使用魔术功能自动执行一组命令/导入

Question

I create a lot of different notebooks during my analyses. However every notebook starts with the same imports and other statements. Most generally the following lines of code are my standard input for the first cell:

import pandas as pd
import numpy as np
.. otherimports

import sys
sys.path.append('../..')
... more imports from this directory

%matplotlib inline
%load_ext autoreload
%autoreload 2

I've grown tired of either copy pasting or manually typing this stuff for every notebook and am now trying to find an automated way of doing this.

I looked at ipython extensions, so i could just call

%load_ext setup_analysis

but can't manage to effectively import stuff into the user environment using such an ipython extension. The extension is running, but imports made in the extension file are not imported into the notebook kernel (which makes sense). I tried using the interactive shell object in combination with user_expressions, but didn't work either.

My question is: What is the easiest way to automate this standard cell, if an extension is the way to go, how can i do an import (and the relevant magic functions) from such an extension?

Answer 1

fixed it:

Probably a bit of a contrived, but it works so I'm happy. Leaving this for future people who want to make their lives easier.

Solution:

Make a ipython extension by following these instructions by following these instructions .

In the .py file for this extension I wrote the following code:

def instructions():
    inst = """
    import pandas as pd
    ... other imports

    sys.path.append('../..')
    ... more imports

    """
    return inst


def load_ipython_extension(ipython):
    compiled_file = "prep_log.py"
    code = compile(instructions(), compiled_file, 'exec')
    ipython.run_code(code)
    ipython.run_line_magic("matplotlib", "inline")
    ipython.run_line_magic("load_ext", "autoreload")
    ipython.run_line_magic("autoreload", "2")

Now i just have the line %load_ext prep on the top of my notebook and I'm ready to go :).