SAVE THE DATE! Next conf call (with test of full integration) on the 9th of March (4pm)

DATE

February 16, 2018

TIME

4pm

LOGISTICS

Skype

AGENDA

Follow-up on barcamp, starting to prepare for the demo

PARTICIPANTS

Sarah Zribi, Yohan Houpert, Sammy Benhamich (LINAGORA), Guokan Shang (LIX/LINAGORA), Pol Meladianos, Xristos Xypolopoulos, Antoine Tixier (LIX)

MINUTES

Yohan, Sammy, OpenPaaS Linagora team:

1) Is the SSL working with the self-signed certificate? If not, what is the solution you propose?

2) this is the link to the Cryptpad version that we need to be deployed on the OpenPaaS platform for the demo https://cryptpad.fr/code/#/1/edit/1Hv3iRVFrvItMvKVgEOb5Q/T2ir4Pwo4ITupXBwjtMpsTn1/

Two warnings: if https finally works, this will have to be changed in the file above. Same if we change the IP of our server.

Note:
- to send us the text from Hublin chat, you can use the same API as for the speech-to-text output, and put the text in the same JSON file

Abdel (Yohan/Sammy) :

- 3) is the integration done for the 2 speech-to-text services: the fast one called in real-time and the more accurate but slower one called at the end of the meeting?

- 4) we need the output of the speech recognition to follow the format that was decided when we did the initial integration last year. The current format you are using (below) does not comply:

Device1-U1-T1-Time_Begin-Time_END Transcription1
Device1-U3-T1-Time_Begin-Time_END Transcription2

We need a tab separated string containing the fields 'from' 'until' 'speaker' 'text'. So, for the first line of output above, that would give Time_Begin tab Time_END tab Device1-U1-T1 tab Transcription1.

- 5) Abdel, is there a domain (politics, business, technical?) on which your speech-to-text works best? (in order to prepare the text that will be read by the 5 users, and pasted into