To give the possibility to split sentences by other characters, not only full stops.

Anonymous Fox 2 feb 2024 💡 Planificado

German uses special characters as quotation marks (i.e. »…« or „…“). When we split long paragraphs to smaller sentences (Automatic Translation: ON, Original Formatting: OFF) each sentence ends at a full stop. German dialogs, where each quotation is ended by « or “ and not by a full stop the end of each line is not correctly recognized, and the whole exchange is seen as a single long sentence (see attachment). It should be possible to choose different "sentence terminators" according to the language. In case of German, a sentence should terminate by full stop, « or “

Comentarios

nhat

The same for me in Chineses, Japaneses, one question happend in my mind is "How the app devines a long paragraphs in to unit to speak tts. Where to customize myself that rule?". Because I notice some Chineses units to speak/translate have 2 sentences (has 2 。characters) and some units have 1 characters. And I suggest the app should be able to import custom scripts for devining. For example loop over paragraphs and do some check conditions (current languages, text content ....)

17 feb 2024

KursX

If you are loading a text file or creating a book, you can specify any characters to split using regular expressions

23 feb 2024

oliver w

In smart book system, when it addresses the long English paragraph, it will show just one sentence no matter how many small sentences in it. But it's really hinder the English learner to study and understand the article quickly. Would it be possible that you could optimize the logic of the splitting sentence? when splitting the sentence, not only by the full stop in the paragraph but also by semicolon, bullet character? I suggest that you can add some more options in the software system such as "split sentences by full stop in the paragraph; split sentences by semicolon in the paragraph; split sentences by bulletcharacter in the paragraph; etc." or even the combination of the options,so the user can choose the appropriate methods in learning. in a nutshell, I hope it can make the long paragraph splitting into shorter sentences when shows original paragraph and translation at the same time.in my opinion, it's really helpful for using bilingual reading material to learn English if you can grasp the specfic meaning the material quickly.

25 oct 2024

Maxim Miller

In Hebrew religious books (Torah, Tanakh, Siddur, Psalms) - there's a different symbol used to separate sentences. At the end of each sentence they use a special "sof pasuk" character instead of a dot (https://en.wikipedia.org/wiki/Sof_passuk). The "Sof pasuk" looks like a colon (:) (U+003A) but the two dots are of diamond shape (׃) (U+05C3), or in some fonts of square shape. You can see the difference if you paste it to MS Word and increase the size of the font to very big. Example of a text https://he.wikisource.org/wiki/%D7%A1%D7%99%D7%93%D7%95%D7%A8/%D7%A0%D7%95%D7%A1%D7%97_%D7%AA%D7%99%D7%9E%D7%9F_%D7%91%D7%9C%D7%93%D7%99/%D7%90%D7%A9%D7%9E%D7%95%D7%A8%D7%95%D7%AA, epub can be downloaded here https://ws-export.wmcloud.org/?format=epub&lang=he&page=%D7%A1%D7%99%D7%93%D7%95%D7%A8%2F%D7%A0%D7%95%D7%A1%D7%97_%D7%A2%D7%93%D7%95%D7%AA_%D7%94%D7%9E%D7%96%D7%A8%D7%97%2F%D7%A1%D7%9C%D7%99%D7%97%D7%95%D7%AA_%D7%9C%D7%99%D7%9E%D7%99_%D7%94%D7%AA%D7%A9%D7%95%D7%91%D7%94. This "Sof pasuk" symbol is not used for anything else, and can be safely added to the regexp for splitting the sentences. It's also not used in other languages. So if you add this symbol to the list of sentence separators, it will work correctly.

152849-176086a2cd584e07b2d98141d2937c3e_Sof-Pasuk-sentence-parsing.jpg

10 sept 2026

Maxim Miller

I also confirm that long paragraphs are a real problem, and need better handling. I made several screenshots, maybe not of such long paragraphs, but those that for some reason include multiple sentences, or can be split to 2-3 parts with no problem at all. Maybe adding a logic that if there's 2 tiny sentences, they can be merged no problem, but if it's a large paragraph - I don't get it why it needs to stay glued together if there's an easy way to split it - quotes, parentheses, semicolon. It will never be perfect, but making it more customizable would help. Or just to make the logic more aware of the length of sentences, and if the length of the sentence is too long - split it more aggressively.

10 sept 2026

Crear ticket