update

.gitignore

@@ -0,0 +1,3 @@
+venv
+**/__pycache__
+.gradio

app.py

@@ -1,71 +1,20 @@
-import pickle
-import gradio as gr
-import shap
-import pandas as pd
-import numpy as np
-import matplotlib.pyplot as plt
-import io
-import base64
+from interface import create_interface
+from model_utils import load_model
 
-# Load model
-with open("best_model.pkl", "rb") as f:
-    model = pickle.load(f)
-
-# Example features - replace with actual features your model uses
-feature_names = ["Age",
-    "weight",
-    "height",
-    "BMI",
-    "gravidity",
-    "parity",
-    "H.Abortion",
-    "living.Child",
-    "Gestational.Age",
-    "Hemoglobin",
-    "hematocrit",
-    "platelet",
-    "MPV.mean.platelet.volume",
-    "PDW.platelet.distribution.width",
-    "neutrophil",
-    "lymphocyte",
-    "NLR.neutrophil.to.lymphocyte",
-    "PLR.platelet.to.lymphocyte.ratio"
-]
-
-
-
-def predict_and_explain(*inputs):
-    # Create a DataFrame for the input
-    input_data = pd.DataFrame([inputs], columns=feature_names)
+def main():
+    print("Loading model...")
+    model = load_model()
     
-    # Prediction
-    prediction = model.predict(input_data)[0]
-
-    # SHAP explanation (on the fly)
-    # explainer = shap.Explainer(model)
-    # shap_values = explainer(input_data)
-    explainer = shap.KernelExplainer(model)
-    shap_values = explainer.shap_values(input_data, nsamples=100)
-
-    # SHAP plot
-    shap_html = shap.plots.force(shap_values[0], matplotlib=False)
-    return f"Risk Prediction: {'High' if prediction else 'Low'}", shap_html
-
-
-# Build the Gradio interface
-input_components = [gr.Number(label=feat) for feat in feature_names]
-output_components = [
-    gr.Textbox(label="Prediction"),
-    gr.HTML(label="Feature Importance")
-]
-
-demo = gr.Interface(
-    fn=predict_and_explain,
-    inputs=input_components,
-    outputs=output_components,
-    title="Pregnancy Risk Analyzer",
-    description="Enter patient data to analyze pregnancy risk and see important features using SHAP"
-)
+    if model is None:
+        print("Warning: Model could not be loaded. Please ensure 'best_model.pkl' exists.")
+    else:
+        print("Model loaded successfully!")
+    
+    print("Creating interface...")
+    demo = create_interface()
+    
+    print("Launching application...")
+    demo.launch(share=True)
 
 if __name__ == "__main__":
-    demo.launch()
+    main()

