Cognitive Asset manager - backend - Project 82038

Manage assets and models built from assets for Cognitive Services

Post-hackathon update

This project has been separated out into a streaming upload server and triggered Azure functions. The streaming server has been moved to a separate repo.

Getting started

1. Fork and clone project

2. Copy env.sample and rename to .env.

3. Create Azure resources

   1. Speech
   2. Translator
   3. Storage

4. Fill in values in .env:

   Variable	Purpose
   SPEECHRESOURCETOKENHOST	westus.api.cognitive.microsoft.com - Token server. TBD - fix when fixing custom endpoints across Cognitive Services.
   SPEECHRESOURCETTSHOST	westus.tts.speech.microsoft.com - TBD - fix when fixing custom endpoints across Cognitive Services.
   SPEECHKEY	Cognitive Services Speech key
   DFBAPISERVERVER	0.1 - versioning of node.js server app
   DFBAPISERVERPORT	3005 - Local node.js server port.
   DFBAPIUPLOADSERVERDIR	upload - Local directory where files are uploaded to.
   DFBAPIDOWNLOADSERVERDIR	out - Local directory to store downloadable files. TBD - Move to Azure files.
   DFBAPIDOWNLOADSERVERURI	localhost - Host used for download URIs.
   DFBAPIDOWNLOADSERVERPORT	3010 - Port used for download URIs. Create MP3 from text then return download URI.
   TRANSLATORKEY	Cognitive Services Translator key, currently hardcoded to westus region.
   AZSTORAGECONNECTIONSTRING	Azure storage connection string
   AZSTORAGECONTAINER	containercogservblob Your custom name for container in Azure Storage
   SECRET	String used to create user authentication token

5. npm start to run server or npm test to run Jest tests.

6. Deploy to Azure web app from local Git. Instructions to redeploy from CLI.

Main URL

Deploy to Azure

Following instructions.

Create user

Do this once, at the beginning, for each local computer you develop with.

1. Create user for local git:

   az webapp deployment source config-local-git --name <app-name> --resource-group <group-name>
   

2. Get credentials

   az webapp deployment list-publishing-credentials --name <app-name> --resource-group <group-name> --query scmUri --output tsv
   

3. Use URL from response, create remote link for git

   git remote add azure <url>
   

Deploy project to azure via local Git command

1. Deploy to azure.

   git push azure master