ocr: pin datasets>=4.0.0 in the 9 recipes that left it unpinned (#70)

An unpinned "datasets" dep lets the resolver backtrack to pre-2.15
releases that still use pa.PyExtensionType, which modern pyarrow has
removed — the job then crashes at import, before inference (see
davanstrien/ocr-bench#18 for a report against dots-ocr.py). The other
17 recipes already pin datasets>=4.0.0; this aligns the remaining 9.
Resolution verified clean for all nine with uv lock --script.


Claude-Session: https://claude.ai/code/session_01FaJ84bsoBpgVXocRzSJ6N5

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>

Author: davanstrien

ocr/deepseek-ocr.py

@@ -1,7 +1,7 @@
 # /// script
 # requires-python = ">=3.11"
 # dependencies = [
-#     "datasets",
+#     "datasets>=4.0.0",
 #     "huggingface-hub",
 #     "pillow",
 #     "torch",

ocr/dots-mocr.py

@@ -1,7 +1,7 @@
 # /// script
 # requires-python = ">=3.11"
 # dependencies = [
-#     "datasets",
+#     "datasets>=4.0.0",
 #     "huggingface-hub",
 #     "pillow",
 #     "vllm>=0.15.1",

ocr/dots-ocr.py

@@ -1,7 +1,7 @@
 # /// script
 # requires-python = ">=3.11"
 # dependencies = [
-#     "datasets",
+#     "datasets>=4.0.0",
 #     "huggingface-hub",
 #     "pillow",
 #     "vllm>=0.15.1",

ocr/falcon-ocr.py

@@ -1,7 +1,7 @@
 # /// script
 # requires-python = ">=3.11"
 # dependencies = [
-#     "datasets",
+#     "datasets>=4.0.0",
 #     "huggingface-hub",
 #     "pillow",
 #     "torch>=2.5",

ocr/firered-ocr.py

@@ -1,7 +1,7 @@
 # /// script
 # requires-python = ">=3.11"
 # dependencies = [
-#     "datasets",
+#     "datasets>=4.0.0",
 #     "huggingface-hub",
 #     "pillow",
 #     "vllm>=0.15.1",

ocr/lighton-ocr.py

@@ -1,7 +1,7 @@
 # /// script
 # requires-python = ">=3.11"
 # dependencies = [
-#     "datasets",
+#     "datasets>=4.0.0",
 #     "huggingface-hub",
 #     "pillow",
 #     "vllm",

ocr/olmocr2-vllm.py

@@ -1,7 +1,7 @@
 # /// script
 # requires-python = ">=3.11"
 # dependencies = [
-#     "datasets",
+#     "datasets>=4.0.0",
 #     "huggingface-hub",
 #     "pillow",
 #     "vllm",

ocr/paddleocr-vl-1.5.py

@@ -1,7 +1,7 @@
 # /// script
 # requires-python = ">=3.11"
 # dependencies = [
-#     "datasets",
+#     "datasets>=4.0.0",
 #     "huggingface-hub",
 #     "pillow",
 #     "torch",

ocr/smoldocling-ocr.py

@@ -1,7 +1,7 @@
 # /// script
 # requires-python = ">=3.11"
 # dependencies = [
-#     "datasets",
+#     "datasets>=4.0.0",
 #     "huggingface-hub",
 #     "pillow",
 #     "vllm",