Add hsanci to nssp

nssp/delphi_nssp/constants.py

@@ -9,6 +9,7 @@
     "state",
     "county",
     "hhs",
+    "hsanci",
 ]
 
 SIGNALS_MAP = {

nssp/delphi_nssp/pull.py

@@ -177,5 +177,5 @@ def pull_nssp_data(
     # Format county fips to all be 5 digits with leading zeros
     df_ervisits["fips"] = df_ervisits["fips"].apply(lambda x: str(x).zfill(5) if str(x) != "0" else "0")
 
-    keep_columns = ["timestamp", "geography", "county", "fips"]
+    keep_columns = ["timestamp", "geography", "county", "fips", "hsa_nci_id"]
     return df_ervisits[SIGNALS + keep_columns]

nssp/delphi_nssp/run.py

@@ -45,6 +45,7 @@ def add_needed_columns(df, col_names=None):
     df = add_default_nancodes(df)
     return df
 
+
 def logging(start_time, run_stats, logger):
     """Boilerplate making logs."""
     elapsed_time_in_seconds = round(time.time() - start_time, 2)
@@ -137,6 +138,11 @@ def run_module(params, logger=None):
                 df = geo_mapper.add_geocode(df, "state_code", "hhs", from_col="state_code", new_col="geo_id")
                 df = geo_mapper.aggregate_by_weighted_sum(df, "geo_id", "val", "timestamp", "population")
                 df = df.rename(columns={"weighted_val": "val"})
+            elif geo == "hsanci":
+                df = df[["hsa_nci_id", "val", "timestamp"]]
+                df = df[df["hsa_nci_id"] != "All"]
+                df = df.groupby(["hsa_nci_id", "timestamp"])["val"].min().reset_index()
+                df = df.rename(columns={"hsa_nci_id": "geo_id"})
             else:
                 df = df[df["county"] != "All"]
                 df["geo_id"] = df["fips"]