init

src/eda.py

@@ -0,0 +1,78 @@
+import streamlit as st 
+import seaborn as sns 
+import matplotlib.pyplot as plt
+from PIL import Image
+from datasets import load_dataset
+import random
+
+def run():
+    st.title('Tomato Leaf Health Classification')
+    st.subheader("this page contains the EDA about tomato leaf health classification")
+    
+    # image = Image.open("./src/credit_card.jpg")
+    # st.image(image, caption="Credit Card")
+    
+    # write
+    st.write("the EDA will explore and analyse classifier tomato leaf health")
+    
+    # fetch dataset 
+    dataset_dict = load_dataset("Heizsenberg/leaf-image-dataset")
+    label_names = dataset_dict["train"].features["label"].names
+
+    
+    dataset_df = dataset_dict['train'].to_pandas()
+    dataset_df["label_name"] = dataset_df["label"].map(dict(enumerate(label_names)))
+    
+    st.write("sample from the dataframe")
+    st.write(dataset_df.sample(15))
+    
+    st.write("content of the dataframe")
+    st.write("Total images:", len(dataset_df))
+    st.write("Total classes:", dataset_df["label"].nunique())  
+    
+    st.write("Tomato Leaf Training dataset class distribution")
+    fig, ax = plt.subplots(figsize=(10,5))
+    sns.countplot(data=dataset_df, x="label_name", order=dataset_df["label_name"].value_counts().index, ax=ax)
+    plt.xticks(rotation=90)
+    plt.title("Class Distribution")
+
+    st.pyplot(fig)
+    
+    st.write("sample image size and mode")
+    sample_path_obj = random.choice(dataset_df["image"].values)
+    sample_path = sample_path_obj['path']
+    img = Image.open(sample_path)
+
+    st.write("Image size:", img.size)
+    st.write("Image mode:", img.mode)
+    
+    st.write("sample from each classes")
+    fig_samp, ax_samp = plt.subplots(4, 3, figsize=(12,12))
+    
+    # samples = dataset_df.sample(10)
+    samples = dataset_df.groupby("label_name").sample(1, random_state=42)
+
+    for ax, (_, row) in zip(ax_samp.flatten(), samples.iterrows()):
+        image_path = row['image']
+        img = Image.open(image_path['path'])
+        ax.imshow(img)
+        ax.set_title(row["label_name"])
+        ax.axis("off")
+
+    plt.tight_layout()
+
+    # Show inside Streamlit
+    st.pyplot(fig_samp)
+    
+    st.write("""
+             ## Insight
+
+                1. dataset contains around 16.011 in 10 classes
+                2. class distribution generally spread evenly with few exceptions on `tomato_tomato_mosaic_virus` has lowest samples and `Tomato_YellowLeaf_curl_virus` having the largest samples, showing complexity in detecting the diseases and easier detection of tomato mosaic virus
+                3. the dataset images is on size (256x256) which needs to be rescaled for lower GPU load
+                4. several samples is shown from 10 different classes, showing both healthy and disease afflicted leaves
+             """)
+
+    
+if __name__ == '__main__':
+    run()

src/prediction.py

@@ -0,0 +1,72 @@
+import numpy as np
+import streamlit as st
+import tensorflow as tf
+from PIL import Image
+
+class_names = [
+ 'Tomato_Bacterial_spot',
+ 'Tomato_Early_blight',
+ 'Tomato_Late_blight',
+ 'Tomato_Leaf_Mold',
+ 'Tomato_Septoria_leaf_spot',
+ 'Tomato_Spider_mites_Two_spotted_spider_mite',
+ 'Tomato__Target_Spot',
+ 'Tomato__Tomato_YellowLeaf__Curl_Virus',
+ 'Tomato__Tomato_mosaic_virus',
+ 'Tomato_healthy'
+]
+
+@st.cache_resource
+def load_my_model():
+    return tf.keras.models.load_model("leaf_detection_model.keras")
+
+
+def predict_image(model, img):
+    # Preprocess
+    img = img.resize((128, 128))
+    img_array = np.array(img) / 255.0
+    img_array = np.expand_dims(img_array, axis=0)
+
+    # 4. Predict
+    predictions = model.predict(img_array)
+    predicted_index = np.argmax(predictions[0])
+    
+    predicted_label = class_names[predicted_index]
+    
+    confidence = np.max(predictions[0])
+
+    return img, predicted_label, confidence
+
+def run():
+    classifier_model = load_my_model()
+    
+    st.write("upload a tomato leaf image to be predicted")
+    
+    uploaded_file = st.file_uploader(
+        "Choose an tomato leaf image to be uploaded",
+        type=["JPG", "jpg", "jpeg"] # Specify accepted file types
+    )
+    
+    if uploaded_file is not None:
+        st.success("File uploaded successfully!")
+        st.write("Filename:", uploaded_file.name)
+        st.write(uploaded_file)
+        
+        st.write("Image")
+        # To read image file buffer as a PIL Image:
+        img = Image.open(uploaded_file)
+
+        image, predicted_class, probs = predict_image(
+            classifier_model, img
+        )
+
+        confidence = np.max(probs) * 100
+        
+        # show result
+        st.image(img, caption="Uploaded Image", use_container_width=True)
+        st.success(f"Tomato Leaf Prediction: {predicted_class}")
+        st.write(f"with confidence level of: {confidence}")
+
+    
+if __name__ == '__main__':
+    run()

src/streamlit_app.py

@@ -1,40 +1,16 @@
-import altair as alt
-import numpy as np
-import pandas as pd
 import streamlit as st
+import eda 
+import prediction
 
-"""
-# Welcome to Streamlit!
+st.set_page_config(
+    page_title="Tomato Leaf Classifier",
+    layout = 'wide',
+    initial_sidebar_state='expanded'
+)
 
-Edit `/streamlit_app.py` to customize this app to your heart's desire :heart:.
-If you have any questions, checkout our [documentation](https://docs.streamlit.io) and [community
-forums](https://discuss.streamlit.io).
+page = st.sidebar.selectbox('Pilih Page: ', ('EDA', 'Prediction'))
 
-In the meantime, below is an example of what you can do with just a few lines of code:
-"""
-
-num_points = st.slider("Number of points in spiral", 1, 10000, 1100)
-num_turns = st.slider("Number of turns in spiral", 1, 300, 31)
-
-indices = np.linspace(0, 1, num_points)
-theta = 2 * np.pi * num_turns * indices
-radius = indices
-
-x = radius * np.cos(theta)
-y = radius * np.sin(theta)
-
-df = pd.DataFrame({
-    "x": x,
-    "y": y,
-    "idx": indices,
-    "rand": np.random.randn(num_points),
-})
-
-st.altair_chart(alt.Chart(df, height=700, width=700)
-    .mark_point(filled=True)
-    .encode(
-        x=alt.X("x", axis=None),
-        y=alt.Y("y", axis=None),
-        color=alt.Color("idx", legend=None, scale=alt.Scale()),
-        size=alt.Size("rand", legend=None, scale=alt.Scale(range=[1, 150])),
-    ))
+if page == 'EDA':
+    eda.run()
+else:
+    prediction.run()