Journaling As Performance Art

by Mike Levin

Okay, I’m going to keep pushing forward. Every journal I keep which is not my main private daily journal (like this) is actually now becoming performance art because of the ease of publishing these days. I keep stuff in Github anyway (still can’t believe it’s now Microsoft) and the github.io publishing system is so slick and easy, it’s hard not to do if you’re already on Github.

This is good for me because I definitely have a predilection to think and plan without actually DOING the follow-up. By virtue of this auto-publishing such as it were, I am actually accomplishing something… actualizing, materializing, what-have-you in a way that touches potentially many people here in this local spacetime region. Sure, non-locality and eternal souls, but also the here-and-now. Sorry to get too wacky, but I’m a wacky guy.

Okay, next! Clearly I need to move everything over to mlseo so that no matter what comes up, even the complex “RankLayer” which is one of my big featured deliverables at work, should be a light and breezy Notebook that I can push this way and that into different deliverable types. See? Everything looks like a nail when all you’ve got is a hammer. But the point is to make it a whopping good hammer. Okay… next! Again…

1, 2, 3… 1? That’s how I instantiate. 1 in this case is re-establishing state from yesterday when I was doing so well. Do I live-cast? No, but embed yesterday’s video into the below entry. Okay, I did that and pushed and checked and it worked. Amazing that’s the world we live in. Just keep making an impression, whether there’s livecasts and videos being produced or not.

Okay, let’s start moving more functions over. Let’s tackle concurrency. I already have httpx imported and I fixed the links on the mikelev.in site so it’s good to do another crawl. This is where I can make a big improvement. In the past I created a get_responses function which got fed a list of URLs and returned the responses. However there are several things wrong with this model. Ugh! Fixing it could be weird API-work.

I want to be able to feed this function either:

• A list of URLs
• A list of tuples or namedtuples

What comes back should be a list of tuples. The first position of the tuple should be whatever was used to make the API-request, be they actual URLs or tuples that got transformed into API-requests. I need to handle both. Interesting!

response = get_responses(list_of_urls)
response = get_responses(list_of_tuples, func_name)

Yes! This thing has to be designed to either receive a list of URLs OR a list of tuples and what to do with the tuples. It’s got to be fairly rigid in the fact that each tuple gets fed to the function. Here’s my test:

def foo():
    return 'bar'

def get_responses(requests, func):
    return requests, func

> get_responses([1, 2, 3], foo)

([1, 2, 3], <function __main__.foo()>)

Okay, now detect if func is func:

def foo():
    return 'bar'

def get_responses(requests, func=None):
    rv = False
    if func and callable(func):
        rv = True
    return rv

get_responses([1, 2, 3], foo)

Okay, now start handling list of URLs when provided:

def get_responses(requests, func=None):

    rv = False

    use_func = False
    if func and callable(func):
        use_func = True

    if not use_func:
        if type(requests) == str:
            requests = [requests]
        if not all([good_url(x) for x in requests]):
            return "Input must be a list of URLs."

    return rv

Okay, and now handle lists of URLs and lists of tuples (plus func) differently:

def get_responses(reqs, func=None):

    rv = False

    use_func = False
    if func and callable(func):
        use_func = True

    if not use_func:
        if type(reqs) == str:
            reqs = [reqs]
        if not all([good_url(x) for x in reqs]):
            return "Input must be a list of URLs."

    resps = []
    for req in reqs:
        if use_func:
            atup = (req, "foo")
        else:
            atup = (req, "bar")
        resps.append(atup)

    rv = resps

    return rv

> get_responses([1, 2, 3], foo)

[(1, 'foo'), (2, 'foo'), (3, 'foo')]

This is what we’re used to for fetching http (web pages).

import requests

r = requests.get(url)

Okay, pushed out a very nice video.