To run from Linux, see the github workflow.
To run from Windows, which I do worryingly much now:
set PYTHONPATH=C:\Users\<<user_name>>\PycharmProjects\quotations_w_unittest\app;%PYTHONPATH%
cd tests/unit
coverage run --omit="../../app/console_io.py" --source="../../app" -m unittest
coverage report -m --fail-under=90The quotes were pulled from kaggle, which I have trusted and enjoyed for years.
The quotations are in json form with a lot of extra fields and duplication which I have no need for.
Who needs quotations?
Who doesn't want inspiration?
I was hoping to use them to demonstrate how to use the unittest module, but they ended up being so malformed, and huge, that I made it my mission here to clean them up.
Even with a script this was arduous. I was going to demo kivy with this but found having 5 keys for 5 solutions to missing punction would be much faster.
As I say arduous. And huge, the unzipped source is 20+ MB. Many of these contain obvious mistakes, especially with spacing. Most notable is a capital in the middle of a word, or more than one. This means several words are joined. Sometimes this happens without a capital. I should run it all through a spell checker. I wonder if the objective of Kaggle here was for us to clean these up.
Anyway they are useful with a meaningful popularity index, that I will discard. In the tests directory I put the top 50 or so for integration testing. There are two occurrences of a capital inside a word. That's pretty dirty for a top 50 quote.
There are 3 methods of cleaning the quotes.
all- Manually pick what to do with each "middle capital". 5 choices. Progress is saved to the final output, which we pick up from next time.csv- Convert to CSVrandomx- Take random sample of x many.
The default output is a json list of quote: author pairs. Duplicates get removed because they were only duplicated so that they could be entered in multiple categories.
Just in case anyone actually wants to try and get some quotation, this is what we should do (taken from kaggle_normaliser.py):
cd C:/Users/<<user_name>>/PycharmProjects/quotations_w_unittest
python app/kaggle_normaliser.py all tests/integration/test-quotes.jsonWe run that in the test/workflow pipeline. The all method must be tested manually or through a test harness because it requires user input (usually quitting).