slack2discord

A Discord client that imports Slack-exported JSON chat history to Discord channel(s).

tl;dr

Donations

If you find this software useful and would like to make a financial contribution, you can make a donation via either:

It would be most appreciated, but you are under no obligation to do so.

History

This started out as thomasloupe/Slackord. I made some contributions (see #7 and #9) to add support for threads. But then my list of additional proposed changes was significant enough, that we mutually decided that me continuing development on a hard fork would be better.

By now the code has changed enough, and even uses a totally different approach for communicating with Discord, that there's probably very little of the original codebase remaining. Nevertheless, I owe inspiration to the original, and it helped me significantly in getting started on this effort.

Note that there also exists a .NET version thomasloupe/Slackord2 by the original author, that contains additional functionality and appears to be more actively maintained than the upstream fork from which this Python project originates.

Prereqs

py3

This assumes Python 3.x. (Actually, currently 3.9+, see mypy notes below.) Python 2.x was EOL'd at the beginning of 2020, and no new project should be using it.

virtualenv

Install the required packages into a Python virtualenv via:

pip install -r requirements.txt

For help creating virtual environments, see the venv docs. If you use Python a lot, you may also want to consider virtualenvwrapper. If you don't want to think much about virtual envs and just want simple Python scripts to work, you could consider pyv.

Usage

Slack export

To export your Slack data as JSON files, see the article at https://slack.com/help/articles/201658943-Export-your-workspace-data

Note that only workspace owners/admins and org owners/admins can use this feature. While this is available on all plans, only public channels are included if you have the Free or Pro version. You need a Business+ or Enterprise Grid plan to export private channels and direct messages (DMs).

I also think (but am not 100% positive) that the export has the same 90 day limit of history imposed on Free plans as of 1 September 2022. (This change was what motivated me to migrate from Slack to Discord and work on this tool.)

If it is a large export, this might take a while. Slack will notify you when it is done.

The export is in the form of a zip file. Download it, and unzip it.

The contents include dirs at the top level, one for each channel. Within each dir is one or more JSON files, of the form YYYY-MM-DD.json, one for each day in which there are messages for that channel. Like so:

channel1
 |- 2022-01-01.json
 |- 2022-01-02.json
 |- ...
channel2
 |- 2022-01-01.json
 |- 2022-01-02.json
 |- ...
...

There is additionally some metadata contained in JSON files at the top level, but they are not (currently) used by this script, and are not shown above.

Discord import

One time setup

For complete instructions, see https://discordpy.readthedocs.io/en/stable/discord.html

Login to the Discord website and go to the Applications page:

https://discordapp.com/developers/applications/
Create a new application. See the implications below (in creating a bot) of choosing an application name that contains the phrase "discord" (like "slack2discord").

New Application -> Name -> Create
Optionally enter a description. For example:

Applications -> Settings -> General Information -> Description

A Discord client that imports Slack-exported JSON chat history to Discord channel(s).

Save Changes
Create a bot

Applications -> Settings -> Bot -> Build-A-Bot -> Add Bot -> Yes, do it!
Unfortunately, if you named your App "slack2discord", Discord disallows the use of the phrase "discord" within a username. So the Username prefix (before the number) will default to "slack2". If you don't like that, you can choose something else:

Applications -> Settings -> Bot -> Build-A-Bot -> Username

like "slack2disc0rd".
Create a token:

Applications -> Settings -> Bot -> Build-A-Bot -> Token -> Reset Token -> Yes, do it!

Copy the token now, as it will not be shown again. This is used below.

(Don't worry if you mess this up, you can always just repeat this step to create a new token.)
Invite the bot to your Discord server:

Applications -> Settings -> OAuth2 -> URL Generator -> Scopes: check "bot"

Bot permissions -> General permissions:

check: "Manage Channels"

Bot permissions -> Text permissions:

additionally check: "Send Messages", "Create Public Threads", "Send Messages in Threads"

This will create a URL that you can use to add the bot to your server.
- Go to Generated URL
- Copy the URL
- Paste into your browser
- Login if requested
- Select your Discord server, and authorize the external application to access your Discord account, confirming the above permissions.
- -> Continue -> Authorize
- Do the Captcha if requested
- Close the browser tab

Discord token

The Discord token (created for your bot above) must be specified in one of the following manners. This is the order that is searched:

On the command line with --token TOKEN
Via the DISCORD_TOKEN env var
Via a .discord_token file placed in the same dir as the slack2discord.py script.

