Upload laion2b_multi_korean_subset_with_image.py

laion2b_multi_korean_subset_with_image.py

@@ -0,0 +1,81 @@
+import json
+from collections import defaultdict
+
+from datasets import (
+    DatasetInfo,
+    GeneratorBasedBuilder,
+    Features,
+    SplitGenerator,
+    Value,
+    Image,
+    Split,
+    DownloadManager,
+)
+
+_URLS = [f"data/{i:0>5}.tar" for i in range(2123)]
+
+
+class Laion2bMultiKoreanSubsetWithImage(GeneratorBasedBuilder):
+
+    VERSION = "0.1.0"
+
+    def _info(self):
+        return DatasetInfo(
+            description="laion2b_multi_korean_subset data with images downloaded successfully",
+            features=Features(
+                {
+                    "image": Image(),
+                    "text": Value("string"),
+                    "width": Value("int32"),
+                    "height": Value("int32"),
+                }
+            ),
+            supervised_keys=None,
+            license="cc-by-4.0",
+        )
+
+    def _split_generators(self, dl_manager: DownloadManager):
+        downloaded = dl_manager.download(_URLS)
+        return [
+            SplitGenerator(
+                name=Split.TRAIN,
+                gen_kwargs={
+                    "archives": [
+                        dl_manager.iter_archive(archive) for archive in downloaded
+                    ]
+                },
+            )
+        ]
+
+    def _generate_examples(self, archives):
+        idx = 0
+        temp = defaultdict(dict)
+
+        for archive in archives:
+            for path, file in archive:
+                if path.endswith(".txt"):
+                    continue
+
+                file_id = path.split(".")[0]
+
+                if path.endswith(".json"):
+                    meta = json.loads(file.read())
+
+                    temp[file_id]["text"] = meta["caption"]
+                    temp[file_id]["width"] = meta["width"]
+                    temp[file_id]["height"] = meta["height"]
+
+                    if "image" in temp[file_id]:
+                        yield idx, temp[file_id]
+                        idx += 1
+                        del temp[file_id]
+
+                if path.endswith(".webp"):
+                    temp[file_id]["image"] = {"path": path, "bytes": file.read()}
+
+                    if "text" in temp[file_id]:
+                        yield idx, temp[file_id]
+                        idx += 1
+                        del temp[file_id]
+
+            temp.clear()