LanguageTool 5.1 ignores random paragraphs about tex2txt

[matzex]

(Possibly, this has been solved with EDIT. Clarified Point 1 with the found problems from LanguageTool?)

No, it has not solved. I just changed my point one and added the part in the brackets "(in the previously missing paragraphs)". With --context -1 all paragraphs are shown and without they are not visible.

I will try to construct an example in the next days and add it to this issue.

[matze-dd]

Thank you for reporting this!

I see two possibilities to approach the reason of the problem.

1. If you specify option --context with a negative argument, do you then see all of the text in the HTML report? If yes, LanguageTool very probablty does not detect mistakes in the text parts not displayed before. If not, then please give a notice.
2. To be on the save side, you could additionally invoke
   python3 tex2txt.py --lang de chapter-4.tex > chapter-4.plain.
   File chapter-4.plain then will contain all the plain text that is sent to LanguageTool by shell.py. Please give me a notice, if something is missing.

From my own experience with LanguageTool I can say that the newer versions sometimes report less problems, since refined rules are better in avoiding false positives.

A final question: Did you already consider switching to YaLafi? This is the successor of Tex2txt that is not devloped anymore. You then would write

python3 -m yalafi.shell --output html --language de-DE chapter-4.tex > report-4test.html

EDIT. Corrected wrong option --html for yalafi.shell.

[matzex]

Thank you very much for your fast and helpful answer!

1. Yes, with --context -1 the file includes all missing paragraphs and identified also more problems (in the previously missing paragraphs) than before.
2. Yes with the text2text.py command and plain text file all missing paragraphs are also included.

I completely missed the YaLafi successor and testet it shortly. Also with YaLafi (using the shell script and changed the path to LanguageTool 5.1) the paragraphs are missing. So it looks like this is unrelated to tex2txt and YaLafi, don't know what the problem here is exactly. I will test this with LanguageTool 5.0 to see if this has the same problem.

EDIT. Clarified Point 1 with the found problems from LanguageTool

[matze-dd]

Thank you very much for your fast and helpful answer!

You are welcome!

Yes, with --context -1 the file includes all missing paragraphs and identified also more problems (in the previously missing paragraphs) than before.

This sounds like a severe bug. With the default value of '--context 2', the HTML report should include all text parts containing a message from LanguageTool, and this should not depend on the LanguageTool version. Could you please try to construct an example text that demonstrates the issue, if it persists?

(Possibly, this has been solved with EDIT. Clarified Point 1 with the found problems from LanguageTool?)

[matze-dd]

@matzex, could you please give a notice whether the problem persists?

Thank you in advance!