Create app.py

app.py

@@ -0,0 +1,188 @@
+import streamlit as st
+from PIL import Image
+from ultralytics import YOLO
+import cv2, os
+from datetime import datetime
+import numpy as np
+from dotenv import load_dotenv
+import nibabel as nib  # Pour traitement NIfTI IRM
+
+# ---------------- Charger config ----------------
+load_dotenv()
+SAVE_LIMIT_FREE = int(os.getenv("SAVE_LIMIT_FREE", 5))
+PREMIUM_KEY = os.getenv("PREMIUM_KEY", "VOTRE_CLE_PREMIUM")
+
+# ---------------- Config générale ----------------
+SAVE_DIR = os.path.join("/tmp", "results")
+os.makedirs(SAVE_DIR, exist_ok=True)
+
+# ---------------- Modèles ----------------
+MODEL_AVC_PATH = "best.pt"
+model_avc = YOLO(MODEL_AVC_PATH)
+
+MODEL_SEG_PATH = "best_seg.pt"
+model_seg = YOLO(MODEL_SEG_PATH)
+
+# ---------------- Etat utilisateur ----------------
+if "uploads_count" not in st.session_state:
+    st.session_state.uploads_count = 0
+if "premium_access" not in st.session_state:
+    st.session_state.premium_access = False
+
+# ---------------- Fonctions utilitaires ----------------
+def _largest_face_bbox(np_img):
+    import mediapipe as mp
+    mp_face_detection = mp.solutions.face_detection
+    h, w = np_img.shape[:2]
+    with mp_face_detection.FaceDetection(min_detection_confidence=0.6) as fd:
+        results = fd.process(cv2.cvtColor(np_img, cv2.COLOR_RGB2BGR))
+        if not results.detections:
+            return None
+        boxes = []
+        for det in results.detections:
+            rel = det.location_data.relative_bounding_box
+            x1 = int(max(0, rel.xmin) * w)
+            y1 = int(max(0, rel.ymin) * h)
+            x2 = int(min(1.0, rel.xmin + rel.width) * w)
+            y2 = int(min(1.0, rel.ymin + rel.height) * h)
+            boxes.append((x1, y1, x2, y2))
+        boxes.sort(key=lambda b: (b[2]-b[0])*(b[3]-b[1]), reverse=True)
+        return boxes[0] if boxes else None
+
+def check_limit():
+    if not st.session_state.premium_access and st.session_state.uploads_count >= SAVE_LIMIT_FREE:
+        st.warning(f"⚠️ Limite gratuite atteinte ({SAVE_LIMIT_FREE} uploads). Passez en mode premium pour continuer.")
+        return False
+    return True
+
+def predict_image(image, conf=0.85, show_labels=True):
+    if not check_limit():
+        return None
+    np_img = np.array(image)
+    face_bbox = _largest_face_bbox(np_img)
+    if face_bbox is None:
+        st.warning("⚠️ Aucun visage humain détecté. Veuillez centrer le visage.")
+        return None
+    if np_img.shape[2] == 4:
+        np_img = cv2.cvtColor(np_img, cv2.COLOR_RGBA2BGR)
+    else:
+        np_img = cv2.cvtColor(np_img, cv2.COLOR_RGB2BGR)
+
+    results = model_avc.predict(source=np_img, conf=conf, verbose=False)
+    if len(results[0].boxes) == 0:
+        return None
+    annotated_image = results[0].plot(labels=show_labels)
+    out_path = os.path.join(SAVE_DIR, f"image_result_{datetime.now().strftime('%Y%m%d_%H%M%S')}.png")
+    cv2.imwrite(out_path, annotated_image)
+    st.session_state.uploads_count += 1
+    return out_path
+
+def predict_video(video_path, conf=0.85, show_labels=True):
+    if not check_limit():
+        return None
+    cap = cv2.VideoCapture(video_path)
+    fourcc = cv2.VideoWriter_fourcc(*'mp4v')
+    out_path = os.path.join(SAVE_DIR, f"video_result_{datetime.now().strftime('%Y%m%d_%H%M%S')}.mp4")
+    fps = cap.get(cv2.CAP_PROP_FPS) or 30
+    width, height = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH)), int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
+    out = cv2.VideoWriter(out_path, fourcc, fps, (width, height))
+    detections = 0
+    while cap.isOpened():
+        ret, frame = cap.read()
+        if not ret:
+            break
+        results = model_avc.predict(frame, conf=conf, verbose=False)
+        if len(results[0].boxes) > 0:
+            detections += 1
+        annotated = results[0].plot(labels=show_labels)
+        out.write(annotated)
+    cap.release()
+    out.release()
+    if detections == 0:
+        return None
