edit errors

app.py

@@ -5,64 +5,58 @@ from autodistill_grounding_dino import GroundingDINO
 from autodistill.utils import plot
 import tempfile
 import cv2
-
 from autodistill.core.custom_detection_model import CustomDetectionModel
 
-# Hardcoded values - ensure to protect the API key
-api_key = "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"
-dino_prompt = "buildings. parks."
+# Hardcoded values
+api_key = "your-hardcoded-api-key"
+dino_prompt = "buildings, parks."
 gpt_prompt = "carbon sinks"
 
 MARKDOWN = """
-# NESGPT-AutoAnnotator
-Grounding DINO for Zero-Shot Carbon Sink Identification and Object Segmentation
-."""
+# DINO-GPT4V
+Use Grounding DINO and GPT-4V to label specific objects.
+Visit [awesome-openai-vision-api-experiments](https://github.com/roboflow/awesome-openai-vision-api-experiments) 
+repository to find more OpenAI Vision API experiments or contribute your own."""
 
 def respond(input_image):
     input_image = cv2.cvtColor(input_image, cv2.COLOR_BGR2RGB)
     with tempfile.NamedTemporaryFile(delete=False, suffix=".jpg") as temp_file:
         cv2.imwrite(temp_file.name, input_image)
 
-        try:
-            DINOGPT = CustomDetectionModel(
-                detection_model=GroundingDINO(
-                    CaptionOntology({dino_prompt: dino_prompt})
-                ),
-                classification_model=GPT4V(
-                    CaptionOntology({k: k for k in gpt_prompt.split(", ")}),
-                    api_key=api_key  # Use the hardcoded API key
-                )
+        DINOGPT = CustomDetectionModel(
+            detection_model=GroundingDINO(
+                CaptionOntology({dino_prompt: dino_prompt})
+            ),
+            classification_model=GPT4V(
+                CaptionOntology({k: k for k in gpt_prompt.split(", ")}),
+                api_key=api_key
             )
+        )
 
-            results = DINOGPT.predict(temp_file.name)
+        results = DINOGPT.predict(temp_file.name)
 
-            # Assuming the results are a list of dictionaries with 'class' keys
-            # Extracting class names from results
-            predicted_classes = [result['class'] for result in results]
-            # Filtering the ontology prompts to include only those that match the predicted classes
-            filtered_ontology = {k: v for k, v in DINOGPT.classification_model.ontology.prompts().items() if k in predicted_classes}
+        if isinstance(results, tuple):
+            # If results are a tuple, handle it accordingly
+            # This is a placeholder, you need to adjust based on the actual structure of the tuple
+            results = results[0]  # Assuming the first item in the tuple is the desired data
 
-            result_image = plot(
-                image=cv2.imread(temp_file.name),
-                detections=results,
-                classes=list(filtered_ontology.keys()),  # Use the filtered ontology
-                raw=True
-            )
-        except ValueError as e:
-            print("An error occurred:", e)
-            # Handle the error appropriately, perhaps by returning an error message or default image
-            result_image = None  # Replace with your error handling logic
+        result = plot(
+            image=cv2.imread(temp_file.name),
+            detections=results,
+            classes=gpt_prompt.split(", "),
+            raw=True
+        )
 
-    return result_image
+    return result
 
 with gr.Blocks() as demo:
     gr.Markdown(MARKDOWN)
     with gr.Row():
         with gr.Column():
-            input_image = gr.Image(type="numpy", label="Input Location")
+            input_image = gr.Image(type="numpy", label="Input Image")
         with gr.Column():
-            output_image = gr.Image(type="numpy", label="Carbon Sink Identifier")
-            submit_button = gr.Button("Identify Carbon Sinks")
+            output_image = gr.Image(type="numpy", label="Output Image")
+            submit_button = gr.Button("Submit")
 
     submit_button.click(
         fn=respond,