Refactor heatmap calculations to categorize the results by opportunity

[EPajares]

Goal of this issue

The calculations on heat-maps was designed to do calculations over all data. In the next step, we need to separate the data by opportunities and do calculations per opportunity.

Resources

Deliverables

Branch to derive

[metemaddar]

To summarize our meeting on this:

1. Here first of all we need to separate the data (grid_ids & traveltime) and add them to their own category (For example as a dictionary). This also can decrease the sort time as we have smaller data series to sort.
2. Then we need to do calculations based on incoming calculation needs as input (Finish new heatmap schema #1786). These calculation names can get connected to their function (Compute mathematical operations on two numpy arrays #1783) to calculate and add needed data to response.

[metemaddar]

When separated grid_ids/travel_times by categories,
The calculations got a bit faster:
Previous Results:

sort takes: 0.36031150817871094 s
mins takes: 0.011055469512939453 s

New results:

sort takes: 0.24292492866516113 s
mins takes: 0.003187417984008789 s

Change

Sort: -36%
Mins: -70%

[EPajares]

Could you run a test with all the poi categories we have? They are listed in the opportunity_config table.

[EPajares]

I think we can close this.