Unable to fetch interactive avatar speech token

[yuting1008]

I am reproducing interactive avatar. After I deployed it on the Azure Static Web App and click login, the following error message occurred. FYR, the local.settings.json has already been uploaded and the environment variables are correct.

[corticalstack]

Same here, 404 on get-speech-token after deploying static web app.

[corticalstack]

@yuting1008 I noticed there is no azure function as defined by repo artefact function_app.py that is deployed into my resource group. Unfmailiar with azure static web apps, I don't know whether my environment setup meant the function was not deployed when I deployed the static web app, or there's a missing step in the setup guide. Do you see this function deployed?

[corticalstack]

@yuting1008 fyi I managed to move beyond the /api 404 issue. I noticed that the api function_app.py had dependencies likes openai. However, in the repo requirements.txt, it only has:

azure-functions
azurefunctions-extensions-http-fastapi

I added:

azure-functions
openai
fastapi
aiohttp
pyodbc
requests
beautifulsoup4

and now the api route, for example get-speech-token, can be consumed. This is confirmed when testing locally with:

swa start src --api-location api --verbose

Unfortunately I've hit a next issue with get-speech-token with a 500 internal server error (have double-checked region and access key).

[yuting1008]

Hi @corticalstack , thank you for reply! However, I still cannot get speech token after updating requirements.txt.

[corticalstack]

@yuting1008 I believe MSFT are working on fixing an issue introduced by the latest commit. This is hoped to be pushed sometime next week.

In the meantime, I also understand that the version 2 commits back works. However, I think this is without the bing function tool.