[Feature] ML model caching for reducing prediction overhead (#156)

The caching is now finished. The cache is created in `settings.py` giving us the most flexibility for using it in the future.

The cache is currently updated/accessed by `tasks.py/get_ml_model` which can be called from whatever task needs to access ml models in this way (currently, `predict` and `predict_simple`).

This implementation currently caches all ml models including the relative reasoning. If we don't want this and only want to cache enviFormer, i can change it to that. However, I don't think there is a harm in having the other models be cached as well.

Co-authored-by: Liam Brydon <62733830+MyCreativityOutlet@users.noreply.github.com>
Reviewed-on: enviPath/enviPy#156
Co-authored-by: liambrydon <lbry121@aucklanduni.ac.nz>
Co-committed-by: liambrydon <lbry121@aucklanduni.ac.nz>

epdb/models.py

@@ -3043,9 +3043,9 @@ class EnviFormer(PackageBasedModel):
     @cached_property
     def model(self):
         from enviformer import load
-
         ckpt = os.path.join(s.MODEL_DIR, "enviformer", str(self.uuid), f"{self.uuid}.ckpt")
-        return load(device=s.ENVIFORMER_DEVICE, ckpt_path=ckpt)
+        mod = load(device=s.ENVIFORMER_DEVICE, ckpt_path=ckpt)
+        return mod
 
     def predict(self, smiles) -> List["PredictionResult"]:
         return self.predict_batch([smiles])[0]
@@ -3059,8 +3059,10 @@ class EnviFormer(PackageBasedModel):
             for smiles in smiles_list
         ]
         logger.info(f"Submitting {canon_smiles} to {self.name}")
+        start = datetime.now()
         products_list = self.model.predict_batch(canon_smiles)
-        logger.info(f"Got results {products_list}")
+        end = datetime.now()
+        logger.info(f"Prediction took {(end - start).total_seconds():.2f} seconds. Got results {products_list}")
 
         results = []
         for products in products_list: