Use iter_archive when streaming

the_pile_stack_exchange.py

@@ -13,8 +13,9 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 """The Stack Exchange Corpus"""
-
+import io
 import os
+import zipfile
 from pathlib import Path
 
 import datasets
@@ -58,24 +59,58 @@ class ThePileStackExchange(datasets.GeneratorBasedBuilder):
         )
 
     def _split_generators(self, dl_manager):
-        dl_dir = dl_manager.download_and_extract(_URL)
-        zips = [str(f) for f in (Path(dl_dir) / "out").iterdir()]
-        extracted = dl_manager.extract(zips, num_proc=os.cpu_count())
-        # non-dir extracteds are zero-size unknown things
-        dirs = [path for path in extracted if os.path.isdir(path)]
-        return [
-            datasets.SplitGenerator(name=datasets.Split.TRAIN, gen_kwargs={"dirs": dirs}),
-        ]
+        if dl_manager.is_streaming:
+            archive = dl_manager.download(_URL)
+            return [
+                datasets.SplitGenerator(
+                    name=datasets.Split.TRAIN,
+                    gen_kwargs={
+                        "files": dl_manager.iter_archive(archive),
+                        "is_streaming": dl_manager.is_streaming,
+                    },
+                ),
+            ]
+        else:
+            dl_dir = dl_manager.download_and_extract(_URL)
+            zips = [str(f) for f in (Path(dl_dir) / "out").iterdir()]
+            extracted = dl_manager.extract(zips, num_proc=os.cpu_count())
+            # non-dir extracteds are zero-size unknown things
+            dirs = [path for path in extracted if os.path.isdir(path)]
+            return [
+                datasets.SplitGenerator(
+                    name=datasets.Split.TRAIN,
+                    gen_kwargs={
+                        "files": dl_manager.iter_files(dirs),
+                        "is_streaming": dl_manager.is_streaming,
+                    },
+                ),
+            ]
 
-    def _generate_examples(self, dirs):
+    def _generate_examples(self, files, is_streaming):
         """Yields examples."""
-        _id = 0
-        for dir in sorted(dirs):
-            txt_files = sorted(Path(dir).glob("**/*.txt"))
-            for txt_file in txt_files:
-                # PosiPath(/home/user/.cache/huggingface/datasets/downloads/extracted/3923d60abeeb876021dc55a897ac2f260b181556f8ca56a7c61e3b8b80afec77/academia.stackexchange_0000000001.txt)
-                domain = txt_file.name.split(".")[0]
-                with txt_file.open(mode="r", encoding="utf-8") as f:
+        if is_streaming:
+            _id = 0
+            for path, file in files:
+                if not path.startswith("out/"):
+                    continue
+                file_content = file.read()
+                with zipfile.ZipFile(io.BytesIO(file_content)) as zip_file:
+                    for name in zip_file.namelist():
+                        if not name.endswith(".txt"):
+                            continue
+                        domain = name.split(".")[0]
+                        with zip_file.open(name, mode="r") as f:
+                            document = f.read().decode(encoding="utf-8")
+                        yield _id, {"domain": domain, "text": document}
+                        _id += 1
+        else:
+            _id = 0
+            for file in files:
+                path = Path(file)
+                if not path.name.endswith(".txt"):
+                    continue
+                domain = path.name.split(".")[0]
+                with path.open(mode="r", encoding="utf-8") as f:
                     document = f.read()
                 yield _id, {"domain": domain, "text": document}
                 _id += 1