Merge pull request #146 from PolicyEngine/350_always_check_for_new_versions

Always check for a new data file version even if one has been downloa…

Author: mikesmit

changelog_entry.yaml

@@ -0,0 +1,4 @@
+- bump: patch
+  changes:
+    fixed:
+    - Always look for new data file versions even if we have a local copy of one.

policyengine/simulation.py

@@ -135,14 +135,13 @@ def _set_data(self):
                         -1
                     ].split("/", 2)
 
-                if not Path(filename).exists():
-                    file_path = download(
-                        filepath=filename,
-                        huggingface_org=hf_org,
-                        huggingface_repo=hf_repo,
-                        gcs_bucket=bucket,
-                    )
-                    filename = str(Path(file_path))
+                file_path = download(
+                    filepath=filename,
+                    huggingface_org=hf_org,
+                    huggingface_repo=hf_repo,
+                    gcs_bucket=bucket,
+                )
+                filename = str(Path(file_path))
             if "cps_2023" in filename:
                 time_period = 2023
             else:

policyengine/utils/data_download.py

@@ -40,10 +40,6 @@ def download(
         except:
             logging.info("Failed to download from Hugging Face.")
 
-    if Path(filepath).exists():
-        logging.info(f"File {filepath} already exists. Skipping download.")
-        return filepath
-
     if data_file.gcs_bucket is not None:
         logging.info("Using Google Cloud Storage for download.")
         download_file_from_gcs(