Update app.py from Colab

app.py

@@ -1,54 +1,47 @@
 #this was in part generated with gemini llm
-import os  # For reading environment variables
-import shutil  # For directory cleanup
-import zipfile  # For extracting model archives
-import pathlib  # For path manipulations
-import tempfile  # For creating temporary files/directories
-
-import gradio  # For interactive UI
-import pandas  # For tabular data handling
+import os  # For reading environment variables
+import shutil  # For directory cleanup
+import zipfile  # For extracting model archives
+import pathlib  # For path manipulations
+import tempfile  # For creating temporary files/directories
+
+import gradio  # For interactive UI
+import pandas  # For tabular data handling
 import PIL.Image # For image I/O
 
 import huggingface_hub # For downloading model assets
-import autogluon.multimodal  # For loading AutoGluon image classifiera
+import autogluon.multimodal  # For loading AutoGluon image classifiera
 from huggingface_hub import HfApi
 
-# 1. Initialize the API object (optional for download, but kept)
+# 1. Initialize the API object
 # Hardcoded Hub model (native zip)
 MODEL_REPO_ID = "FaiyazAzam/24679-image-autolguon-predictor"
-ZIP_FILENAME  = "autogluon_image_predictor_dir.zip"
+ZIP_FILENAME  = "autogluon_image_predictor_dir.zip"
 api = HfApi()
 
 # Local cache/extract dirs
-CACHE_DIR   = pathlib.Path("hf_assets")
+CACHE_DIR   = pathlib.Path("hf_assets")
 EXTRACT_DIR = CACHE_DIR / "predictor_native"
 
 # Download & load the native predictor
 def _prepare_predictor_dir() -> str:
-    CACHE_DIR.mkdir(parents=True, exist_ok=True)
-    
-    # *** MODIFICATION HERE ***
-    # Removed 'token=HF_TOKEN' and used 'token=True' (which is the default)
-    # 'token=True' tells huggingface_hub to automatically look for the token 
-    # in the environment variables (like HF_TOKEN in a Space Secret).
-    local_zip = huggingface_hub.hf_hub_download(
-        repo_id=MODEL_REPO_ID,
-        filename=ZIP_FILENAME,
-        repo_type="model",
-        token=True, # Use default token detection (e.g., HF_TOKEN env var)
-        local_dir=str(CACHE_DIR),
-        local_dir_use_symlinks=False,
-    )
-    # *** END MODIFICATION ***
-
-    if EXTRACT_DIR.exists():
-        shutil.rmtree(EXTRACT_DIR)
-    EXTRACT_DIR.mkdir(parents=True, exist_ok=True)
-    with zipfile.ZipFile(local_zip, "r") as zf:
-        zf.extractall(str(EXTRACT_DIR))
-    contents = list(EXTRACT_DIR.iterdir())
-    predictor_root = contents[0] if (len(contents) == 1 and contents[0].is_dir()) else EXTRACT_DIR
-    return str(predictor_root)
+    CACHE_DIR.mkdir(parents=True, exist_ok=True)
+    local_zip = huggingface_hub.hf_hub_download(
+        repo_id=MODEL_REPO_ID,
+        filename=ZIP_FILENAME,
+        repo_type="model",
+        token=True,
+        local_dir=str(CACHE_DIR),
+        local_dir_use_symlinks=False,
+    )
+    if EXTRACT_DIR.exists():
+        shutil.rmtree(EXTRACT_DIR)
+    EXTRACT_DIR.mkdir(parents=True, exist_ok=True)
+    with zipfile.ZipFile(local_zip, "r") as zf:
+        zf.extractall(str(EXTRACT_DIR))
+    contents = list(EXTRACT_DIR.iterdir())
+    predictor_root = contents[0] if (len(contents) == 1 and contents[0].is_dir()) else EXTRACT_DIR
+    return str(predictor_root)
 
 PREDICTOR_DIR = _prepare_predictor_dir()
 PREDICTOR = autogluon.multimodal.MultiModalPredictor.load(PREDICTOR_DIR)
@@ -59,75 +52,75 @@ CLASS_LABELS = {0: "Pen", 1: "Toy"}
 
 # Helper to map model class -> human label
 def _human_label(c):
-    try:
-        ci = int(c)
-        return CLASS_LABELS.get(ci, str(c))
-    except Exception:
-        return CLASS_LABELS.get(c, str(c))
+    try:
+        ci = int(c)
+        return CLASS_LABELS.get(ci, str(c))
+    except Exception:
+        return CLASS_LABELS.get(c, str(c))
 
 # Do the prediction!
 def do_predict(pil_img: PIL.Image.Image):
-    # Make sure there's actually an image to work with
-    if pil_img is None:
-        return "No image provided.", {}, pandas.DataFrame(columns=["Predicted label", "Confidence (%)"])
+    # Make sure there's actually an image to work with
+    if pil_img is None:
+        return "No image provided.", {}, pandas.DataFrame(columns=["Predicted label", "Confidence (%)"])
 
-    # IF we have something to work with, save it and prepare the input
-    tmpdir = pathlib.Path(tempfile.mkdtemp())
-    img_path = tmpdir / "input.png"
-    pil_img.save(img_path)
+    # IF we have something to work with, save it and prepare the input
+    tmpdir = pathlib.Path(tempfile.mkdtemp())
+    img_path = tmpdir / "input.png"
+    pil_img.save(img_path)
 
-    df = pandas.DataFrame({"image": [str(img_path)]})  # For AutoGluon expected input format
+    df = pandas.DataFrame({"image": [str(img_path)]})  # For AutoGluon expected input format
 
-    # For class probabilities
-    proba_df = PREDICTOR.predict_proba(df)
+    # For class probabilities
+    proba_df = PREDICTOR.predict_proba(df)
 
-    # For user-friendly column names
-    proba_df = proba_df.rename(columns={0: "Pen (0)", 1: "Toy (1)"})
-    row = proba_df.iloc[0]
+    # For user-friendly column names
+    proba_df = proba_df.rename(columns={0: "Pen (0)", 1: "Toy (1)"})
+    row = proba_df.iloc[0]
 
-    # For pretty ranked dict expected by gr.Label
-    pretty_dict = {
-        "Pen": float(row.get("Pen (0)", 0.0)),
-        "Toy": float(row.get("Toy (1)", 0.0)),
-    }
+    # For pretty ranked dict expected by gr.Label
+    pretty_dict = {
+        "Pen": float(row.get("Pen (0)", 0.0)),
+        "Toy": float(row.get("Toy (1)", 0.0)),
+    }
 
-    return pretty_dict
+    return pretty_dict
 
 # Representative example images! These can be local or links.
 EXAMPLES = [
-    ["https://www.penboutique.com/cdn/shop/articles/IMG_6759.jpg?v=1701974210&width=1920"],
-    ["https://media.officedepot.com/images/f_auto,q_auto,e_sharpen,h_450/products/790761/790761_p_pilot_g_2_retractable_gel_pens/790761"],
-    ["https://i5.walmartimages.com/seo/Disney-Pixar-Toy-Story-True-Talkers-Woody-Figure-with-15-Phrases_8c8c4a17-fb26-4f97-a284-1315c48c18ca.c35d5f2d8b932a490db9bb3f40977220.jpeg"]
+    ["https://www.penboutique.com/cdn/shop/articles/IMG_6759.jpg?v=1701974210&width=1920"],
+    ["https://media.officedepot.com/images/f_auto,q_auto,e_sharpen,h_450/products/790761/790761_p_pilot_g_2_retractable_gel_pens/790761"],
+    ["https://i5.walmartimages.com/seo/Disney-Pixar-Toy-Story-True-Talkers-Woody-Figure-with-15-Phrases_8c8c4a17-fb26-4f97-a284-1315c48c18ca.c35d5f2d8b932a490db9bb3f40977220.jpeg"]
 ]
 
 # Gradio UI
 with gradio.Blocks() as demo:
 
-    # Provide an introduction
-    gradio.Markdown("# Pen or Toy?")
-    gradio.Markdown("""
-    This is a simple app that demonstrates how to use an autogluon multimodal
-    predictor in a gradio space to predict the contents of a picture. To use,
-    just upload a photo. The result should be generated automatically.
-    """)
+    # Provide an introduction
+    gradio.Markdown("# Pen or Toy?")
+    gradio.Markdown("""
+    This is a simple app that demonstrates how to use an autogluon multimodal
+    predictor in a gradio space to predict the contents of a picture. To use,
+    just upload a photo. The result should be generated automatically.
+    """)
 
-    # Interface for the incoming image
-    image_in = gradio.Image(type="pil", label="Input image", sources=["upload", "webcam"])
+    # Interface for the incoming image
+    image_in = gradio.Image(type="pil", label="Input image", sources=["upload", "webcam"])
 
-    # Interface elements to show htte result and probabilities
-    proba_pretty = gradio.Label(num_top_classes=2, label="Class probabilities")
+    # Interface elements to show htte result and probabilities
+    proba_pretty = gradio.Label(num_top_classes=2, label="Class probabilities")
 
-    # Whenever a new image is uploaded, update the result
-    image_in.change(fn=do_predict, inputs=[image_in], outputs=[proba_pretty])
+    # Whenever a new image is uploaded, update the result
+    image_in.change(fn=do_predict, inputs=[image_in], outputs=[proba_pretty])
 
-      # For clickable example images
-    gradio.Examples(
-        examples=EXAMPLES,
-        inputs=[image_in],
-        label="Representative examples",
-        examples_per_page=8,
-        cache_examples=False,
-    )
+      # For clickable example images
+    gradio.Examples(
+        examples=EXAMPLES,
+        inputs=[image_in],
+        label="Representative examples",
+        examples_per_page=8,
+        cache_examples=False,
+    )
 
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch()