Discord Transcript Action

This bot connects to Discord and generates HTML transcripts for the #help channel.

Features

• Generates HTML transcripts using the discord-html-transcripts package
• Saves transcripts to a specified output directory

Inputs

• discord-token: The authentication token of the Discord bot. See below for instructions on how to create a bot and get the token.
• guild-id: The guild ID. You can find this by right-clicking on the server icon and selecting "Copy ID".
• transcript-directory: The directory where the transcripts are/will be saved.

If an existing directory is specified, then the action will only fetch threads that have been modified or don't exist yet.

Example

env:
  # Don't forget to put the guild ID in quotes.
  GUILD_ID: '123456789012345678'
---
- name: 'Run the transcript action'
  uses: toitlang/[email protected]
  with:
    discord-token: ${{ secrets.DISCORD_TOKEN }}
    guild-id: { { env.GUILD_ID } }
    transcript-directory: transcripts

Typically, this step is followed by a step that commits the output, and one that uploads the transcript to the gh-pages branch of the repository.

- name: 'Commit the transcripts'
  run: |
    git config --global user.email "github-actions[bot]@users.noreply.github.com"
    git config --global user.name "github-actions[bot]"
    git add transcripts
    git commit -m "Update transcripts"
    git push

- name: 'Upload to gh-pages'
  uses: peaceiris/actions-gh-pages@v4
  with:
    github_token: ${{ secrets.GITHUB_TOKEN }}
    publish_dir: transcripts
    cname: 'example.com'

Output

The transcript directory contains an HMTL file for each thread in the #help forum.

In addition, the output directory contains a index.html file that links to all the transcripts.

Finally, it also produces an index.json file that contains the metadata for each transcript. This can be used to generate a more complex index page.

Here is a screenshot of the generated index page for the Toit Discord page, served at the Toit help page.

Incremental updates

If the transcript directory already exists and contains the index.json file, then the action will only fetch threads that have been modified or don't exist yet.

Run locally

You can also just run the transcript generation locally.

Make sure to run with Node 20. Use, for example, nvm to install it. The repository contains a .nvmrc file that specifies the node version that works. It also contains a .node-version file that is used by GitHub to fetch the correct version.

If you have nvm installed, but not automatically activated in your .bashrc, you will need to do

source /usr/share/nvm/init-nvm.sh

Then run nvm install to install the correct version of node.

Install the dependencies.

npm install

Save the Discord credentials and your input parameters to a .env file. Use the .env.example as a starting point.

INPUT_discord-token=YOUR_DISCORD_TOKEN
INPUT_guild-id=YOUR_GUILD_ID
INPUT_transcript-directory=transcripts

Run the script.

npm run local-action

Discord bot

The following instructions mirror the ones provided on Discord's official quick-start guide

To create a Discord bot, follow these steps:

• Create a new application
• In the 'Installation' tab:
  • Unselect 'User Install'.
  • Switch the Install link to 'None'.
  • Save the changes.
• In the 'Bot' tab:
  • click on 'Reset Token'. This will generate a new token. You will need this token as input to the action. Typically it is saved as a GitHub Action secret.
  • Disable 'Public bot'.
  • Enable 'Message content intent'.
  • Save the changes
• In the 'OAuth2' tab:
  • In the scopes select 'bot'. This opens up the bot permissions.
  • In the bot permissions, select 'View Channels' and 'Read Message History'.
  • Copy the URL and paste it in your browser. This will open a page that allows you to add the bot to a server. Select the server where you want to run the action.

You can uninstall a bot by kicking it from the server.

References

This repository is based on the TypeScript Action template.