You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
is it possible to use the DKPro-Wrapper as a server (similar to coreNLP) or at least as continously running, expecting text on stdin and producing stdout, without closing. The startup time is a heavy overhead if the pipelline is started for each small snippet of text as it occurs and i would prefer not to collect all input and then treat it in a single file.
thank you!
The text was updated successfully, but these errors were encountered:
it is possible to run the pipeline on a folder of text or xml files, the wrapper will then create a filename.csv for each file in the given output folder. Just pass a directory to the -input option.
There are currently no plans on our side to implement a server mode. (Running a full analysis on texts of notable length takes quite a while, so there's a certain risk of timeouts)
A real streaming mode implementation isn't possible since the document must be loaded completely before further analysis steps begin. It might be possible to implement something that could accept multiple documents on stdin separated by sth like \0 …
is it possible to use the DKPro-Wrapper as a server (similar to coreNLP) or at least as continously running, expecting text on stdin and producing stdout, without closing. The startup time is a heavy overhead if the pipelline is started for each small snippet of text as it occurs and i would prefer not to collect all input and then treat it in a single file.
thank you!
The text was updated successfully, but these errors were encountered: