update: annotation app

src/streamlit_app.py

@@ -1,7 +1,8 @@
+import io
 import streamlit as st
 import fitz
 from PIL import Image
-import io
+from datasets import Dataset
 from streamlit_cropper import st_cropper
 
 
@@ -70,3 +71,14 @@ if uploaded_pdf:
         print(f"{len(st.session_state.cropped_images)=}")
         print(f"{st.session_state.cropped_images[-1].size=}")
         print(f"{st.session_state.captions[-1]=}")
+
+huggingface_dataset_address = st.sidebar.text_input("Hugging Face Dataset Address")
+if st.sidebar.button("Save to HuggingFace"):
+    dataset = Dataset.from_dict(
+        {
+            "image": st.session_state.cropped_images,
+            "caption": st.session_state.captions,
+        },
+    )
+    dataset.push_to_hub(huggingface_dataset_address)
+    st.success("Dataset saved to Hugging Face")