Incorrect order of answers when using the wildcard '*' in DAWG

[lifmore]

Hi,

I wonder if there is a small issue in the file automata.py, function __words_with_wildcard, between lines 128 and 147, when the case letter=='*' is processed.

If the dictionary is made of, for example, "CHIAC" and "CHIC", and the query is "CHI*C", the result will be return in an incorrect alphabetical order : "CHIC" then "CHIAC".

This is because the case words_at_current_level is processed before checking the children.

So, for "CHI*C",

• words_at_current_level will first find "CHIC"
• then the loop for child in node.children will find "CHIAC", resulting in an incorrect order of answers.

Any idea? Or maybe did I misunderstood the code?

Best,
Lionel

[lifmore]

I was able to fix it by modifying the __words_with_wildcard function.
The key idea is to process the words_at_current_level case in the middle of the "for child in node.children:" loop, at the right timing, and not before or after the loop.

[aosingh]

Thank you @lifmore

I just saw this. I will try to reproduce the issue and get back to you.