1、新增jupyter notebook解析檔案

首先，建立一個python檔案，命名為Ipynb_importer.ipynb，直接貼上下面的程式碼。  

import io, os,sys,types
from IPython import get_ipython
from nbformat import read
from IPython.core.interactiveshell import InteractiveShell

class NotebookFinder(object):
    """Module finder that locates Jupyter Notebooks"""
    def __init__(self):
        self.loaders = {}

    def find_module(self, fullname, path=None):
        nb_path = find_notebook(fullname, path)
        if not nb_path:
            return

        key = path
        if path:
            # lists aren't hashable
            key = os.path.sep.join(path)

        if key not in self.loaders:
            self.loaders[key] = NotebookLoader(path)
        return self.loaders[key]

def find_notebook(fullname, path=None):
    """find a notebook, given its fully qualified name and an optional path

    This turns "foo.bar" into "foo/bar.ipynb"
    and tries turning "Foo_Bar" into "Foo Bar" if Foo_Bar
    does not exist.
    """
    name = fullname.rsplit('.', 1)[-1]
    if not path:
        path = ['']
    for d in path:
        nb_path = os.path.join(d, name + ".ipynb")
        if os.path.isfile(nb_path):
            return nb_path
        # let import Notebook_Name find "Notebook Name.ipynb"
        nb_path = nb_path.replace("_", " ")
        if os.path.isfile(nb_path):
            return nb_path

class NotebookLoader(object):
    """Module Loader for Jupyter Notebooks"""
    def __init__(self, path=None):
        self.shell = InteractiveShell.instance()
        self.path = path

    def load_module(self, fullname):
        """import a notebook as a module"""
        path = find_notebook(fullname, self.path)

        print ("importing Jupyter notebook from %s" % path)

        # load the notebook object
        with io.open(path, 'r', encoding='utf-8') as f:
            nb = read(f, 4)


        # create the module and add it to sys.modules
        # if name in sys.modules:
        #    return sys.modules[name]
        mod = types.ModuleType(fullname)
        mod.__file__ = path
        mod.__loader__ = self
        mod.__dict__['get_ipython'] = get_ipython
        sys.modules[fullname] = mod

        # extra work to ensure that magics that would affect the user_ns
        # actually affect the notebook module's ns
        save_user_ns = self.shell.user_ns
        self.shell.user_ns = mod.__dict__

        try:
          for cell in nb.cells:
            if cell.cell_type == 'code':
                # transform the input to executable Python
                code = self.shell.input_transformer_manager.transform_cell(cell.source)
                # run the code in themodule
                exec(code, mod.__dict__)
        finally:
            self.shell.user_ns = save_user_ns
        return mod
sys.meta_path.append(NotebookFinder())

 

2、呼叫jupyter notebook module

只要在我們的工作目錄下放置Ipynb_importer.ipynb檔案，就可以正常呼叫所有的jupyter notebook檔案。 這種方法的本質就是使用一個jupyter notenook解析器先對.ipynb檔案進行解析，把檔案內的各個模組載入到記憶體裡供其他python檔案呼叫。  新建一個.ipynb檔案，裡面定義要呼叫的函式，命名為until5.ipynb  (註明：以下是用梯度下降演算法求解線性迴歸)。

def error_function(theta, X, y):
    '''Error function J definition.'''
    diff = np.dot(X, theta) - y
    return (1/2*m) * np.dot(np.transpose(diff), diff)

def gradient_function(theta, X, y):
    '''Gradient of the function J definition.'''
    diff = np.dot(X, theta) - y
    return (1/m) * np.dot(np.transpose(X), diff)

def gradient_descent(X, y, alpha):
    '''Perform gradient descent.'''
    theta = np.array([1, 1]).reshape(2, 1)
    gradient = gradient_function(theta, X, y)
    while not np.all(np.absolute(gradient) <= 1e-5):
        theta = theta - alpha * gradient
        gradient = gradient_function(theta, X, y)
    return theta
 

再新建另一個.ipynb主檔案，該檔案呼叫untiled5.ipynb檔案中的gradient_descent函式和error_function函式

import numpy as np

# Size of the points dataset.
m = 20

# Points x-coordinate and dummy value (x0, x1).
X0 = np.ones((m, 1))
X1 = np.arange(1, m+1).reshape(m, 1)
X = np.hstack((X0, X1))

# Points y-coordinate
y = np.array([
    3, 4, 5, 5, 2, 4, 7, 8, 11, 8, 12,
    11, 13, 13, 16, 17, 18, 17, 19, 21
]).reshape(m, 1)

# The Learning Rate alpha.
alpha = 0.01

import Ipynb_importer
import Untitled5

optimal = Untitled5.gradient_descent(X, y, alpha)
print('optimal:', optimal)
print('error function:', Untitled5.error_function(optimal, X, y)[0,0])