Merge pull request #132 from washingtonpost/release/2.2.2

Release/2.2.2 🎉
washingtonpost · Nov 5, 2024 · ebc3a40 · ebc3a40
2 parents e93515e + 838fd12
commit ebc3a40
Show file tree

Hide file tree

Showing 2 changed files with 14 additions and 0 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,5 +1,8 @@
 # Changelog
 
+# 2.2.2 (11/5/2024)
+- fix: missing `est_correction` column in `VersionedResults` `DataFrame` in the event of bad data [#131](https://github.com/washingtonpost/elex-live-model/pull/131)
+
 # 2.2.1 (11/1/2024)
 - chore: downgrade botocore and s3transfer as per live team dependency [#128](https://github.com/washingtonpost/elex-live-model/pull/128)
 

diff --git a/src/elexmodel/handlers/data/VersionedData.py b/src/elexmodel/handlers/data/VersionedData.py
@@ -127,6 +127,8 @@ def compute_estimated_margin(df):
                         "percent_expected_vote": np.arange(101),
                         "nearest_observed_vote": np.nan * np.ones(101),
                         "est_margin": np.nan * np.ones(101),
+                        "est_correction": np.nan * np.ones(101),
+                        "error_type": "non-monotone percent expected vote",
                     }
                 )
 
@@ -152,6 +154,8 @@ def compute_estimated_margin(df):
                         "percent_expected_vote": np.arange(101),
                         "nearest_observed_vote": np.nan * np.ones(101),
                         "est_margin": np.nan * np.ones(101),
+                        "est_correction": np.nan * np.ones(101),
+                        "error_type": "batch_margin",
                     }
                 )
 
@@ -196,10 +200,17 @@ def compute_estimated_margin(df):
                     "nearest_observed_vote": percent_vote[np.clip(obs_indices + 1, 0, len(percent_vote) - 1)],
                     "est_margin": est_margins,
                     "est_correction": norm_margin[-1] - est_margins,
+                    "error_type": "none",
                 }
             )
 
         results = results.groupby("geographic_unit_fips").apply(compute_estimated_margin).reset_index()
+
+        for error_type in sorted(set(results["error_type"])):
+            if error_type == "none":
+                continue
+            category_error_type = results[results["error_type"] == error_type].geographic_unit_fips.unique()
+            LOG.info(f"# of versioned units with {error_type} error: {len(category_error_type)}")
         return results
 
     def get_versioned_predictions(self, filepath=None):