Expected behavior:
When activated by adding API key (here: https://talkserver.com/admin/configure/moderation) Perspective API should flag/moderate/block obviously offense messages e.g. those with straight profanity.
Actual behavior:
Offensive messages are published, no intervention from Perspective.
Ideally there would be some validation on the config screen to show that the API key and service are functional, but there isn't.
Versions:
Hey Matt! Can you write into our support so we can get some more details and help you troubleshoot this? It has to be an issue with your specific setup.
Also let us know what version of Coral you're using.
Hey @mattlinares, I went through a similar problem and solved it by increasing the timeout for the perspective API by changing the value of the environment variable PERSPECTIVE_TIMEOUT to 1500.
Sadly the response time exceeds 1 second for me most of the time.
Most helpful comment
Hey @mattlinares, I went through a similar problem and solved it by increasing the timeout for the perspective API by changing the value of the environment variable PERSPECTIVE_TIMEOUT to 1500.
Sadly the response time exceeds 1 second for me most of the time.