Script invocation

Briefly, the script is executed via:

./slack2discord.py [--token TOKEN] [--server SERVER] [--no-create] \
    [--users-file USERS_FILE] [--downloads-dir DOWNLOADS_DIR] [--ignore-file-not-found] \
    [-v | --verbose] [-n | --dry-run] <src-and-dest-related-options>

The src and dest related options can be specified in one of three different ways:

--src-file SRC_FILE --dest-channel DEST_CHANNEL

This is for importing a single file from a Slack export, that corresponds to a single day of a single channel.
--src-dir SRC_DIR [--dest-channel DEST_CHANNEL]

This is for importing all of the days from a single channel in a Slack export, one file per day.
--src-dirtree SRC_DIRTREE [--channel-file CHANNEL_FILE]

This is for importing all of the days from multiple (potentially all) channels in a Slack export. One dir per channel, and within each channel dir, one file per day.

For more details, and complete descriptions of all command line options, execute:

./slack2discord.py --help

File attachments

As noted at the end of Slack: How to read Slack data exports, files uploaded to Slack are not directly part of a Slack export. Instead, the export contains URLs that point to the locations of the files on Slack servers. These URLs include a token that gives you the ability access those files. These tokens are listed along with the exports on your Slack workspace's export page, which can be found at https://<workspace-name>.slack.com/services/export

When running the script, such files will first be downloaded from Slack to a local dir (which can be controlled with the --downloads-dir option), and then uploaded to Discord. There are a number of reasons why this operation could fail, and the files listed in the Slack export might not be found. These include:

The file was deleted from Slack after the export was performed
The download token associated with that file export was revoked (this can be done via the export page for the Slack workspace)

In these cases (and for any other HTTP errors related to downloading file attachments from Slack), the default behavior is for the script to fail by raising an HTTPError. This allows you to investigate the situation before deciding how to proceed.

For the special case of HTTP Not Found errors, you can override this behavior, and simply log the not found file as a warning, with the command line option --ignore-file-not-found.

Note that if any files are deleted before the export is created, this state will be reflected within the export (the file will have its mode set to tombstone). Any such files will always be logged as warnings and ignored, regardless of whether or not the ignore option is set for the HTTP Not found case.

Internals

The Discord Python API uses asyncio, so there is the potential to speed up the overall execution time by having multiple Discord HTTP API calls execute in parallel. I have intentionally chosen to not do this.

Within a single channel, we want all of the messages in the channel (and all of the messages within a thread) to be posted in order of timestamp, so that is a reason to serialize those.

A better argument could be made for parallelizing posting to multiple channels. I decided that, at least for the time being, it would be far easier to reason about errors (and potentially restart a failed script, although no such restart support is currently included) if only one channel was imported at a time.

Libraries

This code uses the following libraries:

discord.py (docs, pypi, source) (yes, there really is a .py suffix included in the package name)
decorator (docs, pypi, source)
requests (docs, pypi, source)
tqdm (docs, pypi, source)

External docs

Development

If you want to work on development of this library, besides the packages previously installed (see Prereqs above), you should additionally install the required dev packages into your virtualenv:

pip install -r requirements-dev.txt

This will allow you to run automated tests via pytest:

pytest

As well as mypy static typing checks:

mypy slack2discord.py

I'm using the lower case typing notation for collections (which also removes the need for an import), which requires Python 3.9+ . In the future I may switch to using the | operator rather than Union, which would require Python 3.10+ .

Additionally you can run flake8 style checks:

flake8 slack2discord.py slack2discord

You can automatically run all checks with ./check.sh. GitHub is configured (via .github/workflows/check.yaml and GitHub Actions) to automatically run these checks on any PRs, to require the checks to pass before merging to master, and to automatically run these checks for any pushes to master.

Future work

Some items I am considering:

Better error reporting, so that if an entire import is not successful, it is easier to resume in a way as to avoid duplicates.
Add more automated tests
Ways to optimize file downloads:
- Download multiple files asynchronously via using aiohttp
- Stream file downloads in chunks via Response.iter_content

Feel free to open issues in GitHub if there are any other features you would like to see.

richfromm/slack2discord