Upload streamlit_app.py

src/streamlit_app.py

@@ -11,6 +11,11 @@ import tensorflow.keras.backend as K
 from huggingface_hub import hf_hub_download
 import os
 
+# ✅ Safe directory for Spaces
+os.environ["HF_HOME"] = "/tmp/hf_home"
+os.environ["HF_HUB_CACHE"] = "/tmp/hf_home/hub_cache"
+os.environ["MPLCONFIGDIR"] = "/tmp/matplotlib"
+
 # === Custom Metric ===
 def pearson_corr(y_true, y_pred):
     x = y_true - K.mean(y_true)
@@ -18,19 +23,17 @@ def pearson_corr(y_true, y_pred):
     return K.sum(x * y) / (K.sqrt(K.sum(K.square(x))) * K.sqrt(K.sum(K.square(y))) + K.epsilon())
 
 # === Load model and scaler ===
-# ✅ Force cache to local writable directory
-os.makedirs("./hf_home/hub_cache", exist_ok=True)
-os.environ["HF_HOME"] = "./hf_home"
-os.environ["HF_HUB_CACHE"] = "./hf_home/hub_cache"
 
 model_path = hf_hub_download(
     repo_id="jiinkwan/bmi-model",  # 🔁 This should match your new model repo
-    filename="bmi_model_gender.keras"
+    filename="bmi_model_gender.keras",
+    repo_type="space"
 )
 
 scaler_path = hf_hub_download(
     repo_id="jiinkwan/bmi-model",  # 🔁 This should match your new model repo
-    filename="label_scaler.keras"
+    filename="label_scaler.keras",
+    repo_type="space"
 )
 
 model = tf.keras.models.load_model(model_path, custom_objects={'pearson_corr': pearson_corr})