Incomplete Import

[epic_dolphin208]

Hi, I copied and pasted from 2 HyTek Meet Manager Result PDFs, it looks like it got the main structure, but seems to be missing athletes and results.

SwimSnap URLs -

https://swimsnap.com/results/2011-nc-nma-vs-team-greenville-6aa260

https://swimsnap.com/results/2010-nc-nma-holiday-invit-0181d5

EDIT: The 2010 NMA Holiday looks to be missing All Boys events, and the 2011 NMA vs TEAM looks to be missing "Senior" events (e.g. Boys Senior 200 Yard IM)

Thanks!
-M

[epic_dolphin208]

One more meet to add - https://swimsnap.com/results/2011-3a-western-regionals-6ecdc6

[adam]

Thank you for reporting the issues. I will look into this later today.

[epic_dolphin208]

Couple more I tried today:

https://swimsnap.com/results/nchsaa-3a-state-swimming-and-diving-champions-20a6c0

https://swimsnap.com/results/3a-state-swimming-and-diving-championships-b90f02

https://swimsnap.com/results/3a-state-swimming-diving-championships-7467cc (just a white screen)

https://swimsnap.com/results/2019-sunbelt-championship-meet-fcf890

[adam]

I did a quick check, and it looks like those are high school meet results.

SwimSnap works by recognizing patterns in the text that is pasted into it. Because of that, it isn’t very reliable when the format of the results is different from what it expects. In this case, the high school meet results use a different format than the USA Swimming club results we originally built and tested it for.

SwimSnap was released as a prototype last year and hasn’t seen much usage, so it hasn’t been a priority for us to improve.

I think the better long-term approach is to use AI to understand and extract the data, rather than relying on fixed text patterns. That would make it much more reliable across different result formats.

Thank you for testing it out.

[epic_dolphin208]

Hi Adam, thanks so much for looking into this. The NMA meets are club/USA-S meets, 2019 sunbelt champs a USMS meet, but the rest were indeed High School meets.

I think this is a really great idea, especially if it could lead to a results file generation to preserve older meet results on a limited basis, since PDF results tend to be what remains over time.

AI sounds like a very interesting approach!