Merge branch 'master' of github.com:etalab/csv-detective

datagouv · Mar 3, 2022 · c26fbc6 · c26fbc6
2 parents 254b3c2 + 7c9c918
commit c26fbc6
Show file tree

Hide file tree

Showing 3 changed files with 6 additions and 6 deletions.
diff --git a/csv_detective/detect_labels/FR/geo/code_commune_insee/__init__.py b/csv_detective/detect_labels/FR/geo/code_commune_insee/__init__.py
@@ -6,10 +6,10 @@
 def _is(header):
     '''Returns 1 if the (processed) header matches one of the expected words combination, else 0'''
 
-    words_combinations_list = ['code commune insee', 'code insee', 'codes insee']
+    words_combinations_list = ['code commune insee', 'code insee', 'codes insee', 'code commune']
     processed_header = _process_text(header)
 
     header_matches_words_combination = float(any([words_combination == processed_header for words_combination in words_combinations_list]))
     words_combination_in_header = 0.5*float(any([full_word_strictly_inside_string(words_combination, processed_header) for words_combination in words_combinations_list]))
 
-    return max(header_matches_words_combination, words_combination_in_header)
+    return max(header_matches_words_combination, words_combination_in_header)
diff --git a/csv_detective/detect_labels/FR/geo/code_departement/__init__.py b/csv_detective/detect_labels/FR/geo/code_departement/__init__.py
@@ -6,10 +6,10 @@
 def _is(header):
     '''Returns 1 if the (processed) header matches one of the expected words combination, else 0'''
 
-    words_combinations_list = ['code departement', 'dep'] #'dep': Possible confusion with dep name?
+    words_combinations_list = ['code departement', 'dep', 'departement'] #'dep': Possible confusion with dep name?
     processed_header = _process_text(header)
 
     header_matches_words_combination = float(any([words_combination == processed_header for words_combination in words_combinations_list]))
     words_combination_in_header = 0.5*float(any([full_word_strictly_inside_string(words_combination, processed_header) for words_combination in words_combinations_list]))
 
-    return max(header_matches_words_combination, words_combination_in_header)
+    return max(header_matches_words_combination, words_combination_in_header)
diff --git a/csv_detective/detect_labels/FR/geo/code_region/__init__.py b/csv_detective/detect_labels/FR/geo/code_region/__init__.py
@@ -6,10 +6,10 @@
 def _is(header):
     '''Returns 1 if the (processed) header matches one of the expected words combination, else 0'''
 
-    words_combinations_list = ['code region', 'reg', 'code insee region'] #'reg' : possible confusion with region name?
+    words_combinations_list = ['code region', 'reg', 'code insee region', 'region'] #'reg' : possible confusion with region name?
     processed_header = _process_text(header)
 
     header_matches_words_combination = float(any([words_combination == processed_header for words_combination in words_combinations_list]))
     words_combination_in_header = 0.5*float(any([full_word_strictly_inside_string(words_combination, processed_header) for words_combination in words_combinations_list]))
 
-    return max(header_matches_words_combination, words_combination_in_header)
+    return max(header_matches_words_combination, words_combination_in_header)