Much more deep tests

[yhamoudi]

We should add as much as possible deep tests in data_deep.py. Every time we discover a question well parsed by our algorithm, we should add it immediately into this file. Thus, when we'll make some changes on the algorithm, it will not be necessary to check many examples manually to be sure that we didn't broke a useful rule.

(naturally, it's not necessay to add questions with the same structure, ex: Who is the president of France? and Who is the president of Germany?)

Let's close this issue when there will be 200 tests :)

• 50 tests done
• 100 tests done
• 150 tests done
• 200 tests done

[yhamoudi]

I've shortened the way we write deep tests. We can at least write each triple on a single line.

[Ezibenroc]

👍

But 200 tests with different structures seems a bit ambitious.

[yhamoudi]

there are already a lot of tests in documentation/

[yhamoudi]

But 200 tests with different structures seems a bit ambitious.

We can mix some difficulties. For instance Who is the president of Germany? and Who is the president of the United States? have the same grammatical structure but in the second case there is a merging (United + States)

I've added some achievements in the first post 💰 (and i've discovered the supported emoji)

[Ezibenroc]

👏