config.py

@@ -0,0 +1,58 @@
+MODEL_PATH = 'best_model.pkl'
+
+FEATURE_NAMES = [
+    'Age', 'weight', 'height', 'BMI', 'gravidity', 'parity', 'H.Abortion',
+    'living.Child', 'Gestational.Age', 'Hemoglobin', 'hematocrit', 'platelet',
+    'MPV.mean.platelet.volume', 'PDW.platelet.distribution.width', 
+    'neutrophil', 'lymphocyte', 'NLR.neutrophil.to.lymphocyte', 
+    'PLR.platelet.to.lymphocyte.ratio'
+]
+
+FEATURE_NAMES_FA = [
+    'سن', 'وزن', 'قد', 'BMI', 'تعداد بارداری', 'تعداد زایمان', 'تعداد سقط',
+    'فرزند زنده', 'سن بارداری', 'هموگلوبین', 'هماتوکریت', 'پلاکت',
+    'MPV', 'PDW', 'نوتروفیل', 'لنفوسیت', 'NLR', 'PLR'
+]
+
+APP_TITLE = "🩺 سیستم پیش‌بینی سلامت جنین"
+MODEL_ACCURACY = "95.8%"
+MODEL_AUC = "99.3%"
+
+DEFAULT_VALUES = {
+    'age': None, 'weight': None, 'height': None, 'gravidity': None, 'parity': None,
+    'h_abortion': None, 'living_child': None, 'gestational_age': None,
+    'hemoglobin': None, 'hematocrit': None, 'platelet': None, 'mpv': None,
+    'pdw': None, 'neutrophil': None, 'lymphocyte': None
+}
+
+FIELD_RANGES = {
+    'age': {'min': 15, 'max': 60}, 'weight': {'min': 35, 'max': 150},
+    'height': {'min': 130, 'max': 200}, 'gravidity': {'min': 0, 'max': 15},
+    'parity': {'min': 0, 'max': 12}, 'h_abortion': {'min': 0, 'max': 10},
+    'living_child': {'min': 0, 'max': 12}, 'gestational_age': {'min': 1, 'max': 44},
+    'hemoglobin': {'min': 6.0, 'max': 20.0}, 'hematocrit': {'min': 20.0, 'max': 60.0},
+    'platelet': {'min': 50, 'max': 1000}, 'mpv': {'min': 5.0, 'max': 20.0},
+    'pdw': {'min': 8.0, 'max': 30.0}, 'neutrophil': {'min': 0.5, 'max': 15.0},
+    'lymphocyte': {'min': 0.2, 'max': 8.0}
+}
+
+EXAMPLE_CASES = {
+    "مثال ۱: بیمار کم‌خطر": {
+        'age': 28, 'weight': 68, 'height': 165, 'gravidity': 2, 'parity': 1,
+        'h_abortion': 0, 'living_child': 1, 'gestational_age': 32,
+        'hemoglobin': 12.5, 'hematocrit': 38.0, 'platelet': 280,
+        'mpv': 8.5, 'pdw': 15.2, 'neutrophil': 4.2, 'lymphocyte': 2.1
+    },
+    "مثال ۲: بیمار پرخطر": {
+        'age': 42, 'weight': 85, 'height': 158, 'gravidity': 5, 'parity': 3,
+        'h_abortion': 1, 'living_child': 3, 'gestational_age': 28,
+        'hemoglobin': 9.2, 'hematocrit': 28.5, 'platelet': 450,
+        'mpv': 11.8, 'pdw': 18.7, 'neutrophil': 7.8, 'lymphocyte': 1.2
+    },
+    "مثال ۳: بیمار جوان": {
+        'age': 22, 'weight': 58, 'height': 162, 'gravidity': 1, 'parity': 0,
+        'h_abortion': 0, 'living_child': 0, 'gestational_age': 24,
+        'hemoglobin': 11.8, 'hematocrit': 35.2, 'platelet': 220,
+        'mpv': 9.2, 'pdw': 16.1, 'neutrophil': 3.8, 'lymphocyte': 2.5
+    }
+}

