Additional comments

backend/requirements.txt

@@ -14,7 +14,7 @@ gcsfs==2024.12.0
 google-cloud-storage==2.11.0
 
 # RAG & embeddings
-pytorch-lightning==2.5.1           
+pytorch-lightning==2.5.1
 langchain
 langchain-huggingface
 sentence-transformers
@@ -44,3 +44,14 @@ HuggingFace
 huggingface_hub
 huggingface_hub[hf_xet]
 
+# ————————————————————————————————————————————————
+
+joblib
+shap
+matplotlib
+scipy
+scikit-learn
+imbalanced-learn
+xgboost
+evidently
+optuna

data_enhancement.py

@@ -33,7 +33,6 @@ DATASETS = [
 ]
 
 # ==== REGEX FOR CLEANING =====================================================
-
 _PROJECT_ID_RE = re.compile(r"^(?:19|20)\d{2}")
 _GENERIC_NUM_RE = re.compile(r"\d{4}")
 

nginx.conf

@@ -22,7 +22,7 @@ http {
         listen       [::]:4444 default_server;
         server_name  _;
 
-        # Proxy the React/Vite build on port 8000
+        # Proxy React/Vite build on port 8000
         location / {
             proxy_pass         http://127.0.0.1:8000;
             proxy_http_version 1.1;
@@ -36,7 +36,7 @@ http {
             proxy_redirect     off;
         }
 
-        # Proxy your FastAPI app on port 7860
+        # Proxy FastAPI app on port 7860
         location /api/ {
             proxy_pass         http://127.0.0.1:7860/api/;
             proxy_http_version 1.1;

predictive_modelling.py

@@ -34,7 +34,6 @@ from evidently.presets import DataDriftPreset
 import optuna
 
 # --- Custom Transformers ---
-
 # Transformer for binarizing multi-label columns (lists of categories per row)
 class MultiLabelBinarizerTransformer(BaseEstimator, TransformerMixin):
     """Custom transformer to binarize multi-label (list of strings) columns.