Tips: For those coming from R: Silent In Place Replacement

Silent, in place assignment updating an object This tripped me up even though it’s consistent with how I’ve seen other objects behave. I needed an attribute to hold data extracted from a collection of files in a directory and created a class for this.

class hps_search_experiment:
    def __init__(self, path="", trial_type=''):
        self.path = path
        self.trial_type = trial_type
        self.hps_best_trial = None
        
    def process_hps_files(self):
        # ...
        
        self.hps_best_trial = hps_best_trial

However, running like so fails.

test = hps_search_experiment(
    path = './hps_search_intermediates_G/', 
    trial_type = 'rnr')
    
test = test.process_hps_files()
test.hps_best_trial

#> AttributeError: 'NoneType' object has no attribute 'hps_best_trial'

This had me baffeld because I was thinking with R’s norms of data <- data %>% funciton() where in place replacement is the exception. Instead I needed to be thinking with python’s base object norms (e.g. a_list.extend(['b', 'c']) ). This fails because I overwrote test with the output of the method, which returns notthing since it’s overwriting attributes within test’s scope.

These would also work to update the attribute:

self.hps_best_trial = hps_best_trial

hps_search_experiment.__setattr__(self, "hps_best_trial", hps_best_trial)

# if it's initialized as a list
self.hps_best_trial.append([hps_best_trial]) 
# if a dict is initialized for data
self.data = {'a':1}
self.data.update({'hps_best_trial':hps_best_trial})