Open udomai opened 3 years ago
Only 2 little things had to be fixed. Running this change on the test server to test the effect on the diff.
In Premium, there were only 24 problems. The premium test branch is here: https://github.com/languagetooler-gmbh/languagetool-premium/commit/74299403a7ff6e74eee6c61d3aac021f4aa5760b
Testing this, too.
The above PR is missing one thing: Adding words to spelling.txt
that will no longer be in the dictionary after this (e.g. "Manne"). Is there a smart way of checking for which of the new entries in removed.txt
that is necessary?
I think I have a solution. Trying now.
I wrote a python script that counted how many times the entries in the above list are in the German dictionary. I converted those who only had one entry (i.e. the one we removed) into a list that I can now put into spelling.txt.
old-dat-forms-for-spelling.txt
This all helps improve Suggestions in rules like DE_AGREEMENT, see https://github.com/languagetool-org/languagetool/issues/5470
We have 2,996 obsolete forms for
SUB:DAT:SIN:MAS
in the dictionary. They're a problem especially when we want to synthesize that form and have to suggest the antiquated form as well.starkeSUBSINMAS.txt
I'll test this in a branch to see how many rules will have to be changed.