interface.py

@@ -0,0 +1,145 @@
+import gradio as gr
+from model_utils import predict_outcome
+from styles import RTL_CSS, HTML_HEAD
+from config import (
+    APP_TITLE, MODEL_ACCURACY, MODEL_AUC, 
+    DEFAULT_VALUES, FIELD_RANGES, EXAMPLE_CASES
+)
+
+def create_patient_info_section():
+    with gr.Column():
+        gr.Markdown("### 📝 اطلاعات بیمار")
+        
+        age = gr.Number(label="سن", value=DEFAULT_VALUES['age'], 
+                       minimum=FIELD_RANGES['age']['min'], maximum=FIELD_RANGES['age']['max'])
+        weight = gr.Number(label="وزن (کیلوگرم)", value=DEFAULT_VALUES['weight'], 
+                          minimum=FIELD_RANGES['weight']['min'], maximum=FIELD_RANGES['weight']['max'])
+        height = gr.Number(label="قد (سانتی‌متر)", value=DEFAULT_VALUES['height'], 
+                          minimum=FIELD_RANGES['height']['min'], maximum=FIELD_RANGES['height']['max'])
+        
+        with gr.Row():
+            gravidity = gr.Number(label="تعداد بارداری", value=DEFAULT_VALUES['gravidity'], 
+                                 minimum=FIELD_RANGES['gravidity']['min'], maximum=FIELD_RANGES['gravidity']['max'])
+            parity = gr.Number(label="تعداد زایمان", value=DEFAULT_VALUES['parity'], 
+                              minimum=FIELD_RANGES['parity']['min'], maximum=FIELD_RANGES['parity']['max'])
+        
+        with gr.Row():
+            h_abortion = gr.Number(label="تعداد سقط", value=DEFAULT_VALUES['h_abortion'], 
+                                  minimum=FIELD_RANGES['h_abortion']['min'], maximum=FIELD_RANGES['h_abortion']['max'])
+            living_child = gr.Number(label="فرزند زنده", value=DEFAULT_VALUES['living_child'], 
+                                    minimum=FIELD_RANGES['living_child']['min'], maximum=FIELD_RANGES['living_child']['max'])
+        
+        gestational_age = gr.Number(label="سن بارداری (هفته)", value=DEFAULT_VALUES['gestational_age'], 
+                                   minimum=FIELD_RANGES['gestational_age']['min'], maximum=FIELD_RANGES['gestational_age']['max'])
+    
+    return age, weight, height, gravidity, parity, h_abortion, living_child, gestational_age
+
+def create_lab_tests_section():
+    with gr.Column():
+        gr.Markdown("### 🧪 آزمایشات خون")
+        
+        hemoglobin = gr.Number(label="هموگلوبین", value=DEFAULT_VALUES['hemoglobin'], 
+                              minimum=FIELD_RANGES['hemoglobin']['min'], maximum=FIELD_RANGES['hemoglobin']['max'])
+        hematocrit = gr.Number(label="هماتوکریت", value=DEFAULT_VALUES['hematocrit'], 
+                              minimum=FIELD_RANGES['hematocrit']['min'], maximum=FIELD_RANGES['hematocrit']['max'])
+        platelet = gr.Number(label="پلاکت", value=DEFAULT_VALUES['platelet'], 
+                            minimum=FIELD_RANGES['platelet']['min'], maximum=FIELD_RANGES['platelet']['max'])
+        
+        with gr.Row():
+            mpv = gr.Number(label="MPV", value=DEFAULT_VALUES['mpv'], 
+                           minimum=FIELD_RANGES['mpv']['min'], maximum=FIELD_RANGES['mpv']['max'])
+            pdw = gr.Number(label="PDW", value=DEFAULT_VALUES['pdw'], 
+                           minimum=FIELD_RANGES['pdw']['min'], maximum=FIELD_RANGES['pdw']['max'])
+        
+        with gr.Row():
+            neutrophil = gr.Number(label="نوتروفیل", value=DEFAULT_VALUES['neutrophil'], 
+                                  minimum=FIELD_RANGES['neutrophil']['min'], maximum=FIELD_RANGES['neutrophil']['max'])
+            lymphocyte = gr.Number(label="لنفوسیت", value=DEFAULT_VALUES['lymphocyte'], 
