search

hugomflavio / actel

Standardised analysis of acoustic telemetry data from fish moving through receiver arrays
https://hugomflavio.github.io/actel-website
26 stars 4 forks source link

Allow allocating multiple transmitter ids (the whole code-space-signal) to a single animal #98

Open hugomflavio opened 2 months ago

hugomflavio commented 2 months ago

Stems from: https://github.com/hugomflavio/actel/issues/97#issuecomment-2200001109

Currently, actel supports multiple signals in the same code-space, but will stop if the same signal is found in different code spaces.

To properly implement this, I think the biometrics input needs to be changed. I.e. we'll no longer have a CodeSpace and a Signal column (where signals can be separated by |), but instead the biometrics will have a single "transmitter" column, where full transmitter id can be split with |. actel can still check for stray signals on non-reported codespaces internally.

Should ask @jdpye if this has been an issue elsewhere and how it has been handled.

jdpye commented 2 months ago

Definitely has been an issue for anyone who's separated codespace and tag ID for datasets of larger size. my recommended practice is as you describe, to keep frequency/codespace/code in one unique field. A69-xxxx-xxxxx , etc.

We'll support larger/stranger codes when they come along but in a perfect world you'll always need to specify what frequency and codeset a given ID number is coming from.

jdpye commented 2 months ago

the H170-xxx-xxxxx in Ine's example there is acceptable and is the way we'd express it in our detection extract reports, for example.

hugomflavio commented 2 months ago

Ok, thanks for the input! The way forward then is to create a transmitter column in the biometrics, so users can input the whole thing right away. I'll have to do some work to keep the backwards compatibility with the old actel method, but it should all be contained within the pre-processing of the biometrics.