Wrapper

The implementation of krmining.external.Wrapper Beta version¶

In [1]:

from krmining.external import Wrapper
from krmining.clustering import KMeans
from krmining.datasets import get_example_covid_id

import dill

from krmining.external import Wrapper from krmining.clustering import KMeans from krmining.datasets import get_example_covid_id import dill

Create the model¶

In [2]:

# Make model and save
df = get_example_covid_id()

for column in df.columns:
    df[column] = df[column] / df[column].max()

kmeans = KMeans(k=3, init="kmeans++")
kmeans.fit(df)

test_object = dill.dumps(kmeans)

# Make model and save df = get_example_covid_id() for column in df.columns: df[column] = df[column] / df[column].max() kmeans = KMeans(k=3, init="kmeans++") kmeans.fit(df) test_object = dill.dumps(kmeans)

c:\Users\Bina Umat\Desktop\Project\krmining_project\keyar-mining\krmining\clustering\_kmeans.py:22: UserWarning: The model still in maintaining in slow or extended memory
  warnings.warn(

Create the wrapper¶

In [3]:

# make wrapper

wrapper_object = Wrapper()

@wrapper_object.save_func("data_normalize")
def data_normalize(df, binary_model):
    for column in df.columns:
        df[column] = df[column] / df[column].max()

    return df, binary_model

@wrapper_object.save_func("cluster_predict")
def cluster_predict(df, binary_model):
    import dill

    model = dill.loads(binary_model)

    cluster = model.predict(df)

    return cluster

wrapper_object.add_step("data_normalize")
wrapper_object.add_step("cluster_predict")

# make wrapper wrapper_object = Wrapper() @wrapper_object.save_func("data_normalize") def data_normalize(df, binary_model): for column in df.columns: df[column] = df[column] / df[column].max() return df, binary_model @wrapper_object.save_func("cluster_predict") def cluster_predict(df, binary_model): import dill model = dill.loads(binary_model) cluster = model.predict(df) return cluster wrapper_object.add_step("data_normalize") wrapper_object.add_step("cluster_predict")

data_normalize is saved
cluster_predict is saved

Load the wrapper¶

In [4]:

cluster = wrapper_object.load_func(df, test_object)
cluster

cluster = wrapper_object.load_func(df, test_object) cluster

----------------
Before execution
----------------
- Time to execute data_normalize function: 0.0019981861114501953
- Time to execute cluster_predict function: 0.0050427913665771484
---------------
After execution
---------------

Out[4]:

array([2, 0, 0, 2, 0, 1, 2, 2, 1, 1, 1, 2, 0, 2, 0, 2, 2, 2, 2, 2, 2, 2,
       2, 2, 2, 0, 2, 0, 2, 2, 2, 0, 0, 0], dtype=int64)

Saving the wrapper and load¶

In [5]:

saved_wrapper = wrapper_object.save_object_to_var()
load_wrapper = dill.loads(saved_wrapper)

load_wrapper.load_func(df, test_object)

saved_wrapper = wrapper_object.save_object_to_var() load_wrapper = dill.loads(saved_wrapper) load_wrapper.load_func(df, test_object)

----------------
Before execution
----------------
- Time to execute data_normalize function: 0.0019526481628417969
- Time to execute cluster_predict function: 0.0029952526092529297
---------------
After execution
---------------

Out[5]:

array([2, 0, 0, 2, 0, 1, 2, 2, 1, 1, 1, 2, 0, 2, 0, 2, 2, 2, 2, 2, 2, 2,
       2, 2, 2, 0, 2, 0, 2, 2, 2, 0, 0, 0], dtype=int64)