skip to main content

@tonyfast s notebooks

site navigation
notebook summary
title
ipython profile updates
description
when i want to write, i really want to write! i don't want to import modules i know i should have access to. i only care about this in interactive computing circumstances. i need to satisfy my compulsion to write new things, and avoid the old things. a revision stage after this initial burst will formalize the reuse of the module.
cells
6 total
3 code
state
executed in order
kernel
Python [conda env:p311] *
language
python
name
conda-env-p311-py
lines of code
38
outputs
2
table of contents
{"kernelspec": {"display_name": "Python [conda env:p311] *", "language": "python", "name": "conda-env-p311-py"}, "language_info": {"codemirror_mode": {"name": "ipython", "version": 3}, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.11.3"}, "widgets": {"application/vnd.jupyter.widget-state+json": {"state": {}, "version_major": 2, "version_minor": 0}}, "title": "ipython profile updates", "description": "when i want to write, i really want to write! i don't want to import modules i know i should have access to. i only care about this in interactive computing circumstances. i need to satisfy my compulsion to write new things, and avoid the old things. a revision stage after this initial burst will formalize the reuse of the module."}
notebook toolbar
Activate
cell ordering

ipython profile updates

when i want to write, i really want to write! i don't want to import modules i know i should have access to. i only care about this in interactive computing circumstances. i need to satisfy my compulsion to write new things, and avoid the old things. a revision stage after this initial burst will formalize the reuse of the module.

we're going to supply ourselves with better literate programming interfaces by defining a custom ipython configuration file that is loaded by default. once the configuration is applied we have a superior literate and functional programming vocabulary than we are normally offered.

programmatically discover the default ipython profile location which is where we need to write or update our configuration at. 

config = !ipython profile locate
config = pathlib.Path(*config, "ipython_config.py")

config_source contains the lines of code we want to write to our configuration. 

    def config_source():        
        source = """from nobook.utils import Index, Series, DataFrame, doctest
        import midgy,  anyio, pyperclip, importnb
        from pathlib import Path
        from IPython.display import *
        from dataclasses import dataclass, field
        from toolz.curried import *
        with importnb.Notebook(): from __llm_workflow import load_ipython_extension
        load_ipython_extension(get_ipython())
        del load_ipython_extension
        [get_ipython().user_ns.setdefault(k, v) for k, v in __import__("sys").modules.items() if "." not in k]
        print("my config loaded")
        """
                
        c.InteractiveShellApp.exec_lines = source.splitlines()

write the config_source to disc 

config.write_text(textwrap.dedent(
    "".join(inspect.getsourcelines(config_source)[0][1:])
))

when we restart our kernel we should have some new variables that we can use by default.

  • i always want my pandas dataframes because i use them as first class citizens 

            >>> assert all((Index, Series, DataFrame))

  • we loaded in all the modules we know about from sys modules. this enriched namespace makes it easier to continue flowing thoughts. 

            >>> assert all((pathlib,))

  • toolz is the shit for functional programming, especially the curried versions. we want them to shred faster. 

            >>> assert all((operator, compose, pipe))

all these modifications make it possible run this without any imports. we start with imports, but could remove them after our configuration was defined properly.