+                                  minimum=FIELD_RANGES['lymphocyte']['min'], maximum=FIELD_RANGES['lymphocyte']['max'])
+    
+    return hemoglobin, hematocrit, platelet, mpv, pdw, neutrophil, lymphocyte
+
+def predict_with_explanation(age, weight, height, gravidity, parity, h_abortion, 
+                           living_child, gestational_age, hemoglobin, hematocrit, 
+                           platelet, mpv, pdw, neutrophil, lymphocyte):
+    
+    required_fields = [age, weight, height, gravidity, parity, h_abortion, 
+                      living_child, gestational_age, hemoglobin, hematocrit, 
+                      platelet, mpv, pdw, neutrophil, lymphocyte]
+    
+    if any(field is None or field == "" for field in required_fields):
+        return "⚠️ لطفاً تمام فیلدها را پر کنید", "برای پیش‌بینی دقیق، تمام اطلاعات مورد نیاز است.", None
+    
+    result, detailed_report = predict_outcome(
+        age, weight, height, gravidity, parity, h_abortion, 
+        living_child, gestational_age, hemoglobin, hematocrit, 
+        platelet, mpv, pdw, neutrophil, lymphocyte
+    )
+    
+    return result, detailed_report
+
+def clear_all_fields():
+    return tuple([None] * 17)
+
+def load_example(example_name):
+    example_data = EXAMPLE_CASES[example_name]
+    return tuple(example_data[key] for key in [
+        'age', 'weight', 'height', 'gravidity', 'parity', 'h_abortion', 
+        'living_child', 'gestational_age', 'hemoglobin', 'hematocrit', 
+        'platelet', 'mpv', 'pdw', 'neutrophil', 'lymphocyte'
+    ])
+
+def create_interface():
+    with gr.Blocks(title=APP_TITLE, theme=gr.themes.Soft(), css=RTL_CSS, head=HTML_HEAD) as demo:
+        
+        gr.Markdown(f"""
+        # {APP_TITLE}
+        
+        این سیستم با استفاده از مدل هوش مصنوعی **AdaBoost**، احتمال بروز عوارض در بارداری را پیش‌بینی می‌کند.
+        
+        **📊 عملکرد مدل:** دقت {MODEL_ACCURACY} | AUC {MODEL_AUC}
+        
+        🔍 **ویژگی‌های سیستم:**
+        - پیش‌بینی دقیق با استفاده از هوش مصنوعی
+        - تحلیل SHAP برای توضیح تأثیر هر ویژگی
+        - گزارش تفصیلی و قابل فهم برای پزشکان
+        
+        📝 **راهنما:** تمام فیلدها را پر کنید یا از مثال‌های آماده استفاده کنید.
+        """)
+        
+        with gr.Row():
+            patient_inputs = create_patient_info_section()
+            lab_inputs = create_lab_tests_section()
+        
+        with gr.Row():
+            predict_btn = gr.Button("🔍 پیش‌بینی", variant="primary", size="lg")
+            clear_btn = gr.Button("🗑️ پاک کردن", variant="secondary")
+        
+        with gr.Row():
+            with gr.Column(scale=2):
+                result_text = gr.Textbox(label="نتیجه پیش‌بینی", lines=2)
+                detailed_report = gr.Markdown(label="گزارش تفصیلی")
+        
+        gr.Markdown("---")
+        gr.Markdown("## 📚 مثال‌های آماده")
+        
+        with gr.Row():
+            for example_name in EXAMPLE_CASES.keys():
+                example_btn = gr.Button(f"📋 {example_name}", variant="secondary")
+                example_btn.click(
+                    fn=lambda name=example_name: load_example(name),
+                    outputs=list(patient_inputs) + list(lab_inputs)
+                )
+        
+        predict_btn.click(
+            fn=predict_with_explanation,
+            inputs=list(patient_inputs) + list(lab_inputs),
+            outputs=[result_text, detailed_report]
+        )
+        
+        clear_btn.click(
+            fn=clear_all_fields,
+            outputs=list(patient_inputs) + list(lab_inputs) + [result_text, detailed_report]
+        )
+    
+    return demo

