Fix bug in subsetting script

Signed-off-by: Danila Fedorin <danila.fedorin@gmail.com>
This commit is contained in:
Danila Fedorin 2025-02-23 13:16:33 -08:00
parent fb46142e9d
commit a8c2b1d05a

View File

@ -16,8 +16,7 @@ def get_used_characters(files):
"""Collect unique characters from all .html files in the given directory."""
char_set = set()
for file in files:
full_path = os.path.join(root, file)
text = extract_text_from_html(full_path)
text = extract_text_from_html(file)
char_set.update(text)
return "".join(sorted(char_set))