GEOGRID vs GEOGRIDATA

[agrignard]

In this table https://cityio.media.mit.edu/api/table/dungeonmaster/

Inside GEOGRID each block has this properties

"properties":{"color":[106,61,154],"height":4,"interactive":"Web"}

and inside GEOGRIDDATA each block has this properties

{"color":[255,255,51],"height":4,"id":0,"interactive":"Web","name":"Residential"}

Do we need to repeat the information in both field? Isn't it redundant? Also the color value are different

[crisjf]

It is redundant, but it gives you the option of querying GEOGRID only once because it never changes and just calling GEOGRIDDATA when you detect the table update. Ignore the properties in GEOGRID and use the ones in GEOGRIDDATA. GEOGRID should be used when you need the coordinates of the table (you need them only at init) and to get the type definitions under GEOGRID/properties.

[RELNO]

Yes, as @crisjf mentioned the GEOGRID is state "0" of this table (i.e taboola rasa). Later states are GEOGRIDDATA

[agrignard]

OK

And is it necessary to store state 0 properties in GEOGRID? I mean in term of storing size (there was a time we were arguing wether we should use int or float to save space;-) )

Also Instead of writing issue for this kind of question is there an entry point for the meta structure of this json so I can look at it before to post an issue?

[RELNO]

And is it necessary to store state 0 properties in GEOGRID?

It's not about storing state 0. GEOGRID is the baseline data for both geometry and meta for each grid cell. This is how we initiate a table with all its data in one place (vs the multi fields we used to have).
GEOGRIDDATA is lighter, as it's only meta without geometry (think= old grid field). This allow iterations in real time with small(ish) overhead.

[crisjf]

So as we think about the Handler agent in GAMA, the GEOGRID endpoint should be called on init, and the GEOGRIDDATA endpoint should be used for everything else.

[crisjf]

I think we can close this, right?

[RELNO]

Closing for now, @agrignard open at will.