Stt adapter IBM - andrew d anderson #123
Conversation
Co-Authored-By: AndrewDAnderson <[email protected]>
cleaned up the code making it more in line with the logic in the other adapter, and better setup for accomodating change over time
|
Need to double check the tests are passing, but other then that ready for merge to master. cleaned up the code making it more in line with the logic in the other adapter, and better setup for accommodating changes and adjustment over time. Seems like IBM speaker diarization might actually be quiet good, in this example it recognised it as being just one speaker. So rather then grouping by speaker, which would have left with one giant paragraph and indured draftjs performance, with virtual dom etc.. I left some comments in the code on where that could be changed, eg it could break paragraphs every minute. This also loses some of the clever stuff that was done in the previous PR about adding punctuation to IBM STT which doesn't seem to have it (?). Seems like a common workaround is to add punctuation after every IBM line but will leave that for a later PR. |
was missing speaker info
|
fixed the test, ready for merge
|
There was a problem hiding this comment.
Just had a couple of tiny things but looks good @AndrewDAnderson & @pietrop 👍
Is your Pull Request request related to another issue in this repository ?
from original PR #60 by @AndrewDAnderson
Describe what the PR does
Adds support for IBM Watson STT (cc @seanmtracey)
State whether the PR is ready for review or whether it needs extra work
Updated PR #60 with current master, and doing some final tweaks before merging to master, ahead off reorg (PR #122 - issue #109)
Additional context
NA