app.py

# import streamlit as st
# import cv2
# from streamlit_drawable_canvas import st_canvas
# from keras.models import load_model
# import numpy as np

# # Sidebar controls
# st.sidebar.title("Canvas Settings")
# drawing_mode = st.sidebar.selectbox("Drawing tool:", ("freedraw", "line", "rect", "circle", "transform"))
# stroke_width = st.sidebar.slider("Stroke width: ", 1, 25, 10)
# stroke_color = st.sidebar.color_picker("Stroke color hex: ", "#000000")  # black
# bg_color = st.sidebar.color_picker("Background color hex: ", "#FFFFFF")  # white
# bg_image = st.sidebar.file_uploader("Background image:", type=["png", "jpg"])
# realtime_update = st.sidebar.checkbox("Update in realtime", True)

# # Load model with caching
# @st.cache_resource
# def load_mnist_model():
#     return load_model("mnist_model.keras")

# model = load_mnist_model()

# st.title("🖌️ Mindist: Draw a Number, Predict Instantly")

# # Create a two-column layout
# col1, col2 = st.columns([1, 1])

# with col1:
#     st.subheader("Draw Here 👇")
#     canvas_result = st_canvas(
#         fill_color="rgba(255, 165, 0, 0.3)",
#         stroke_width=stroke_width,
#         stroke_color=stroke_color,
#         background_color=bg_color,
#         update_streamlit=realtime_update,
#         height=280,
#         width=280,
#         drawing_mode=drawing_mode,
#         key="canvas",
#     )

# with col2:
#     if canvas_result.image_data is not None:
#         st.subheader("Original Drawing")
#         st.image(canvas_result.image_data, use_column_width=True)

# # Below the two columns: Show preprocessing and prediction
# if canvas_result.image_data is not None:
#     st.markdown("---")
#     st.subheader("Preprocessed Image & Prediction")
    
#     img = cv2.cvtColor(canvas_result.image_data.astype("uint8"), cv2.COLOR_RGBA2GRAY)
#     img = 255 - img  # Invert colors
#     img_resized = cv2.resize(img, (28, 28))
#     img_normalized = img_resized / 255.0
#     final_img = img_normalized.reshape(1, 28, 28, 1)
    
#     col3, col4 = st.columns([1, 1])
#     with col3:
#         st.image(img_resized, caption="28x28 Preprocessed", clamp=True, channels="GRAY")
#     with col4:
#         prediction = model.predict(final_img)
#         predicted_digit = np.argmax(prediction)
#         st.markdown(f"### 🧠 Predicted Digit: **{predicted_digit}**")
import streamlit as st
import cv2
import numpy as np
from streamlit_drawable_canvas import st_canvas
from keras.models import load_model

# Sidebar controls
st.sidebar.title("Canvas Settings")
drawing_mode = st.sidebar.selectbox("Drawing tool:", ("freedraw", "line", "rect", "circle", "transform"))
stroke_width = st.sidebar.slider("Stroke width: ", 1, 25, 10)
stroke_color = st.sidebar.color_picker("Stroke color hex: ", "#000000")  # black
bg_color = st.sidebar.color_picker("Background color hex: ", "#FFFFFF")  # white
bg_image = st.sidebar.file_uploader("Background image:", type=["png", "jpg"])
realtime_update = st.sidebar.checkbox("Update in realtime", True)

# Mode selection
mode = st.sidebar.radio("Select Prediction Mode", ["Single Digit", "Multi Digit"])

# === Load models ===
@st.cache_resource
def load_single_digit_model():
    return load_model("mnist_model.keras")

@st.cache_resource
def load_multi_digit_model():
    return load_model("best_model.keras")  # Your multi-digit model

model_single = load_single_digit_model()
model_multi = load_multi_digit_model()

# === Streamlit UI ===
st.title("🧠 Digit Recognition App")
st.subheader(f"✏️ Mode: {mode}")

# Create drawing canvas
canvas_result = st_canvas(
    fill_color="rgba(255, 165, 0, 0.3)",
    stroke_width=stroke_width,
    stroke_color=stroke_color,
    background_color=bg_color,
    update_streamlit=realtime_update,
    height=280,
    width=280,
    drawing_mode=drawing_mode,
    key="canvas",
)

# Prediction Section
if canvas_result.image_data is not None:
    st.markdown("---")
    st.subheader("🧪 Prediction Results")

    # Preprocess drawing
    img = cv2.cvtColor(canvas_result.image_data.astype("uint8"), cv2.COLOR_RGBA2GRAY)
    img = 255 - img  # Invert
    _, thresh = cv2.threshold(img, 30, 255, cv2.THRESH_BINARY)

    if mode == "Single Digit":
        resized = cv2.resize(thresh, (28, 28))
        norm = resized.astype("float32") / 255.0
        input_img = norm.reshape(1, 28, 28, 1)

        prediction = model_single.predict(input_img)
        digit = np.argmax(prediction)

        col1, col2 = st.columns(2)
        with col1:
            st.image(resized, width=200, caption="28x28 Preprocessed")
        with col2:
            st.success(f"🧠 Predicted Digit: **{digit}**")

    elif mode == "Multi Digit":
        resized = cv2.resize(thresh, (80, 28))  # Resize to match your model (width=80, height=28)
        norm = resized.astype("float32") / 255.0
        input_seq = norm.reshape(1, 28, 80, 1)

        preds = model_multi.predict(input_seq)

        # Decode predictions for each digit
        predicted_digits = [np.argmax(p[0]) for p in preds]
        predicted_str = ''.join(str(d) for d in predicted_digits)

        col1, col2 = st.columns(2)
        with col1:
            st.image(resized, width=300, caption="80x28 Multi-digit Input")
        with col2:
            st.success(f"🧠 Predicted Sequence: **{predicted_str}**")