Skip to content

Commit cc083f4

Browse files
committed
Removing aliases that are in iso639
1 parent 7db3e8b commit cc083f4

File tree

1 file changed

+0
-4
lines changed

1 file changed

+0
-4
lines changed

scripts/1-fetch/internetarchive_fetch.py

Lines changed: 0 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -214,7 +214,6 @@ def normalize_language(raw_language):
214214
cleaned = normalize_key(cleaned_for_match.replace("-", " "))
215215

216216
ALIAS_MAP = {
217-
"english": "English",
218217
"engrish": "English",
219218
"english_handwritten": "English",
220219
"enlgish": "English",
@@ -240,17 +239,14 @@ def normalize_language(raw_language):
240239
"es_es": "Spanish",
241240
"mandarin": "Chinese",
242241
"nederlands": "Dutch",
243-
"dutch": "Dutch",
244242
"swahili": "Swahili",
245243
"no language (english)": "Undetermined",
246244
"whatever we play it to be": "Undetermined",
247245
"english & chinese subbed": "Multiple languages",
248-
"mis": "Uncoded languages",
249246
"n/a": "Undetermined",
250247
"none": "Undetermined",
251248
"und": "Undetermined",
252249
"unknown": "Undetermined",
253-
"und": "Undetermined",
254250
"no language (english)": "Undetermined",
255251
"no speech": "Undetermined",
256252
"no spoken language": "Undetermined",

0 commit comments

Comments
 (0)