+    st.session_state.uploads_count += 1
+    return out_path
+
+def load_nii_as_image(file):
+    nii = nib.load(file)
+    data = nii.get_fdata()
+    slice_idx = data.shape[2] // 2  # coupe centrale
+    slice_img = data[:, :, slice_idx]
+    slice_img = ((slice_img - slice_img.min()) / (slice_img.max() - slice_img.min()) * 255).astype(np.uint8)
+    return Image.fromarray(slice_img)
+
+# ---------------- Interface Streamlit ----------------
+st.title("🧠 Stroke-IA – Détection AVC par IA")
+
+# ---------------- Sidebar ----------------
+st.sidebar.header("⚙️ Paramètres utilisateur")
+conf_threshold = st.sidebar.slider("Seuil de confiance AVC", 0.1, 1.0, 0.85, 0.05, key="conf_slider")
+show_labels = st.sidebar.checkbox("Afficher les labels", value=True, key="labels_checkbox")
+
+st.sidebar.header("🧠 Paramètres IRM")
+conf_threshold_irm = st.sidebar.slider("Seuil de confiance IRM", 0.1, 1.0, 0.8, 0.05, key="conf_slider_irm")
+
+st.sidebar.header("🔑 Premium / Essai")
+if not st.session_state.premium_access:
+    user_key = st.sidebar.text_input("Entrez votre clé premium :", type="password", key="premium_input")
+    if user_key == PREMIUM_KEY:
+        st.session_state.premium_access = True
+        st.sidebar.success("✅ Mode premium activé ! La limitation est levée.")
+        st.rerun()
+
+if not st.session_state.premium_access:
+    st.sidebar.info(f"📊 Utilisation gratuite : {st.session_state.uploads_count}/{SAVE_LIMIT_FREE}")
+
+# ---------------- Upload vidéo ----------------
+st.header("🎥 Détection sur vidéo")
+video_file = st.file_uploader("Uploader une vidéo", type=["mp4", "mov"], key="video_uploader")
+if video_file and st.button("Analyser la vidéo", key="video_button"):
+    temp_path = os.path.join(SAVE_DIR, "temp_video.mp4")
+    with open(temp_path, "wb") as f:
+        f.write(video_file.read())
+    result_path = predict_video(temp_path, conf=conf_threshold, show_labels=show_labels)
+    if result_path is None:
+        st.success(f"✅ Aucun AVC détecté ou limite gratuite atteinte.")
+    else:
+        st.video(result_path)
+
+# ---------------- Upload image ----------------
+st.header("🖼️ Détection sur image")
+image_file = st.file_uploader("Uploader une image", type=["jpg", "jpeg", "png"], key="image_uploader")
+if image_file and st.button("Analyser l'image", key="image_button"):
+    image = Image.open(image_file)
+    result_path = predict_image(image, conf=conf_threshold, show_labels=show_labels)
+    if result_path is None:
+        st.success(f"✅ Aucun AVC détecté ou limite gratuite atteinte.")
+    else:
+        st.image(result_path, caption="Image annotée", use_container_width=True)
+
+# ---------------- Upload IRM ----------------
+st.header("🧪 Détection sur IRM")
+irm_file = st.file_uploader("Uploader une IRM", type=["jpg", "jpeg", "png", "nii"], key="irm_uploader")
+if irm_file and st.button("Analyser l'IRM", key="irm_button"):
+    image = Image.open(irm_file) if not irm_file.name.endswith(".nii") else load_nii_as_image(irm_file)
+    np_img = np.array(image)
+    if np_img.shape[2] == 4:
+        np_img = cv2.cvtColor(np_img, cv2.COLOR_RGBA2BGR)
+    else:
+        np_img = cv2.cvtColor(np_img, cv2.COLOR_RGB2BGR)
+    results = model_seg.predict(np_img, conf=conf_threshold_irm, verbose=False)
+    if len(results[0].masks.data) == 0:
+        st.warning("⚠️ Aucune anomalie détectée sur cette IRM.")
+    else:
+        annotated_img = results[0].plot()
+        out_path = os.path.join(SAVE_DIR, f"irm_result_{datetime.now().strftime('%Y%m%d_%H%M%S')}.png")
+        cv2.imwrite(out_path, annotated_img)
+        st.image(out_path, caption="IRM annotée", use_container_width=True)
+
+# ---------------- Disclaimer ----------------
+st.markdown(f"""
+---
+👨‍💻 **Badsi Djilali** — Ingénieur Deep Learning  
+🚀 Créateur de **Stroke_IA_Detection**  
+🧠 (Détection d'asymétrie faciale & AVC par IA + anomalies IRM)
+
+⚠️ **Disclaimer :** Stroke-IA est une démo technique, pas un avis médical.  
+© {datetime.now().year} — Badsi Djilali.
+""")