Twitter Bot Service with Raindrop Integration

Service to automatically tweet selection of raindrop.io items at regular intervals via the @lcptuk twitter account.

How to configure

The following features are customisable by changing their values in the config file:

• running: a boolean, true or false, turns the bot on and off
• interval: an integer, the amount of minutes between posting tweets. eg. 60. Minimum is every 5 minutes; Bot won't take into account a tweeting interval shorter than the polling interval (5min)
• raindrop_tag: a string, only library items with this tag will be selected in the Raindrop collection. eg. 'post-to-twitter'
• tweet_format: a string, the content of the tweet with placeholders with the library item attributes.
• default_image_url: a string, the link to the image to display with the tweet if the library item doesn't have a corresponding image. The image will be reframed by twitter. If the image url doesn't work, the bot will gracefully fall back onto not displaying an image with the tweet.

The bot polls the config to check for changes every 5 minutes. Therefore there may be up to 5 minutes before any change is implemented. If the tweeting interval has changed to X min and there are <X min since the last tweet before the config was changed, the next tweet will be X min after the latest tweet.

Tweet format

The tweet will always display the raindrop cover image of the raindrop library item when possible.

Available library item attributes are:

• {title}
• {url}
• {description}
• {tags} that will show as #tag1 #tag2 #tag3

The most up-to-date available library item attributes can be seen in the content dictionary

The tweet_format string follows the rules of python3 string formatting: https://pyformat.info/.

A tweet_format example is '{title:.100}\n\n{url}' where the title is truncated after 100 characters.

⚠️ A tweet won't be posted if it is over 280 characters.

How to run

The following instructions require knowing how to type commands into the terminal.

It assumes Python version >3.8 is already installed. You can check with:

$ python --version

Locally

You may want to run this bot locally to test it, or if you don't need the features of a dedicated production server.

1. Import the project:

$ git clone [email protected]:nwspk/lcpt_twitter_bot.git
$ cd lcpt_twitter_bot

2. Create a credentials.sh file with the twitter and raindrop authentication required inside the lcpt_twitter_bot folder.

$ touch ./credentials.sh

The credential.sh file should be laid out like this:

export RAINDROP_CLIENT_ID=<paste here>
export RAINDROP_CLIENT_SECRET=<paste here>
export RAINDROP_TOKEN=<paste here>

export TWITTER_KEY=<paste here> # twitter consumer key
export TWITTER_SECRET=<paste here> # twitter consumer secret
export TWITTER_ACCESS_TOKEN=<paste here> # twitter access token key
export TWITTER_ACCESS_SECRET=<paste here> # twitter access token secret

3. Install required external libraries.

$ pip3 install -r ./requirements.txt

4. Start the bot.

$ python3 ./run.py

Stop the bot by pressing Ctrl-c.

5. To run locally in the background (only works as long as you don't turn your computer off):

Check screen is installed:

$ screen --version

If it returns Screen version <etc> you're good to go.

$ screen -S bot
$ python3 ./run.py

Wait a few seconds to check it's running fine. You shouldn't see any Errors returned.

Press Ctrl-a then d to let the bot run in the background.

The bot will be stopped automatically if its process is killed, if you turn your computer off for example, and it won't start again until you follow the above step 5 again. One way to stop the bot while your computer is still running is to type:

$ screen -r bot

Then press Ctrl-c to stop the bot, and press Ctrl-a then b to return to normal.

Remotely on a server

These are instructions on how to admin the already running bot instance that is currently hosted on DigitalOcean.

1. Find the remote host (=server) IP address:

Sign in to your Digital Ocean account and under Projects click on raindrop twitter bot. Under Resources, you should see DROPLETS (1) with the IP address. Keep track of the IP address.

2. Open a terminal and connect to the remote server:

$ ssh webmaster@<ip address>

Enter the password when asked.

3. The bot is running in a screen.

$ screen -r bot