app1

app.py

@@ -45,6 +45,15 @@ from sksurv.metrics import concordance_index_censored
 from sksurv.linear_model import CoxnetSurvivalAnalysis, CoxPHSurvivalAnalysis
 from sksurv.preprocessing import OneHotEncoder
 from sksurv.util import Surv
+from xgboost import XGBRegressor, DMatrix, train as xgb_train
+from lifelines import CoxPHFitter
+from mlflow import sklearn as mlflow_sklearn
+import mlflow.sklearn
+import boto3
+import io
+
+warnings.filterwarnings("ignore", category=UndefinedMetricWarning)
+set_config(display="text")
 
 
 from sksurv.ensemble import GradientBoostingSurvivalAnalysis
@@ -99,26 +108,7 @@ def load_df_merge():
     url = 'https://fireprojectbislead.s3.us-east-1.amazonaws.com/dataset/historique_incendies_avec_coordonnees.csv'
     return pd.read_csv(url, sep=';', encoding='utf-8')
 #------------------------------------------------------- ----------------Notre produit#_________________________________________________
-import streamlit as st
-import pandas as pd
-import numpy as np
-import plotly.express as px
-import warnings
-from sklearn.exceptions import UndefinedMetricWarning
-from sklearn import set_config
-from sklearn.model_selection import train_test_split
-from sklearn.pipeline import Pipeline
-from sklearn.preprocessing import StandardScaler
-from sklearn.impute import SimpleImputer
-from xgboost import XGBRegressor, DMatrix, train as xgb_train
-from lifelines import CoxPHFitter
-from sksurv.util import Surv
-from sksurv.metrics import concordance_index_censored
-from mlflow import sklearn as mlflow_sklearn
-import mlflow.sklearn
 
-warnings.filterwarnings("ignore", category=UndefinedMetricWarning)
-set_config(display="text")
 
 # ────────────────────────────────────────────────
 # 1) FONCTION DE CHARGEMENT DU CSV BRUT
@@ -133,12 +123,17 @@ def load_raw_data() -> pd.DataFrame:
 # 2) FONCTION D’ENTRAÎNEMENT + PRÉDICTIONS
 # ────────────────────────────────────────────────
 
-
-mlflow.set_tracking_uri(os.getenv("BACKEND_STORE_URI"))  # URI NeonDB
 os.environ["MLFLOW_DEFAULT_ARTIFACT_ROOT"] = os.getenv("MLFLOW_DEFAULT_ARTIFACT_ROOT")  # S3
 os.environ["AWS_ACCESS_KEY_ID"] = os.getenv("AWS_ACCESS_KEY_ID")
 os.environ["AWS_SECRET_ACCESS_KEY"] = os.getenv("AWS_SECRET_ACCESS_KEY")
 
+def load_model_from_s3(bucket: str, key: str):
+    s3 = boto3.client("s3")
+    buffer = io.BytesIO()
+    s3.download_fileobj(bucket, key, buffer)
+    buffer.seek(0)
+    model = joblib.load(buffer)
+    return model
 
 # ────────────────────────────────────────────────
 # 2) FONCTION DE PREDICTIONS (sans entraînement)
@@ -168,8 +163,10 @@ def train_model_and_predict(df_raw: pd.DataFrame) -> pd.DataFrame:
     features = [f for f in features if f in df.columns]
 
     # c) Chargement du modèle depuis MLflow
-    model_uri = "models:/fire_survival/Production"  # ou /1 pour la v1
-    pipe = mlflow.sklearn.load_model(model_uri)
+    pipe = load_model_from_s3(
+    bucket=os.getenv("S3_BUCKET"),
+    key="mlflow/models/xgboost_survivalCOX_model_2f26eb52269844a2b01f13974185f102.joblib"  # ton chemin exacts
+)
 
     # d) Estimation baseline hazard avec Cox factice (comme avant)
     y_struct = Surv.from_dataframe("event", "duration", df)