2s(GeneratorBasedBuilder):
                 "DE_transcript": Value("string"),
                 "EN_transcript": Value("string"),
             }),
+            homepage="https://huggingface.co/datasets/PedroDKE/LibriS2S",
+            citation="""@inproceedings{jeuris-niehues-2022-libris2s,
+               title = "{L}ibri{S}2{S}: A {G}erman-{E}nglish Speech-to-Speech Translation Corpus",
+               author = "Jeuris, Pedro and Niehues, Jan",
+               booktitle = "Proceedings of the Thirteenth Language Resources and Evaluation Conference",
+               year = "2022",
+               url = "https://aclanthology.org/2022.lrec-1.98",
+               pages = "928--935"
+            }""",
+            license="cc-by-nc-sa-4.0",
         )
     
     def _split_generators(self, dl_manager):
         """Returns SplitGenerators."""
-        if self.config.data_dir is None:
-            raise ValueError(
-                "This dataset requires you to specify the data_dir parameter in the config. "
-                "You can do this by passing --data_dir when loading the dataset."
-            )
-            
         return [
             SplitGenerator(
                 name=Split.TRAIN,
                 gen_kwargs={
-                    "filepath": os.path.join(self.config.data_dir, "alignments/all_de_en_alligned_cleaned.csv"),
-                    "audio_dir": self.config.data_dir,
+                    "split": "train",
+                    "cache_dir": dl_manager.manual_dir if dl_manager.manual_dir else ".",
                 }
             ),
         ]
 
-    def _generate_examples(self, filepath, audio_dir):
+    def _generate_examples(self, split, cache_dir):
         """Yields examples."""
-        with open(filepath, encoding="utf-8") as f:
+        csv_path = os.path.join(cache_dir, "alignments/all_de_en_alligned_cleaned.csv")
+        
+        with open(csv_path, encoding="utf-8") as f:
             reader = csv.DictReader(f)
             for idx, row in enumerate(reader):
                 # Handle paths for both DE and EN audio
-                de_audio_path = os.path.join(audio_dir, row["DE_audio"])
-                en_audio_path = os.path.join(audio_dir, row["EN_audio"])
-                
-                # Check if files exist
-                if not os.path.exists(de_audio_path) or not os.path.exists(en_audio_path):
-                    print(f"Warning: Missing audio file(s) for index {idx}")
-                    print(f"DE audio path: {de_audio_path}")
-                    print(f"EN audio path: {en_audio_path}")
-                    continue
+                de_audio_path = os.path.join(cache_dir, row["DE_audio"])
+                en_audio_path = os.path.join(cache_dir, row["EN_audio"])
                 
                 yield idx, {
                     "book_id": int(row["book_id"]),
@@ -57,4 +68,4 @@ class Libris2s(GeneratorBasedBuilder):
                     "score": float(row["score"]),
                     "DE_transcript": row["DE_transcript"],
                     "EN_transcript": row["EN_transcript"],
-                }
+                }