hey there is an error in the sample generator spark code(for first batch tests), but its about saving numpy array imho and reading of it,  since all logic there were tested.

hmm should i fix it now.

what i were thinking of: 

either reading some anthropology topics that i had not started reading a cultural anthropology book excluding a few pages yet,

hmm or fixing this error and continueing the topology revising task. 

i think i would first still apply some analysis on geometry of the group via topology knowledge even before moving to neural networks, hmm or generate shapes of integers  to 10000 :D

i mean group specific properties of such integers.

hmm before nns lets check geometrical similarities or possible evolution of that shapes of geometries.

it would be similar to wedges of Sumerian cuniform  since shapes are like that from initial check not exactly like that.

but there is possibility of having that shapes internal system evolving with some logic pattern that needs to be searched. so that would eventually could be accomplished best with various nn types of temporal evolution patterns, or static patterns, whatever pattern there if exists to find the groupbehaviour of  integers.

of course it could turn out that nn is not well suited for such analysis task. but lets check. (e..g maybe modeling with functions specific functions investigations on dyanmics of that patterns evolution might be more apporp than nn based check. but would start with nn after first would again continue some manual batch checks on regarding topological data of the patterns which might give some info on static behaviors of repeated patterns maybe if exists if its not dynamically changing patterns along with some model function or something to model through nn. 

hmm what i decided to follow on tonight: 

after some rest -> would continue with topology revising to fastly be able to check manual check static patterns in the sample data (after fixing file writing error). 

i decided to go with algebraic and geometric topological analysis on data before passing the task to neural network to do more through analysis that could also incorporate temporality unlike me whom is not expert of special math functions that might be needed to utilize in analysis of topological structure (specifically change of topological structure part) but still some analysis I could do with the to be revised topological information to check.

so I postpone for tonight anthropology reading but then instead focus on specific topology revising studying period to fastly code the pattern analyzer and run with batch with spark to fastly have an insight to the data with also topological definitions.