From 4c7d729cf7e1a04a0a0c95d29da425efc8f6bbc0 Mon Sep 17 00:00:00 2001
From: Emma Nechamkin <97977170+emma-nechamkin@users.noreply.github.com>
Date: Wed, 26 Jan 2022 13:57:45 -0500
Subject: [PATCH] Issue 1140 loss rate rounding (#1170)

* updated loss rate rounding

* fixing a typo in variable name

* fixing typo in variable name

* oops, now ready to push

* updated pickle with float for loss rate columns

* updated a typo, now multiplies all loss rates by 100 consistent with other pcts

* updated with final pickles, all tests passing

* updated incorporating lucas pr comments

* changed literal to field name
---
 .../data_pipeline/etl/score/constants.py      |  12 ++++++++++--
 .../data_pipeline/etl/score/etl_score_post.py |  10 ++++++++++
 .../snapshots/downloadable_data_expected.pkl  | Bin 13685 -> 13511 bytes
 3 files changed, 20 insertions(+), 2 deletions(-)

diff --git a/data/data-pipeline/data_pipeline/etl/score/constants.py b/data/data-pipeline/data_pipeline/etl/score/constants.py
index 1bd16abc..d38a56c2 100644
--- a/data/data-pipeline/data_pipeline/etl/score/constants.py
+++ b/data/data-pipeline/data_pipeline/etl/score/constants.py
@@ -69,12 +69,20 @@ PERCENT_PREFIXES_SUFFIXES = [
     "percent",
     "Percentage",
     "Energy burden",
-    "loss rate",
     "greater than or equal to 18 years",
     field_names.PERCENTILE_FIELD_SUFFIX,
 ]
-
 TILES_ROUND_NUM_DECIMALS = 2
+
+# FEMA rounding columns
+FEMA_ROUND_NUM_COLUMNS = [
+    field_names.EXPECTED_BUILDING_LOSS_RATE_FIELD,
+    field_names.EXPECTED_AGRICULTURE_LOSS_RATE_FIELD,
+    field_names.EXPECTED_POPULATION_LOSS_RATE_FIELD,
+]
+
+TILES_FEMA_ROUND_NUM_DECIMALS = 4
+
 # Tiles data: full field name, tile index name
 TILES_SCORE_COLUMNS = {
     field_names.GEOID_TRACT_FIELD: "GTF",
diff --git a/data/data-pipeline/data_pipeline/etl/score/etl_score_post.py b/data/data-pipeline/data_pipeline/etl/score/etl_score_post.py
index c42d7b84..084ab495 100644
--- a/data/data-pipeline/data_pipeline/etl/score/etl_score_post.py
+++ b/data/data-pipeline/data_pipeline/etl/score/etl_score_post.py
@@ -259,6 +259,16 @@ class PostScoreETL(ExtractTransformLoad):
                     pd.to_numeric(df_100, errors="coerce")
                 ).astype("Int64")
                 df[column] = df_int
+
+            elif column in constants.FEMA_ROUND_NUM_COLUMNS:
+                # Convert loss rates by multiplying by 100 (they are percents)
+                # and then rounding appropriately.
+                df_100 = df[column] * 100
+                df[column] = floor_series(
+                    series=df_100.astype(float64),
+                    number_of_decimals=constants.TILES_FEMA_ROUND_NUM_DECIMALS,
+                )
+
             else:
                 # Round all other floats.
                 df[column] = floor_series(
diff --git a/data/data-pipeline/data_pipeline/etl/score/tests/snapshots/downloadable_data_expected.pkl b/data/data-pipeline/data_pipeline/etl/score/tests/snapshots/downloadable_data_expected.pkl
index 5608343f42a9ba3290dcecdc7dc2a7020a3241ce..1043c8bdbc2c1cc4b3efe9c726d167b3eaf05c8e 100644
GIT binary patch
delta 239
zcmeyGbv%=`fo19*lZ~u@SSLG(i%l-)(v-G;u)V<~_vkfu`={-b?@s@womw&_DZ|E_
z)0=CuA)o8y4Q!bpwbB_IJ7cG4xZ5vT@6X%f%jj;u?PbN4zn`}Ol~1l=-wINr1XH7B
z_UN^Yv#z^6lmRu8-&+7=<mLw)GZ+~+PM*#6geha=<Q8rawK<nZgc-zb;|Gi83xKHz
zK``Yg1XiOZtiu9Q8!gVn!MI{`oV=ze<EqIWCMGQFSQr>4U*s2?e9gp~k!`Y&X)*w1
C4O_wh

delta 193
zcmX?}`8A8Rfn{oz=|<K+tdr;SC{I2o$-&4lnUT|)nbVtVvLn0e<OOV*s4|o7*f*kz
zY*yiTz&KfgO=B__*L~KkOhyKV$*;MdKv<hEa*HrS1)uXl1t*{7hp{#bz*sW{VOr{i
nbXcIe*NQQ5Fdp8#K~7VY@#y4R#wLs>C-a+xFe*+?F-ZmhILtY$