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

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

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 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

#pip install pandas geopandas arcgis arcgis2geojson

import json
import geopandas as gpd
from geopandas import GeoDataFrame
from arcgis.features import FeatureLayer
from arcgis2geojson import arcgis2geojson

def get_arcgis_layer_as_geopandas(url:str) -> GeoDataFrame:
    layer = FeatureLayer(url)

thehappycheese

import threading
from http.server import BaseHTTPRequestHandler, HTTPServer

def serve_once(data:str, port=8000, host="127.0.0.1"):
    """
    Serves the given data on the given port and host, then returns after
    responding to one request.
    
    This is a helpful way to send HTML to a browser being controled using Selenium.
    """

thehappycheese

function prompt {
    $currentPath = (Get-Location).Path
    $driveLetter = $currentPath.Substring(0,1)
    $folderName = Split-Path $currentPath -Leaf
    "$driveLetter:/.../$folderName> "
}

thehappycheese

from pathlib import Path
import keyring
import pandas as pd
import os
import sys
# You need access to the compiled .NET AdomdClient .dll
# You can get it using nuget
# ```powershell
# nuget.exe install Microsoft.AnalysisServices.AdomdClient.retail.amd64
# nuget.exe install Microsoft.AnalysisServices.Tabular

thehappycheese

import pandas as pd

def split_dataframe_by_weighted_sequence(
        df:pd.DataFrame,
        split_column:str,
        weight_column:str,
        skew:float=0.5
    ):
    """
    Split a dataframe into two parts by a sequence of values and a sequence of weights.

thehappycheese

# pairwise coefficient of interpolation
Interp ← ÷:⊃(-≡⊢|≡/-)◫2

MarchingSquares ← (
  :
  ⊃(
    ≡(◫2)Interp
  | ◫2⍜⍉Interp
  | ≡≡(°⋯♭)◫2_2> # types
  | ≡°⍉⍉⊞⊟∩(⇡-1)⊃(⊢⇌|⊢)△