About **Programmatic Answer Generation** Part

[yc1999]

Thanks for your great work! Are the python codes in Programmatic Answer Generation generated by LLM or written by human?

[yc1999]

And one more thing, it looks like that you normalize the format of time to XX:XX in flight-easy.json:

However in Combined_Flights_2022.csv, the format is as follows, which will cause mismatch and decrease of EM :

Thanks :)