change split name and file reading

paradocs.py

@@ -111,7 +111,7 @@ class ParaDocs(datasets.GeneratorBasedBuilder):
 
         return [
             datasets.SplitGenerator(
-                name="shard",
+                name="train",
                 gen_kwargs={
                     "filepaths": dl_manager.download(data_sources[lang])
                 }
@@ -139,7 +139,7 @@ class ParaDocs(datasets.GeneratorBasedBuilder):
         for filepath in filepaths:
             # logger.info("Generating examples from = %s", filepath)
             try:
-                with open(filepath, "r", encoding="utf-8") as f:
+                with xz.open(filepath, "r", encoding="utf-8") as f:
                     rstream = csv.DictReader(f,
                                                 delimiter='\t',
                                                 fieldnames = [
@@ -166,5 +166,7 @@ class ParaDocs(datasets.GeneratorBasedBuilder):
                     for example in rstream:
                         yield id_, example
                         id_ += 1
-            except:
+            except Exception as e:
+                print(e, filepath)
                 print("Error reading file:", filepath)
+