Update app.py

app.py

@@ -1,7 +1,8 @@
 import gradio as gr
 from PIL import Image
 import torch
-import tempfile
+import base64
+import io
 
 from transformers import (
     pipeline,
@@ -9,11 +10,6 @@ from transformers import (
     AutoImageProcessor
 )
 
-from staticmap import StaticMap, CircleMarker
-
-# -------------------------
-# Model yükle
-# -------------------------
 MODEL_ID = "cagrigungor/fire-prediction"
 
 model = AutoModelForImageClassification.from_pretrained(MODEL_ID)
@@ -29,69 +25,68 @@ pipe = pipeline(
 # -------------------------
 # Inference
 # -------------------------
-def predict(image: Image.Image):
+def predict_from_image(image):
     if image is None:
-        return {}
+        return None
 
     image = image.convert("RGB")
     results = pipe(image)
 
     return {r["label"]: float(r["score"]) for r in results}
 
-# -------------------------
-# OSM'den görüntü üret
-# -------------------------
-def osm_to_image(lat, lon, zoom):
-    m = StaticMap(224, 224)
-    marker = CircleMarker((lon, lat), "red", 6)
-    m.add_marker(marker)
-
-    image = m.render(zoom=zoom)
-    return image
 
 # -------------------------
-# OSM + Prediction
+# Base64 (haritadan gelen)
 # -------------------------
-def predict_from_osm(lat, lon, zoom):
-    img = osm_to_image(lat, lon, zoom)
-    preds = predict(img)
-    return img, preds
+def predict_from_base64(base64_str):
+    if base64_str is None:
+        return None
 
-# -------------------------
-# UI
-# -------------------------
-with gr.Blocks(title="🔥 Wildfire Detection + OSM") as app:
-    gr.Markdown("## 🔥 Wildfire Detection (Upload + OpenStreetMap)")
-
-    with gr.Tab("📁 Manuel Görüntü Yükle"):
-        upload_img = gr.Image(type="pil", label="Görüntü Yükle")
-        upload_btn = gr.Button("Tahmin Et")
-        upload_out = gr.Label(num_top_classes=2)
-
-        upload_btn.click(
-            fn=predict,
-            inputs=upload_img,
-            outputs=upload_out
-        )
-
-    with gr.Tab("🗺 OpenStreetMap"):
-        lat = gr.Number(label="Latitude", value=39.0)
-        lon = gr.Number(label="Longitude", value=35.0)
-        zoom = gr.Slider(5, 15, value=10, step=1, label="Zoom")
+    image_bytes = base64.b64decode(base64_str.split(",")[1])
+    image = Image.open(io.BytesIO(image_bytes)).convert("RGB")
+    return predict_from_image(image)
 
-        osm_btn = gr.Button("Haritadan Tahmin Et")
-
-        osm_img = gr.Image(label="OSM Görüntüsü")
-        osm_out = gr.Label(num_top_classes=2)
-
-        osm_btn.click(
-            fn=predict_from_osm,
-            inputs=[lat, lon, zoom],
-            outputs=[osm_img, osm_out]
-        )
 
 # -------------------------
-# Run
+# UI
 # -------------------------
-if __name__ == "__main__":
-    app.launch()
+with gr.Blocks(title="🔥 Wildfire Detection with Map") as app:
+    gr.Markdown("## 🔥 Wildfire Detection (OSM + Image Upload)")
+
+    with gr.Tabs():
+        # -------------------------
+        # TAB 1: Manual Upload
+        # -------------------------
+        with gr.Tab("📤 Resim Yükle"):
+            img_input = gr.Image(type="pil", label="Görüntü Yükle")
+            btn1 = gr.Button("Tahmin Et")
+            out1 = gr.Label(num_top_classes=2)
+
+            btn1.click(
+                fn=predict_from_image,
+                inputs=img_input,
+                outputs=out1
+            )
+
+        # -------------------------
+        # TAB 2: Map
+        # -------------------------
+        with gr.Tab("🗺️ Harita (OSM)"):
+            gr.HTML("""
+            <iframe
+              src="/map.html"
+              style="width:100%; height:600px; border:none;">
+            </iframe>
+            """)
+
+            base64_input = gr.Textbox(visible=False)
+            btn2 = gr.Button("Haritadan Tahmin Et")
+            out2 = gr.Label(num_top_classes=2)
+
+            btn2.click(
+                fn=predict_from_base64,
+                inputs=base64_input,
+                outputs=out2
+            )
+
+app.launch()