Currently finders use two strings. One contains a literal example of a header and the other contains special characters that define how the header can vary. It works well for UK Teletext/Oracle/Ceefax but has problems with other broadcasts from countries:
No way to define numeric month/year dates.
Text patterns expect the English spellings.
The system is quite confusing to use.
One possible way to improve the finders would be to use regular expressions. This would allow any arbitrary type of header to be matched. However the finder needs to do fuzzy matching with ranking based on the type of match. Also, regular expressions would probably be even harder to understand - although they are at least a standard rather than something I have made up.
https://github.com/ali1234/vhs-teletext/blob/088da3911d610910fbe9b383d9382b4c602c1617/teletext/finders.py#L144
Currently finders use two strings. One contains a literal example of a header and the other contains special characters that define how the header can vary. It works well for UK Teletext/Oracle/Ceefax but has problems with other broadcasts from countries:
One possible way to improve the finders would be to use regular expressions. This would allow any arbitrary type of header to be matched. However the finder needs to do fuzzy matching with ranking based on the type of match. Also, regular expressions would probably be even harder to understand - although they are at least a standard rather than something I have made up.