An issue previously alluded to in #8 , cropping up again. As the transliteration convention of each line is detected independently, some ITRANS lines without any distinctive ITRANS characters can get detected as HK, screwing up the metre (e.g. "sh" is two consonants in HK and one in ITRANS).
Previously dealt with "Sh" as a special case and can do the same with "sh" (does स्ह occur in Sanskrit?), but should probably switch to detecting the whole input at once.
An issue previously alluded to in #8 , cropping up again. As the transliteration convention of each line is detected independently, some ITRANS lines without any distinctive ITRANS characters can get detected as HK, screwing up the metre (e.g. "sh" is two consonants in HK and one in ITRANS).
Previously dealt with "Sh" as a special case and can do the same with "sh" (does स्ह occur in Sanskrit?), but should probably switch to detecting the whole input at once.