Skip to content

Latest commit

 

History

History
180 lines (131 loc) · 6.34 KB

How to reconstruct the solution.md

File metadata and controls

180 lines (131 loc) · 6.34 KB

Reconstruct the solution

Prerequisities

Instructions

Auth0

  1. Go to Auth0 website. If you do not have an account, sign up. You can log in using your microsoft account also

  2. Create a Machine To Machine application. Name it technicaldocuindexer.

  3. Click on your newly created application and from the settings subpage copy:

    • Domain

    • Client ID

    • Client Secret

    • Go to APIs subpage and copy API IDENTIFIER

      audience

  4. Go back to settings subpage, scroll down to Allowed Callback URLs and paste there following values:

    https://localhost:44378/callback, https://TechnicalDocuIndexerWeb.azurewebsites.net/callback

  5. In Allowed Logout URLs paste following values:

    https://localhost:44378, https://TechnicalDocuIndexerWeb.azurewebsites.net

  6. Go to Rules subtab and Create new Rule.

  7. In new rule paste following code and save it.

    • function (user, context, callback) { const assignedRoles = (context.authorization || {}).roles; const idTokenClaims = context.idToken || {};

      idTokenClaims['https://schemas.quickstarts.com/roles'] = assignedRoles;

      callback(null, user, context); }

  8. In Users & Roles subtab, go to Roles:

    • Create Role FileUploader and SearchReader

  9. Go to Users and create 2 new users:

    • searchreader@
    • fileuploader@

  10. After creating both users, Click on their details, go to subtab Roles and add:

    • SearchReader for searchreader@ user
    • FileUploader for fileuploader@ user

  11. Go to APIs subtab on the left and select your technicaldocuindexer API.

  12. In settings tab, make sure you have enabled RBAC Settings and Add Permissions in the Access Token.

Azure portal

  1. Log in to Azure portal and create a new resource group.
  2. Create Function App resource "TDICustomSkills" , where we will store our serverless functions (custom skills).
  3. Create an Azure Cognitive Search Resource and name it cogni-technical-docu-indexer01.
  4. In newly created resource, create a new index.
  5. Copy from newly created search service:
    • From Keys section, Primary admin key
    • Search Service name, which should be cogsearchtechnicaldocuindexer01
    • Search index name azureblob-index
  6. Copy from Cognitive Services resource:
    • From keys and endpoint copy KEY 1

Azure DevOps

  1. Go to Azure DevOps portal, log in using your Microsoft account and create a new project.
  2. Decide whether your code will be hosted in Github or Azure Repos - we assume the code will be in Azure Repos.
  3. Click on Repos tab and create a new repository.
  4. In newly created repository cloned Azure Knowledge mining repository and push it to master branch.
  5. Next, go to Pipelines and create a new variable group tdi-base in your repository.
  6. Fill up variables:
    • Auth0Settings.Audience - API Identifier
    • Auth0Settings.ClientId
    • Auth0Settings.ClientSecret
    • Auth0Settings.Domain
    • Resource.Group.Name
    • Search.APIKey
    • Search.IndexName
    • Search.QueryKey
    • Search.Service - search service name
  7. Save them.
  8. Go to Pipelines subtab in Azure Pipelines.
  9. Create a new pipeline, with code in Azure Repos Git. It should auto-select an existing azure-pipelines.yml file from your repo.
  10. Submit creation of new pipeline and run it.
  11. Pipeline should have created all necessary resources for the application to work.

Deploy Azure Functions

  1. In your local repository, launch CustomFunctions solution.
  2. Right click on the project, and Publish the code to the cloud, to TDICustomSkills FunctionApp.
  3. Repeat the proces for BingFunction solution.

Add skillset to Search Service

  1. In your cogsearchtechnicaldocuindexer01 click on skillsets and create a new skillset.

  2. In skill definition Templates, select Custom Web API Skill - Azure Functions

  • For Template recognizing function paste

     "inputs": [
        {
          "name": "extension",
          "source": "/document/metadata_storage_file_extension"
        },
        {
          "name": "content",
          "source": "/document/content"
        }
      ],
      "outputs": [
        {
          "name": "foundServices",
          "targetName": "foundServices"
        }

  • For word count paste:

    "inputs": [
        {
          "name": "extension",
          "source": "/document/metadata_storage_file_extension"
        },
        {
          "name": "content",
          "source": "/document/content"
        }
      ],
      "outputs": [
        {
          "name": "wordCount",
          "targetName": "wordCount"
        },
        {
          "name": "timeToRead",
          "targetName": "timeToRead"
        }
      ]
  1. Save new skillset.

Upload Files for Indexer

  1. Go to your resource group.
  2. Click on sttechnicaldocuindexer Storage Account.
  3. In Containers subtab, create a new container initial-load and paste there files you would like to be searched upon (pdf, json, markdown files, png).
  4. Go to your cogsearchtechnicaldocuindexer01 and create new Indexer in Indexer tab.
  5. Fill out neccesary parameters. Choose an existing connection for Storage Connection string and pinpoint initial-load container.

Create Index

  1. Go to Indexes tab and create a new Index.
  2. After creating new Index, remember to click on it and set CORS allowed origin type to Custom and set the URL of the webapp.

Navigate to https://technicaldocuindexerweb.azurewebsites.net, login using searchreader user and test your app!