add the example 'caching_and_replay' which uses Mesa-Replay

[Logende]

No description provided.

[rht]

I tried to lint the README.md via https://vale.sh/, as an experiment. I found that it said "cachable" is not a thing. According to mikebronner/laravel-model-caching#159, "cacheable" occurs ~9x more often than "cachable". And so, I think we should use "cacheable".

As a side note, I think we should have a prose linter for the repo. Haven't decided yet which one to use. Contenders: https://github.com/errata-ai/vale#functionality.

[Corvince]

So cool to see this. It is always nice to see libraries coming up to populate the mesa ecosystem! I am excited to take a closer look at the library and am positive about this PR being mergeable.

[rht]

Pressed the wrong button. Wanted to say that one more use case would be to resume execution of a simulation run at a checkpoint.

[Logende]

I tried to lint the README.md via https://vale.sh/, as an experiment. I found that it said "cachable" is not a thing. According to GeneaLabs/laravel-model-caching#159, "cacheable" occurs ~9x more often than "cachable". And so, I think we should use "cacheable".

As a side note, I think we should have a prose linter for the repo. Haven't decided yet which one to use. Contenders: https://github.com/errata-ai/vale#functionality.

Renamed cachable to cacheable in both the mesa-replay repo and in this example.

[Logende]

Pressed the wrong button. Wanted to say that one more use case would be to resume execution of a simulation run at a checkpoint.

I will add something like a "go_to_step(step: int)" function and a step_backwards() function to cacheable_model. This could then be integrated into the UI to allow the user to jump to some step or move backwards.

Edit: I think I will go with just a function to resume at some checkpoint because moving backwards would make this a little more complicated. It would be easily possible with the regular CacheableModel, but it would not be with StreamingCacheable model.

[rht]

Pressed the wrong button. Wanted to say that one more use case would be to resume execution of a simulation run at a checkpoint.

I will add something like a "go_to_step(step: int)" function and a step_backwards() function to cacheable_model. This could then be integrated into the UI to allow the user to jump to some step or move backwards.

No worries about this one for this PR. I will open an issue in https://github.com/Logende/mesa-replay.

[Logende]

Pressed the wrong button. Wanted to say that one more use case would be to resume execution of a simulation run at a checkpoint.

I will add something like a "go_to_step(step: int)" function and a step_backwards() function to cacheable_model. This could then be integrated into the UI to allow the user to jump to some step or move backwards.

No worries about this one for this PR. I will open an issue in https://github.com/Logende/mesa-replay.

Already implemented and tested 😸
Logende/mesa-replay@ff94c06

[rht]

cc: @snunezcr. Finally, we are having a model replay feature!

[Tortar]

I want just to say that we should really try to merge mesa-replay inside mesa itself in the future overcoming the last difficulties, because it can be a very useful functionality, for users to benefit the most it should be in the main repo :-)

[rht]

I want just to say that we should really try to merge mesa-replay inside mesa itself in the future overcoming the last difficulties, because it can be a very useful functionality, for users to benefit the most it should be in the main repo :-)

That's the plan. Currently, it's a separate repo so that it's more convenient to do breaking changes. Also, the cache size is too big at the moment.

@jackiekazil any comments?

[rht]

Merging now. Thank you @Logende !