NANs for CAN:

* add missing columns, allow nan values through

Author: dshemetov

covid_act_now/delphi_covid_act_now/run.py

@@ -12,12 +12,34 @@
 from delphi_utils import (
     create_export_csv,
     get_structured_logger
+    S3ArchiveDiffer,
+    Nans
 )
 
 from .constants import GEO_RESOLUTIONS, SIGNALS
 from .geo import geo_map
 from .pull import load_data, extract_testing_metrics
 
+def add_nancodes(df, signal):
+    """Add nancodes to the dataframe."""
+    # Default missingness codes
+    df["missing_val"] = Nans.NOT_MISSING
+    df["missing_se"] = Nans.NOT_MISSING if signal == "pcr_tests_positive" else Nans.NOT_APPLICABLE
+    df["missing_sample_size"] = (
+        Nans.NOT_MISSING if signal == "pcr_tests_positive" else Nans.NOT_APPLICABLE
+    )
+
+    # Mark any nans with unknown
+    val_nans_mask = df["val"].isnull()
+    df.loc[val_nans_mask, "missing_val"] = Nans.OTHER
+    if signal == "pcr_tests_positive":
+        se_nans_mask = df["se"].isnull()
+        df.loc[se_nans_mask, "missing_se"] = Nans.OTHER
+        sample_size_nans_mask = df["sample_size"].isnull()
+        df.loc[sample_size_nans_mask, "missing_sample_size"] = Nans.OTHER
+
+    return df
+
 def run_module(params):
     """
     Run the CAN testing metrics indicator.
@@ -59,6 +81,7 @@ def run_module(params):
         df = geo_map(df_county_testing, geo_res)
 
         # Export 'pcr_specimen_positivity_rate'
+        df = add_nancodes(df, "pcr_tests_positive")
         exported_csv_dates = create_export_csv(
             df,
             export_dir=export_dir,
@@ -69,6 +92,7 @@ def run_module(params):
         df["val"] = df["sample_size"]
         df["sample_size"] = np.nan
         df["se"] = np.nan
+        df = add_nancodes(df, "pcr_tests_total")
         exported_csv_dates = create_export_csv(
             df,
             export_dir=export_dir,

covid_act_now/tests/test_run.py

@@ -21,6 +21,7 @@ def test_output_files(self, clean_receiving_dir):
         run_module(self.PARAMS)
         csv_files = set(listdir("receiving"))
         csv_files.discard(".gitignore")
+        today = pd.Timestamp.today().date().strftime("%Y%m%d")
 
         expected_files = set()
         for signal in SIGNALS:
@@ -30,7 +31,11 @@ def test_output_files(self, clean_receiving_dir):
         # All output files exist
         assert csv_files == expected_files
 
+        expected_columns = [
+            "geo_id", "val", "se", "sample_size",
+            "missing_val", "missing_se", "missing_sample_size"
+        ]
         # All output files have correct columns
         for csv_file in csv_files:
             df = pd.read_csv(join("receiving", csv_file))
-            assert (df.columns.values == ["geo_id", "val", "se", "sample_size"]).all()
+            assert (df.columns.values == expected_columns).all()