thehappycheese

def calculate_nickmapbi_offset(carriageway, xsp):
    return {
        "L":{
            f"L{num+1}": +3.5/2 - 3.5*num for num in range(0,7)
        },
        "R":{
            f"R{num+1}": -3.5/2 + 3.5*num for num in range(0,7)
        },
        "S":{
            f"L{num+1}": -3.5/2 - 3.5*num for num in range(0,7)

thehappycheese

def fuzzy_column_name_match(list1, list2):
    from fuzzywuzzy import process
    threshold = 60
    response = {}
    for name_to_find in list1:
        resp_match =  process.extractOne(name_to_find ,list2)
        if resp_match[1] > threshold:
            response[name_to_find] = resp_match[0]
    return response
  

thehappycheese

def run_in_executor(f):
    """
    This is a hack to turn a legacy blocking funciton into an async function.
    Thanks to balki https://stackoverflow.com/a/53719009/1782370
    
    Example:
    
    The following example shows how to use use a blocking
    `azure.identity` credential type with `pandas.read_parquet()`.
    Pandas normally requires that you use one of the limited

thehappycheese

for group_index, group in df.groupby(["road_number","cway"]):

    blank_rows = group[group["cluster"].isna()]
    filled_rows = group[group["cluster"].notna()]

    for blank_row_index, blank_row in blank_rows.iterrows():

        # find distance by looing for minimum "signed overlap"
        overlap_min = np.maximum(filled_rows["slk_from"], blank_row["slk_from"])
        overlap_max = np.minimum(filled_rows["slk_to"], blank_row["slk_to"])

thehappycheese

import pandas as pd
from azure.identity.aio import DefaultAzureCredential
CONTAINER = "..."
STORAGE_ACCOUNT_NAME = "..."
pd.read_parquet(
    path            = f"abfss://{CONTAINER}@{STORAGE_ACCOUNT_NAME}.dfs.core.windows.net/some/path/example.parquet",
    storage_options = {"credential":DefaultAzureCredential()}
)

thehappycheese

# Must use the async io variant of Azure Credentials
from azure.identity.aio import DefaultAzureCredential
# fsspec directly, use this library which implements it and gives better type hints and autocompletion
import adlfs

cloud_filesystem = adlfs.AzureBlobFileSystem(
    account_name="<STORAGE_ACCOUNT_NAME>",
    credential=DefaultAzureCredential()
)

thehappycheese

# Take a snippet of python code that defines at least one function at the top-level, then returns the last defined function.
import ast
def compile_function(source:str):
    # parse first, so we can automatically find the funciton name later
    parsed = ast.parse(source)
    # compile in specified `scope` dictionary
    exec(compile(source, "", "exec"), scope:={})
    # return the last function definition
    for item in reversed(parsed.body):
        if isinstance(item, ast.FunctionDef):

thehappycheese

import numpy as np
from numpy import typing as npt
def overlap(a:npt.NDArray, b:npt.NDArray, x:npt.NDArray, y:npt.NDArray):
   """Compute the signed distance between lists of intervals"""
    overlap_min = np.maximum(a, x.reshape(-1,1))
    overlap_max = np.minimum(b, y.reshape(-1,1))
    signed_overlap_len = overlap_max - overlap_min
    return signed_overlap_len
  

thehappycheese

import pandas
from zipfile import ZipFile
zip_file_path = "some_zip.zip"
# some_zip.zip/
# ├─ part1.csv
# ├─ part2.csv
# ├─ part3.csv
zip_file = ZipFile(zip_file_path)
extracted_data = pd.concat([
    pandas.read_csv(

thehappycheese

# inspired by https://stackoverflow.com/questions/1524126/how-to-print-a-list-more-nicely

# needs refinement before I post as answer though. I'll update this at some point

from typing import Iterable, Literal
def print_columns(data:Iterable, columns:int=3, sep:str=" ", alignment:Literal[">","<","^"]=">"):
    """Prints a list of objects in columns.
    
    `data` should be an iterable object, such as a list. Each element of data will be converted to a string using the built in `str()`
    `sep` is a string used to separate the columns. defaults to `' '`