feat: add examples and image cropping method

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text

app.py

@@ -1,3 +1,5 @@
+import os
+
 import spaces  # Must be before any CUDA/torch imports
 import gradio as gr
 
@@ -5,6 +7,14 @@ import abc_utils
 import config
 from inference import inference
 
+_EXAMPLE_EXTENSIONS = (".png", ".jpg", ".jpeg", ".webp")
+_examples_dir = os.path.join(config.APP_DIR, "examples")
+_example_paths = []
+if os.path.isdir(_examples_dir):
+    for name in sorted(os.listdir(_examples_dir)):
+        if name.lower().endswith(_EXAMPLE_EXTENSIONS):
+            _example_paths.append([os.path.join(_examples_dir, name)])
+
 with gr.Blocks(theme=gr.themes.Soft(), title="LEGATO OMR Demo") as demo:
     gr.Markdown("""
     # 🎼 LEGATO: Large-scale End-to-end Generalizable Approach to Typeset OMR
@@ -33,6 +43,12 @@ with gr.Blocks(theme=gr.themes.Soft(), title="LEGATO OMR Demo") as demo:
 
     gr.Markdown("### ✨ Try it")
     inp = gr.Image(type="pil", label="📤 Upload score image")
+    if _example_paths:
+        gr.Examples(
+            examples=_example_paths,
+            inputs=[inp],
+            label="Example scores",
+        )
     with gr.Row():
         out = gr.Textbox(label="📝 ABC transcription", lines=10, buttons=["copy"])
         with gr.Accordion("🎵 Rendered ABC notation", open=True):

image_utils.py

@@ -5,12 +5,12 @@ from config import LETTER_ASPECT
 
 def pad_to_portrait_letter(pil_image: Image.Image) -> Image.Image:
     """If aspect ratio is wider than letter, pad at the bottom to match letter aspect."""
-    w, h = pil_image.size
-    if w / h < LETTER_ASPECT:
-        return pil_image
-    new_h = int(round(w / LETTER_ASPECT))
-    canvas = Image.new("RGB", (w, new_h), (255, 255, 255))
     if pil_image.mode != "RGB":
         pil_image = pil_image.convert("RGB")
+    w, h = pil_image.size
+    pil_image = pil_image.resize((1050, 1050*h//w))
+    if pil_image.height >= 1485:
+        return pil_image
+    canvas = Image.new("RGB", (1050, 1485), (255, 255, 255))
     canvas.paste(pil_image, (0, 0))
     return canvas