Update app.po

app.po

@@ -460,6 +460,41 @@ from PIL import Image
 # -------------------------------------------------------
 # Soil Recognizer
 # -------------------------------------------------------
+import torch
+import torch.nn as nn
+import torchvision.transforms as transforms
+from PIL import Image
+import streamlit as st
+
+# ----------------------------
+# Load Soil Recognition Model
+# ----------------------------
+@st.cache_resource
+def load_soil_model():
+    try:
+        model = torch.load("soil_best_model.pth", map_location=torch.device("cpu"))
+        model.eval()
+        return model
+    except Exception as e:
+        st.error(f"⚠️ Could not load soil model: {e}")
+        return None
+
+soil_model = load_soil_model()
+
+# Define soil classes (adjust if your model has different labels)
+SOIL_CLASSES = ["Sand", "Silt", "Clay", "Gravel", "Peat"]
+
+# Image preprocessing pipeline
+transform = transforms.Compose([
+    transforms.Resize((224, 224)),   # match training input size
+    transforms.ToTensor(),
+    transforms.Normalize([0.485, 0.456, 0.406],  # ImageNet normalization
+                         [0.229, 0.224, 0.225])
+])
+
+# ----------------------------
+# Soil Recognizer UI
+# ----------------------------
 def soil_recognizer_ui():
     st.header("🖼️ Soil Recognizer (Image / OCR)")
     site = get_active_site()
@@ -467,29 +502,34 @@ def soil_recognizer_ui():
     col1, col2 = st.columns(2)
 
     with col1:
-        uploaded = st.file_uploader("Upload soil image", type=["jpg","jpeg","png"])
-        if uploaded:
-            img = Image.open(uploaded)
+        uploaded = st.file_uploader("Upload soil image", type=["jpg", "jpeg", "png"])
+        if uploaded is not None:
+            img = Image.open(uploaded).convert("RGB")
             st.image(img, caption="Uploaded soil image", use_column_width=True)
-            # TODO: integrate your trained soil recognition model
-            st.success("Soil recognizer placeholder: model inference to be integrated.")
-    
-    with col2:
-        st.subheader("📑 OCR Extraction")
-        if easyocr is None:
-            st.warning("easyocr not installed. Add `easyocr` to requirements.txt.")
-        else:
-            ocr_file = st.file_uploader("Upload photo of question/text (OCR)", type=["jpg","jpeg","png"], key="ocr_input")
-            if ocr_file:
-                reader = easyocr.Reader(['en'])
-                results = reader.readtext(np.array(Image.open(ocr_file)))
-                extracted_text = " ".join([r[1] for r in results])
-                st.text_area("Extracted text", extracted_text, height=150)
-                # TODO: parse extracted numbers for classification if possible
-                site["ocr_pending"] = True
-                save_active_site(site)
-                st.success("OCR text extracted. Parsed values will be linked to classifier soon.")
 
+            if soil_model:
+                try:
+                    # Preprocess
+                    img_t = transform(img).unsqueeze(0)  # add batch dimension
+                    with torch.no_grad():
+                        outputs = soil_model(img_t)
+                        probs = torch.softmax(outputs, dim=1)
+                        conf, pred = torch.max(probs, 1)
+                        predicted_class = SOIL_CLASSES[pred.item()]
+                        confidence = conf.item()
+
+                    # Save to site
+                    site["Soil Profile"] = predicted_class
+                    site["classifier_inputs"]["image_confidence"] = confidence
+                    save_sites(SITES)
+
+                    st.success(f"✅ Predicted: **{predicted_class}** ({confidence:.2%} confidence)")
+
+                except Exception as e:
+                    st.error(f"❌ Inference error: {e}")
+            else:
+                st.warning("⚠️ Soil model not loaded. Please check `soil_best_model.pth`.")
+    
 # -------------------------------------------------------
 # Soil Classifier
 # -------------------------------------------------------