Mastodon experiment

Install

yarn install
npm link eventlog-server
npm link wikijs-cli

We assume a working installation of mastodon-bot is available at the path ../mastodon-bot.

Config

Set the mastodon instances and mastodon account identifiers to track in ./config/accounts.txt.

Format:

<mastodon host>;<account id>

The account id for a user can be found with the npx mastodon-cli account command:

$ npx mastodon-cli account --url https://mathstodon.xyz hochstenbach | jq '.id'
112881408245106756

Set a demonstrator researcher profile page

export DEMO_MODE=NO_TOOTS
export DEMO_PROFILE=https://wiki.mycontributions.info/en/researcher/orcid/0000-0000-0000-0010

Optional preparation

Create a backup of the cache database

npx eventlog-server export > ./export-2024-11-04.json

Import a cache database export

npx eventlog-server remove-all
npx eventlog-server import ./export-2024-11-04.json

Start the daemons

pm2 start ecosystem.config.js

Fetch toots for the configured accounts

Fetch some toots for the configured accounts. By default, 20 toots will be gathered per account. Using a MASTODON_HANDLER configuration handler we create and count only the toots that reference at least one URL.

The default account file to be processed is ./config/account.txt.

The values in brackets [...] show optional arguments with their default values.

yarn fetch-all [max-toots-per-account:-20] [account_file:-./config/accounts.txt]

Run the experiment

Running the experiment means moving all accepted toots to the mastodon bot environment and starting all handlers to find metadata for the URLs mentioned in the toots and updating the Wiki.js page.

yarn run-experiment

We can also test the workflow moving only one file to the mastodon bot environment.

yarn run-one

Monitor the daemons

pm2 ps

pm2 logs exp-accepted

Statistics

./bin/stats.sh

Stop the daemons

pm2 stop ecosystem.config.js

Export all experiment data

./bin/end_experiment.sh

Clean all

./bin/clean_experiment.sh