add jamp

llm-jp-eval.py

@@ -6,6 +6,7 @@ from typing import Generator
 
 import datasets
 
+from .base import BaseDatasetProcessor
 from .chabsa import ChabsaDatasetProcessor
 from .coreference import WikipediaCoreferenceDatasetProcessor
 from .dependency import WikipediaDependencyDatasetProcessor
@@ -129,7 +130,7 @@ class LlmJpEval(datasets.GeneratorBasedBuilder):
         elif dataset_name == "jsts":
             processor = JSTSDatasetProcessor
 
-        output_dir = "./data_llm-jp-eval"
+        output_dir = Path("./data_llm-jp-eval")
         processor(output_dir, version_name).pipeline()
 
         eval_dir = output_dir / version_name / "evaluation"