search

lardemua / atom

Calibration tools for multi-sensor, multi-modal robotic systems
GNU General Public License v3.0
252 stars 26 forks source link

organization of code of results for a paper or robot #282

Closed miguelriemoliveira closed 3 years ago

miguelriemoliveira commented 3 years ago

Hi @danifpdra,

About the new results for mmtbot. I already had this in mind for the papers of @eupedrosa and @aaguiar96 . The problem is that the results, unlike everything else, are lost in the computers of the person which is taking the results. I discussed with @eupedrosa some alternatives, but in the end we did not advance.

This is what I think we need.:

  1. a single place where the data, the scripts to process the data (or configurations for standard scripts are located)
  2. Typically the data is very large, so perhaps it could be on a google drive or one drive and there should be a script to download it to a standard location

So I propose that we create a git repository (or add a new folder or package to existing repos) which would contain this.

The mmtbot example:

A new mmtbot_evaluaton folder should contain

Any other ideas? Something missing here?

miguelriemoliveira commented 3 years ago

@danifpdra what is your opinion about this? If you agree I can reorganize, no problem.

danifpdra commented 3 years ago

Hi @miguelriemoliveira ,

I agree with you. I think the data should be available for all. Your solution seems good.

miguelriemoliveira commented 3 years ago

Hi @danifpdra ,

do you want me to start with this reorganization? I can give it a try ...

danifpdra commented 3 years ago

hi @miguelriemoliveira,

As you wish. Do you prefer if I do it? I can do it if you want.

miguelriemoliveira commented 3 years ago

I will do the first round and then you adjust as you want ok? Do you have everything pushed?

danifpdra commented 3 years ago

I've just pushed everything now.

miguelriemoliveira commented 3 years ago

Hi @danifpdra ,

I have the first proposal. All the things are in mmtbot_experiments. There is a global readme and a readme for each experiment.

What is missing:

  1. Checking the datasets and bagfiles. You are not using mine, we are already desynced which is another reason to get this up and running.
  2. Adding a folder with the results. That folder we were putting inside the dataset older should be in this repo I think ...

Can you do these two?

... and give comments and suggestion to improve ...

miguelriemoliveira commented 3 years ago

Closing...