Update the parmap calls to use the new parmap arguments

[zeehio]

Hi Davide,

I had to make some changes in parmap arguments and now all parmap arguments are prefixed with pm_ (as you can see from the changes in this pull request). Old arguments still work as before, although if they are used they give a deprecation warning. I plan to support both the old and the new arguments for a while, but eventually I will have to break backwards compatibility.

If you want to test this, make sure you have installed the latest parmap version with pip install -U parmap. This last version works with both the new syntax as well as the old one so any code you have will still work.

This pull request updates your parmap calls to the new arguments.

The last parmap version I released allows for keyword arguments to be passed to parallelized functions like this:

parmap.map(yourfunc, your_list, your_arg=your_val)

In order to prevent confusion between parmap arguments and keyword arguments passed to yourfunc, it was necessary to use a common prefix for all parmap parameters.

parmap.map(yourfunc, your_list, your_arg=your_val, pm_parallel=True)

With this new syntax, it is easy for anyone to see that keyword arguments starting with pm_ are for parmap, while other arguments (here your_arg) will be passed to the parallelized function.

That's why I had to change parallel to pm_parallel.

On the other hand, the best way to use parmap with a specific number of CPUs is through the pm_processes argument, instead of creating a multiprocessing pool manually. This has been changed in this commit as well.

Currently the old parmap arguments are still valid, and they give deprecation warnings so users can update the syntax. However it is best to transition to the new arguments, as a future parmap 2.0 version might break this backwards compatibility.

The parmap dependency in setup.py has been updated accordingly as well.

Sorry for the inconvenience :sweat_smile: and have a nice day

[zeehio]

By the way, parmap now supports showing a progress bar. If you are interested, pass pm_pbar=True and make sure you have the tqdm package (pip install tqdm)

[dgerosa]

Thanks!