Convert a python string to an sklearn classifier object

Question

If i have a string:

"ExtraTreesClassifier(bootstrap=False, class_weight=None, max_depth=None, max_features=11, min_samples_leaf=2, min_samples_split=8, n_estimators=20, criterion="gini", n_jobs=4, random_state=12)"

What is the best way to convert it to an sklearn classifier so that i can fit my dataset to it?

Desired output:

<class 'sklearn.ensemble.forest.ExtraTreesClassifier'>

EDIT

In my case, the string is not wrapped in hyphens (understand that without the hypens i will be able to call the sklearn function), rather the variable is of type string.

SOLUTION

I have managed to solve it by first extracting the parameters to dictionary form, then placing it within the sklearn estimator:

import ast

ExtraTreesClassifier(**ast.literal_eval("{"+i.split("(", 1)[1][:-1].replace("=",":")+"}"))

i.split("(", 1)[1][:-1] removes "ExtraTreesClassifier(" and ")" from the string.

To format string to the same syntax as a dictionary, use: .replace("=",":") then wrap brackets "{"

Because this is still a string representation of a dictionary, to pass it as parameters to the sklearn estimator, this is needed to be wrapped around the dictionary: **ast.literal_eval() -> see this link