The bot parses RSS feeds of scientific journals, downloads new publications in HTML/XML and then builds a word cloud image of the extracted text.
https://twitter.com/AtmosSciBot
A word cloud can be created if AtmosSciBot is mentioned in a tweet and a URL to the HTML or XML page of an open-access publication is included. Note that the URL should directly lead to the HTML page (NOT a PDF) with the full text of the publication and NOT an abstract (so don't request EGU articles that are still in discussion state)!
The tweet has to contain a short name of the corresponding journal as a hashtag. Only the journals that are in the list of journals are allowed: otherwise the bot would not know how to extract the text from HTML. The tweet also has to contain the following words in any order: "make", "word", "cloud", "please".
Example:
@AtmosSciBot make wordcloud please #JGRA http://onlinelibrary.wiley.com/doi/10.1002/2015JD024680/full
Note only open-access articles can be processed, otherwise the bot will reply that it is unable to retrieve text.
Note you may have to wait for a couple of hours until the request is processed.
A on-demand generation request can contain a name of Google Font (https://fonts.google.com) and the wordcloud will use the chosen font if the request is correct.
To choose font, include [font=<name of the font>] in the tweet. The word cloud will use the "Regular" (400) style of the font.
This option is made possible thanks to this google-font-download script.
Example:
@AtmosSciBot make wordcloud please [font=Raleway] #QJRMS https://rmets.onlinelibrary.wiley.com/doi/full/10.1002/qj.2911
or (note the spaces)
@AtmosSciBot make wordcloud please [font=M PLUS 1p] #QJRMS https://rmets.onlinelibrary.wiley.com/doi/full/10.1002/qj.2911
If something is wrong with the font-related request, the wordcloud is created using the default font (https://fonts.google.com/specimen/Chicle).