Create dataset.py

dataset.py

@@ -0,0 +1,54 @@
+import json
+import datasets
+
+_DESCRIPTION = """
+LEMAS Eval Dataset (Multilingual)
+
+Each sample contains:
+- audio: speech audio
+- txt: normalized transcript
+- align: word-level alignment (json string)
+"""
+
+class LEMASEval(datasets.GeneratorBasedBuilder):
+    VERSION = datasets.Version("1.0.0")
+
+    def _info(self):
+        return datasets.DatasetInfo(
+            description=_DESCRIPTION,
+            features=datasets.Features(
+                {
+                    "key": datasets.Value("string"),
+                    "audio": datasets.Audio(sampling_rate=None),
+                    "dur": datasets.Value("float32"),
+                    "txt": datasets.Value("string"),
+                    "align": datasets.Value("string"),
+                }
+            ),
+            supervised_keys=None,
+        )
+
+    def _split_generators(self, dl_manager):
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TEST,
+                gen_kwargs={
+                    "jsonl_path": "eval.jsonl",
+                },
+            )
+        ]
+
+    def _generate_examples(self, jsonl_path):
+        with open(jsonl_path, "r", encoding="utf-8") as f:
+            for idx, line in enumerate(f):
+                obj = json.loads(line)
+
+                yield idx, {
+                    "key": obj["key"],
+                    "audio": obj["audio"], 
+                    "dur": float(obj["dur"]),
+                    "txt": obj["txt"],
+                    # 把 struct → string
+                    "align": json.dumps(obj["align"], ensure_ascii=False),
+                }
+