model_utils.py

@@ -0,0 +1,79 @@
+import numpy as np
+import joblib
+import warnings
+from config import MODEL_PATH
+
+warnings.filterwarnings('ignore')
+
+model = None
+
+def load_model():
+    global model
+    try:
+        model = joblib.load(MODEL_PATH)
+        return model
+    except Exception as e:
+        print(f"Error loading model: {e}")
+        return None
+
+def calculate_derived_features(age, weight, height, neutrophil, lymphocyte, platelet):
+    height_m = height / 100
+    bmi = weight / (height_m ** 2)
+    nlr = neutrophil / lymphocyte if lymphocyte > 0 else 0
+    plr = platelet / lymphocyte if lymphocyte > 0 else 0
+    return bmi, nlr, plr
+
+def predict_outcome(age, weight, height, gravidity, parity, h_abortion, 
+                   living_child, gestational_age, hemoglobin, hematocrit, 
+                   platelet, mpv, pdw, neutrophil, lymphocyte):
+    global model
+    
+    if model is None:
+        model = load_model()
+        if model is None:
+            return "خطا: مدل بارگذاری نشد", ""
+    
+    try:
+        bmi, nlr, plr = calculate_derived_features(age, weight, height, neutrophil, lymphocyte, platelet)
+        
+        input_data = np.array([[
+            age, weight, height, bmi, gravidity, parity, h_abortion,
+            living_child, gestational_age, hemoglobin, hematocrit, platelet,
+            mpv, pdw, neutrophil, lymphocyte, nlr, plr
+        ]])
+        
+        prediction_proba = model.predict_proba(input_data)[0]
+        prediction = model.predict(input_data)[0]
+        
+        if prediction == 0:
+            result = f"🟢 پیش‌بینی: سالم (احتمال سالم بودن: {prediction_proba[0]*100:.1f}%)"
+            risk_level = "کم"
+        else:
+            result = f"🔴 پیش‌بینی: پرخطر (احتمال عوارض: {prediction_proba[1]*100:.1f}%)"
+            risk_level = "بالا"
+        
+        detailed_report = f"""
+📊 **گزارش تفصیلی پیش‌بینی**
+
+**نتیجه کلی:** {result}
+
+**سطح ریسک:** {risk_level}
+
+**ویژگی‌های محاسبه شده:**
+- BMI: {bmi:.2f}
+- NLR (نسبت نوتروفیل به لنفوسیت): {nlr:.2f}
+- PLR (نسبت پلاکت به لنفوسیت): {plr:.2f}
+
+⚠️ **توجه:** این پیش‌بینی صرفاً جهت کمک به تشخیص است و نباید جایگزین نظر پزشک شود.
+        """
+        
+        return result, detailed_report
+        
+    except Exception as e:
+        return f"خطا در پردازش: {str(e)}", ""
+
+def get_model():
+    global model
+    if model is None:
+        model = load_model()
+    return model

requirements.txt

@@ -1,6 +1,10 @@
 gradio
-scikit-learn
 pandas
 numpy
+scikit-learn
+joblib
 shap
 matplotlib
+xgboost
+catboost
+lightgbm

styles.py

@@ -0,0 +1,80 @@
+RTL_CSS = """
+/* RTL Support for Persian/Arabic */
+.gradio-container {
+    direction: rtl !important;
+    text-align: right !important;
+}
+
+/* Fix input fields alignment */
+.gr-textbox, .gr-number, .gr-slider {
+    direction: rtl !important;
+    text-align: right !important;
+}
+
+/* Fix labels */
+label {
+    direction: rtl !important;
+    text-align: right !important;
+}
+
+/* Fix buttons */
+.gr-button {
+    direction: rtl !important;
+}
+
+/* Fix markdown content */
+.gr-markdown {
+    direction: rtl !important;
+    text-align: right !important;
+}
+
+/* Fix specific input elements */
+input[type="number"], input[type="text"], textarea {
+    direction: rtl !important;
+    text-align: right !important;
+}
+
+/* Fix column layouts */
+.gr-column {
+    direction: rtl !important;
+}
+
+/* Fix row layouts */
+.gr-row {
+    direction: rtl !important;
+}
+
+/* Fix slider component */
+.gr-slider input {
+    direction: ltr !important;
+}
+
+/* Ensure proper spacing for Persian text */
+body {
+    font-family: 'Tahoma', 'Arial', sans-serif !important;
+    direction: rtl !important;
+}
+
+/* Fix any remaining LTR elements */
+* {
+    direction: inherit;
+}
+
+/* Special fixes for gradio components */
+.wrap.svelte-1116kco {
+    direction: rtl !important;
+}
+
+.container.svelte-1116kco {
+    direction: rtl !important;
+}
+"""
+
+HTML_HEAD = """
+<meta charset="UTF-8">
+<meta name="viewport" content="width=device-width, initial-scale=1.0">
+<style>
+    @import url('https://fonts.googleapis.com/css2?family=Vazir:wght@300;400;500;600&display=swap');
+    body { font-family: 'Vazir', 'Tahoma', Arial, sans-serif !important; }
+</style>
+"""