How to make Python' function to see imported libraries

Question

How to make imported self-written function to see libraries which are in the main file?

I have the main file written in Jupiter Notebook in the root directory with code. This file contains many libraries, including Numpy and the others.

Along with the external libraries, I import in my file self-written library with functions: import functions_lib. Self-written library is also located in the main root.

But the the imported functions (eg functions_lib.xgb_plot_k_fold_roc_curve(X_train, y_train, nfold, seed) don't see the libraries imported from the main file (eg Numpy).

#main file:
import pandas as pd
import numpy as np
import functions_lib

functions_lib.xgb_plot_k_fold_roc_curve(X_train, y_train, nfold, seed)

# improterd library code
def xgb_plot_k_fold_roc_curve(X_train, y_train, nfold, seed):
    ...
    mean_tpr = np.mean(tprs, axis=0)
    ...


NameError                                 Traceback (most recent call last)
<ipython-input-15-1a9972b490ae> in <module>
      2 
      3 
----> 4 functions_lib.xgb_plot_k_fold_roc_curve(X_train, y_train, nfold, seed)

C:\projects\Python-model\functions_lib.py in xgb_plot_k_fold_roc_curve(X_train, y_train, nfold, seed)
     39     tprs = []
     40     aucs = []
---> 41     mean_fpr = np.linspace(0, 1, 100)
     42     plt.figure(figsize=(10,10))
     43     i = 0

NameError: name 'np' is not defined

Answer 1

It is useful to understand variables and their scope . Consider np as a variable. You can use it only in the module where it was defined at the module level. If you load module inside a function as:

def func():
    import numpy as np
    pass

you can not use np outside the function since it is a local variable.

It is not like a preprocessor directive #include in C/C++. Thus insert import numpy as np in each module where you use it.