Command line tool to create a tweet based on a #100DaysOfCode log message.
The log from which the tweet will be generated has to have a format like my log.
You can find my log at https://log100days.lpld.io/log.md.
My log is based on the original #100DaysOfCode log repo.
But I am diverting from that format a little.
Specifically, I am using sub-headings (of the third level ### ...
) instead of strong text **...**
to delimit the daily sub-sections ("Today's Progress", "Thoughts", "Links").
Check out my raw log for details.
This is a markdown log, that is converted into an HTML page.
I have created a little Quart/Flask app that converts my log to an HTML site using the markdown2
package.
But other Markdown converters should work similar.
Once you have an HTML document with h2
day headers and h3
sections for "Today's Progress" and "Link(s)", you can point the tool at the URL of your HTML log page and generate a Tweet from it.
To actually enable the tweeting, you need to create a Twitter developer account and get an API key, API secret, Access Token and Access Token Secret.
The tweet will also contain the first link that you define in the "Link(s)" section under today's log. To save some space, the link is shortened. By default, I use my own link shortener.
If you want to, you can also use the Bit.ly service. Their links might be a bit shorter and they provide you with some engagement and tracking info. My link shortener does not do that. To use the Bitly service, you need to generate an API key through them and add it to the config (see below).
If you want to create a tweet for a different day than today, you can do so with the --offset
command line flag.
The offset is defined in integer days relative to today.
So to generate a tweet for yesterday use -o -1
.
If you want to suppress the actual tweeting and only see the message in the console, use the --testmode
command line flag.
I recommend pipx
to install python scripts and other tools in isolated virtual environments. This keeps the your platform Python installation clean and you don't have to worry about activating a particular virtual environment to use a tool/script.
pipx install logtweet
This way you will have a clean environment and the tool still available on the command line.
But if you want to, you should be able to install it with pip install logtweet
.
You need a configuration file for the script to work.
The config.ini
can either be in the current working directory or in ~/.config/logtweet/
.
Because I can not figure out how to define a "post-install hook" that is run by pip
, I can not generate an example config at the defined location.
This means you have to put the config there manually.
You can find an example config on GitHub.
In that config file you define the URL where your log can be found and the API keys and access tokens that are needed for Twitter and Bit.ly.
Install with
python -m pip install -e ".[develop]"
This installs the app dependencies as well as tools to develop and distribute the package.
Run tests with
$ pytest
This project makes use of code linters to keep the code quality consistent.
Run flake8
for general code quality check.
Run mypy .
to check type hints in the code.
To keep an eye on the current code quality, I would suggest installing the appropriate plugin for you editor/IDE to show the lints while working on it.
To distribute the package on PyPi, first create the latest build with:
python setup.py sdist bdist_wheel
This creates the new distribution files in the /dist
directory.
Upload the created files (e.g. for version 0.1.4) with twine
.
twine upload dist/logtweet-0.1.4*