Save IPython Variables

You use IPython. It's pretty great. But things happen! Your computer crashes, your IPython shell crashes (in theory), your computer runs out of memory and the process becomes unreachable. You can lose your whole session, and potentially hours of work.

Sure, you can work out of a file, using %edit to keep your code on disk, but the variables that can take hours to build up -- whether it's through querying databases, scraping the web, or executing long-running algorithms -- can just vanish.

This is where this module comes into play: save-ipython-variables lets you do just that -- save your global IPython variables to disk easily, and load them back into the global namespace when you need them again, even in a whole new IPython session.

Example Usage

In [1]: from save_ipython_variables import load_all_variables, save_variable

In [2]: data = build_data_dictionary_from_db()

In [3]: save_variable('data', data)

In [4]: save_variable('number', 5)

In [5]: exit()

And now, in a new session:

In [1]: from save_ipython_variables import load_all_variables, save_variable

In [2]: load_all_variables()
Loaded the following variables: ['data', 'number']

In [3]: if data:
   ...:     print 'Loaded Successfully!'
Loaded Successfully!

You can also choose to load select variables:

In [1]: from save_ipython_variables import load_all_variables, save_variable

In [2]: load_all_variables(['number'])
Loaded the following variables: ['number']

In [3]: number
Out[3]: 5

In [3]: data
---------------------------------------------------------------------------
NameError                                 Traceback (most recent call last)
<ipython-input-8-2cd6ee2c70b0> in <module>()
----> 1 data

NameError: name 'data' is not defined

As you can see, number was successfully loaded, but data was ignored.

Details

Each variable is saved to disk in the current directory as <variable_name>.pkl. Variable names are also added to the set saved_var_names, which is saved to disk like all other variables.

Variables are loaded into scope by adding them to the __builtins__ dictionary.