The debug container simulates a stream of incoming textual data. It is similar to using a microphone, but relies on transcribed data. To choose a different dataset, change settings (CHOOSE_DEBUG_DATA) in the main project directory. Please put your own data in the folder 00_custom/debug_data. To be used succesfully by the framework, the data must be a csv-file with 3 variables: id_speaker;id_time;text. The formatting is as follows:
For demonstration purposes, we provide two files:
Streaming interval can be set in settings.txt - use DEBUG_TIME_DELAY variable. Value indicates the seconds for pushing new turns. The value should be greater than 10, since most functions run only every 10 seconds. We have made good experiences with a value of 25. If left blank, the timestamp in the debug data (variable id_time) is used to push turns, mirroring incoming turns in real-time.
All functions are based on R.
Felix Süttmann, Valentin Gold