The parser splits name formats on the ASCII comma only (parser.py, self._full_name.split(",")), so the "Lastname, Firstname" format is never detected when the input uses the Arabic comma ، (U+060C) — the standard comma in Arabic, Persian, and Urdu text. The character survives into the parsed output glued to a name piece.
>>> from nameparser import HumanName
>>> HumanName("سلمان، محمد") # Arabic comma: expect first='محمد' last='سلمان'
<HumanName : [
title: ''
first: 'سلمان،'
middle: ''
last: 'محمد'
...
]>
>>> HumanName("سلمان, محمد") # same input with ASCII comma parses correctly
<HumanName : [
title: ''
first: 'محمد'
middle: ''
last: 'سلمان'
...
]>Two effects: the lastname-comma format detection fails (first/last are swapped), and the comma character pollutes the output value.
Possible approach: normalize comma variants to , during pre-processing, or split on a small set of comma characters. If treated as a set, it could live in config alongside the other constants; the fullwidth comma ， (U+FF0C, used in CJK text) would be a natural sibling to include.
Observed on v1.3.0.
The parser splits name formats on the ASCII comma only (
parser.py,self._full_name.split(",")), so the "Lastname, Firstname" format is never detected when the input uses the Arabic comma،(U+060C) — the standard comma in Arabic, Persian, and Urdu text. The character survives into the parsed output glued to a name piece.Two effects: the lastname-comma format detection fails (first/last are swapped), and the comma character pollutes the output value.
Possible approach: normalize comma variants to
,during pre-processing, or split on a small set of comma characters. If treated as a set, it could live in config alongside the other constants; the fullwidth comma，(U+FF0C, used in CJK text) would be a natural sibling to include.Observed on v1.3.0.