Consider: Adding a detailed and ubiquitous "event" logging system

[christianechevarria]

Is your feature request related to a problem? Please describe.
Unit tests are great, stack traces are great, but they only give you a snapshot of a behavior that is unexpected and a function call, respectively. During testing, the question: "what series of steps caused this error?" often arises. There's great value in being able to understand the conditions under which problems arise and test to see if the same conditions always lead to the same problems.

Describe the solution you'd like
The full-neat rewrite is introducing a network.mutations array that logs the series of mutations that have been applied to an individual network, this is very useful given the evolutionary nature of the library but ideally we could devise a way to track all operations on a per-network / per-neuron / per-unit level and inspect them later in a convenient centralized way or maybe even attaching one standard top level data-structure to each unit that we could be conditioned to check in cases of failure or questions of performance. Although we have more work to do to get the hard metrics of testing down, tools to develop quick heuristics about improving the library are also necessary.

Describe alternatives you've considered
The quick and dirty version of this is just to keep adding specialized tracking arrays to each unit, nodes, connections, networks, populations and modify the relevant functions to track changes so that we can then access those tracking arrays later. This seems inelegant, but at least for now it would help debugging and developing tremendously.

[GavinRay97]

A really solid reference implementation for this would be the walkthrough and tiny amount of code needed to implement undo/redo mutations and time-travel debugging in Redux:

https://github.com/reduxjs/redux/blob/master/website/docs/recipes/ImplementingUndoHistory.md

You can apply the same sort of logic here, since Redux is just a paradigm for managing mutating stateful data.

Something that will be REALLY helpful here is the new Firefox Replay that should be released shortly (it is currently available only on Firefox Nightly for OSX):

https://firefox-replay.com/
https://developer.mozilla.org/en-US/docs/Mozilla/Projects/WebReplay

Mozilla is constantly putting its efforts into improving Firefox’s devtools. One such effort is Firefox Replay, an experimental tool that allows Firefox content processes to record their behavior so that it can be replayed and rewound later.

The main highlight of Firefox Replay is the “code timeline” that enables you to scan through every code execution at a glance. Along with execution points, the timeline also shows exceptions, events, and network requests in real-time. It also allows you to save your recordings and pick up where you left afterward.

How Firefox Replay works:

The record and replay behavior is achieved by “controlling the non-determinism in the browser.” Initially, it records non-deterministic behaviors (intra-thread and inter-thread) and then replays it later to “force the browser to behave deterministically.”

Firefox Replay includes IPC integration to enable communication between a recording or replaying process and the chrome process. Its rewind infrastructure allows a replaying process to restore a previous state. Its debugger integration enables the JS debugger to read the required information from a replaying process and control the process’s execution.

[christianechevarria]

Love the Redux undo-history, hot-linked the demo gif below just because it was so impressive

Same concept but with mutation methods, disconnections, node removals, and crossovers / network merges to start. Will work excellently as part of testing / debugging suite @luiscarbonell is working on. I'd love to see the visualization implemented with @teh-mICON code for network visualizations so we can track network changes backwards and forwards in time.

Edit adding the awesome visualization by @teh-mICON for reference

[micon4sure]

Could easily add mutation history, maybe even ancestry to the visualization.
Would even be possible to visually step through forward and/or backward passes.

[christianechevarria]

Could easily add mutation history, maybe even ancestry to the visualization.
Would even be possible to visually step through forward and/or backward passes.

💯 Imagine having left and right arrows where you could just click through the forward / back passes and the mutations and have smooth animations of the changes that happened along with seeing a step id + general metadata on each step. This would be killer as a teaching tool