[patch] Introduce a flag for _not_ mangling the name of wrapped function jobs

[liamhuber]

Uses a private interface and defaults false, so this is totally backwards compatible.

def my_function(x):
    return x

job = self.project.wrap_python_function(my_function)
job.input["x"] = 42
job._mangle_name_on_save = False
job.save()
assert(job.job_name, my_function.__name__)

[liamhuber]

The osx 12 tests failed, but it was a timeout and I have trouble seeing in the log but it looks like a completely unrelated test. I'm trying to just re-run, but the button doesn't do anything on my browser...

/Users/runner/work/pyiron_base/pyiron_base/pyiron_base/jobs/datamining.py:651: PerformanceWarning: 
your performance may suffer as PyTables will pickle object types that it cannot
map directly to c-types [inferred_type->mixed,key->block1_values] [items->Index(['name', 'array'], dtype='object')]

  self.pyiron_table._df.to_hdf(
.
  0%|          | 0/4 [00:00<?, ?it/s]
100%|██████████| 4/4 [00:00<00:00, 41.70it/s]
TestWithProject: Setting up test project
.
The job test_a was saved and received the ID: 6
Loading and filtering jobs:   0%|          | 0/5 [00:00<?, ?it/s]
The job test_b was saved and received the ID: 7
Loading and filtering jobs: 100%|██████████| 5/5 [00:00<00:00, 645.32it/s]
The job test_c was saved and received the ID: 8

The job test_d was saved and received the ID: 9
The job test_table was saved and received the ID: 10
Error: The action has timed out.

[liamhuber]

The osx 12 tests failed, but it was a timeout and I have trouble seeing in the log but it looks like a completely unrelated test. I'm trying to just re-run, but the button doesn't do anything on my browser...

Indeed, the same timeout error appears on #1349. It went through OK on #1352 though, so I'll probably try just re-running a few times...

[jan-janssen]

Suggested change

	self._mangle_name_on_save = True # Automatically rename job using function and
	self._overwrite_job_name_by_hash = True

I feel when we have to explain the variable name in a comment, then maybe the variable name is not intuitive.

[liamhuber]

Yeah, I was aware of this deficiency -- I fixed it in the stacked PR, but it was still unprofessional of me to just leave it knowingly-bad here.

I find your suggestion loses information on when/why we would rename this, and focuses on the implementation (hashing) instead of the functionality (leveraging input values), so I opted to go with _automatically_rename_on_save_using_input.

This is definitely an improvement, although in the stacked PR I'm still inclined to leave to comment to give even more depth to why we're doing this.

[jan-janssen]

In principle I agree with the changes, I just find the variable name _mangle_name_on_save hard to understand and I do not understand why this information is stored in the HDF5 file.

[jan-janssen]

Looks good to me