Words marked as "known" in one text aren't also considered known in new texts.

[jzohrab]

I used LingL to process a short text, marking some words as known. I then copied and pasted part of this into a new text, and all of the words in the new text again had status = unknown, even though they should have been marked as known as I'd already seen them in the first text.

Initial text:

New text:

All words are again marked as unknown even though the exact same words were processed in the first text:

Expected behavior

The words that were marked as known in the first text should have again been known in the second text; they're the same word.

Notes:

• This is LingL Version 2021.08.01 (commit 7813d8aaeec5ec65f56 = master, cloned from source)
• running the django server on mac

[jzohrab]

Ouch: it looks like importing new texts just copies all words in there as totally new terms.

sqlite> select text_id, wordtextLC from words where text_id in (1, 5) and wordtextLC = 'futuro' order by wordtextLC;
text_id|wordtextLC
1|futuro
5|futuro

Given that, I don't really follow how this program will let me accumulate old vocab and only deal with new vocab in new texts. :-( Have I misunderstood something, @gustavklopp ?

[njusunyi]

Ouch: it looks like importing new texts just copies all words in there as totally new terms.

sqlite> select text_id, wordtextLC from words where text_id in (1, 5) and wordtextLC = 'futuro' order by wordtextLC;
text_id|wordtextLC
1|futuro
5|futuro

Given that, I don't really follow how this program will let me accumulate old vocab and only deal with new vocab in new texts. :-( Have I misunderstood something, @gustavklopp ?

It seems that it considers a known word in a new sentence as "new". It makes sense as a word can have different meanings in different contexts. But I still think it would be better if it considers a reoccurred word as known/learn by default and let the user decide if it has a new meaning in the current sentence.

[njusunyi]

Ouch: it looks like importing new texts just copies all words in there as totally new terms.

sqlite> select text_id, wordtextLC from words where text_id in (1, 5) and wordtextLC = 'futuro' order by wordtextLC;
text_id|wordtextLC
1|futuro
5|futuro

Given that, I don't really follow how this program will let me accumulate old vocab and only deal with new vocab in new texts. :-( Have I misunderstood something, @gustavklopp ?

It seems that it considers a known word in a new sentence as "new". It makes sense as a word can have different meanings in different contexts. But I still think it would be better if it considers a reoccurred word as known/learn by default and let the user decide if it has a new meaning in the current sentence.

I think the issue is the "I know all" function. Any words marked as known by the "I know all" button are not recognized in new text.