Ogaday

Dictionary puzzle of the week:

The objective is to write an expression or function that, given a dictionary in which each value is a list, transpose it so the result is list of dictionaries, where:

• The keys of each new dictionary are the same as the keys of the original dictionary
• The values of the ith dictionary correspond to the ith elements of the values of the original dictionary

If that doesn't make sense, here's an example:

# Input:

Ogaday

from datetime import datetime, timedelta
from typing import Iterator


def date_range(
    start: datetime,
    stop: datetime,
    step: timedelta
) -> Iterator[datetime]:
    """

Ogaday

On my set up at least:

shift + alt gr ' <vowel>: áéíóú  
shitf + alt gr ` <vowel>: àèìòù  
shift + alt gr - <vowel>: āēīōū

Ogaday

Comparison using the python date format spec to parse RFC-822 date strings vs using dateutil.

In [1]: from timeit import timeit 
   ...: from time_rfc_822 import parse_RFC_822_naive, parse_RFC_822_dateutil 
   ...: sample_date_string = 'Fri, 06 Mar 2020 14:22:22 GMT' 
   ...: %timeit parse_RFC_822_naive(sample_date_string) 
   ...: %timeit parse_RFC_822_dateutil(sample_date_string)                                                                                                                                                                                                                                
17.5 µs ± 805 ns per loop (mean ± std. dev. of 7 runs, 100000 loops each)
154 µs ± 12.6 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)

Ogaday

# Python 3.5
import numpy as np
import matplotlib.pyplot as plt

from keras.models import Sequential
from keras.layers.core import Dense, Activation
from keras.optimizers import SGD

from sklearn.metrics import mean_squared_error

Ogaday

Mocking HTTP Requests in Click

Simple MVP to test mocking of HTTP requests made in Click commands using the responses library.

In simple_post.py we have a Click command that makes a request to https://httpbin.org, a useful API for testing requests.

In test_simple_post.py we have a test suite which can be run by PyTest. In the first

Ogaday

Prefect on Dask on K8s

Intro

A "Hello, world!" Prefect flow running on an ephemeral Dask Cluster on Kubernetes.

The Prefect Core docs suggest running flows on a Dask cluster via Dask Distributed and an article from the MET Office Informatics Lab demonstrates running an adaptive Dask cluster on k8s. This example is inspired by those sources, as well as the respective docs for the technologies used.

Running Instructions

Ogaday

"""
A native implementation of argmax.

https://towardsdatascience.com/there-is-no-argmax-function-for-python-list-cd0659b05e49

Tests::

    pytest --doctest-modules -vvv argmax.py

Linting::

Ogaday

Function Register

A Lightweigt function registry implementation.

Eg:

register = Register()

@register("hello")

Ogaday

Function Composition

A minimal compose implementation.

Given a list of functions, apply them squentially to an argument. This is equivalent to:

funcs = [h, ..., g]
compose(x, funcs) == g(...(h(x)))