Discourse Import Command

[zakhap]

Description

How the sausage gets made:

• What should the Discourse Import Script Do?
  • Accounts/Profiles
    • Create Accounts for all users in the Discourse Imports.
    • Create Profiles for each Account imported, copying over the Profile Name + Avatar if it exists. No other identifying informations should ever be made public on the Profile, esp. emails.
    • If “Accounts Claimable” is set to true, Accounts created by the Discourse Import should be claimable by signing into Common with the same email.
      • If “Accounts Claimable” is set to false, Accounts created by the Discourse Import should not transfer over the email (to User Account) nor be claimable.
  • For each Category, Topic, Comment + Reaction from Discourse, the script should make the Common equivalent: Category ⇒ Topic, Topic ⇒ Thread, Comment ⇒ Comment, Reaction ⇒ Reaction.
    • Proper Account+Profile ownership of each created Thread, Comment, and Reaction should be maintained. Going to a Profile created during the import should show all the Threads + Comments and Threads/Comments/Reactions should be properly labeled with Profile information across the Forum.

Project Owner

@zakhap

Acceptance Criteria

• Should successfully perform the above import.

Additional context

Inputs required for this script should come from #7123.

[rbennettcw]

Status update:

• The PR implements a route that successfully takes a Discourse dump file URL (can be s3, etc.) and imports the entities into the specified CW community.
• The last piece of the puzzle is making sure the QA/prod Heroku environments are configured correctly, since the implementation requires the Node app spinning up a psql child process to handle the import. As a binary dependency, there might be some gotchas with the Heroku Node.js runtime.

[rbennettcw]

@jnaviask We need to use the psql binary to import sql files because the dump files contain syntax that’s only supported by the binary psql client (not a JS client). And the only way to access the psql binary in Heroku is via buildpacks– basically Heroku’s version of Docker.

[jnaviask]

@rbennettcw that's interesting -- when i do heroku run bash I can access the psql binary (/usr/bin/psql). Is the app itself running in a different env somehow?

[rbennettcw]

Bumping to 8 since the staleness of the branch broke types, imports and models. Fixing some logic that broke due to the profile model changes.