deeplcom / deepl-java Goto Github PK

Hello,
Thanks for the 0.2.1 version.
I'm integrating your library 0.2.1 in a maven project and here is what I notice :

• the you cite in the README works but does not automatically include a dependency to gson, which needs to be included by hand with another .
• deepl-java-0.2.1 doesn't work with last version of gson (2.9.1). I make it work with gson 2.8.5, already used by deepl-java-0.2.0
  Thierry

Hi,

in the README.md of https://github.com/DeepLcom/deepl-java and the class header javadoc of Translator I could not find directly an information whether a Translator instance is thread-safe or not and can be shared between different threads. Would it be possible to add a note for that?

Thanks and best regards,
Philipp Förmer

Hello, I've been using this library for a while now, and I noticed that in my shadowed jar, there were assets for the library org.apache.commons:commons-math3. Out of curiosity, I checked the source code to see how it is used, but nowhere in the code the library is actually used.
Is it possible to remove it from the dependencies ?

Ahoy! 👋

As the topic says, use Dependabot for automation. Some dependencies are 2-3 years out of date. Let me know if you find such PR useful.

Hi people,

I've played a bit with the library and found a couple of inconveniences.

The obvious - if you pass a big text to translate, it is impossible to guess when you will get an error response that the size of the request is over the limit. I know it is 128k but the headers and other data are probably included in the calculation. As the client, I need help determining when to split the data.

There seems to be no implementation of glossary management, is there any plan to implement it in the future?
Or is it possible to make a contribution?

openapi has it.

As per-title, the only place reading the SentenceSplittingMode ignores the All value, which is useful when setting the tag handing to html.

Maybe it's because the default has changed since november 2022 as stated in https://www.deepl.com/docs-api/translate-text/translate-text/:

The use of nonewlines as the default value for text translations where tag_handling=html is new behavior that was implemented in November 2022, when HTML handling was moved out of beta.

I have tried to request a translation through a curl request for the same input and the result is just fine.

The latest version available on Maven is 1.1.0

It is probably nothing to the java library but a general DeepL translation question.

At the weekend, I tried to translate a bunch of strings and composed one big text to make fewer requests. The format of the string was:

<hash of the string to translate>######<text to tranlsate>^^^^^^^
......

That worked for some strings. However, DeepL returned a different integer for some hashes or even appended some strings.
That, of course, broke the whole algorithm.

What do you think about how to handle such a translation? Would DeepL ignore an integer from translation if it is part of XML, for example

<tr hash="!hash of string to translate!">!string for the translation!</tr>
...

Do you have any recommendations or thoughts?