chore(fuzzywuzzy): use thefuzz, fix levenshtein (#747)

* chore * clean * fix type check after updating pyright * Update pyproject.toml Co-authored-by: Marigold <[email protected]>
owid · Jan 9, 2023 · 1f1b887 · 1f1b887
1 parent aa735a3
commit 1f1b887
Show file tree

Hide file tree

Showing 5 changed files with 600 additions and 731 deletions.
diff --git a/etl/match_variables.py b/etl/match_variables.py
@@ -14,7 +14,7 @@
 
 import click
 import pandas as pd
-from fuzzywuzzy import fuzz
+from thefuzz import fuzz
 
 from etl import db
 

diff --git a/etl/scripts/countries-regions/init.py b/etl/scripts/countries-regions/init.py
@@ -1,12 +1,13 @@
 import os
 
 import pandas as pd  # noqa
-from IPython import get_ipython
+from IPython import get_ipython  # type: ignore
 from sqlalchemy import create_engine
 
 engine = create_engine(os.environ["MYSQL_URL"])
 
 ipython = get_ipython()
+assert ipython
 ipython.magic("load_ext rich")
 ipython.magic("load_ext autoreload")
 ipython.magic("autoreload 2")

diff --git a/etl/steps/data/garden/emdat/2022-11-24/shared.py b/etl/steps/data/garden/emdat/2022-11-24/shared.py
@@ -527,7 +527,7 @@ def add_region_aggregates(
             .reset_index()
             .assign(**{country_column: region})
         )
-        data = pd.concat([data, region_df], ignore_index=True)
+        data = pd.concat([data, region_df], ignore_index=True)  # type: ignore
 
     return data