Skip to content

Instantly share code, notes, and snippets.


Alexis Gallagher algal

  • San Francisco
View GitHub Profile
algal /
Created May 29, 2021
binomial coefficients with dynamic programming
# binomial coefficients
Recurrence relation:
C(n,k) is the count of all k-subsets in a set of n items.
C(n,k) = C(n-1,k-1) + C(n-1,k)
Intuition underlying the relation:
1. pick any item from the n-set, call it the special item.
2. Every k-subset either includes the special item or does not.

24 bit color in emacs, under

What's the problem?

The normal way to enable 24 bit color on a recent emacs (27.1) and a recent OS distribution (one which already has the terminal type definition xterm-direct), is just to set TERM=xterm-direct before running emacs.

But this does not work with, probably because of some incompatibility in its underlying library, hterm.

What's the solution

algal / isSubcollection.swift
Last active Mar 13, 2021
Look for a subarray in an array in Swift
View isSubcollection.swift
extension Collection {
/// True if `self` is a range within `collection`
/// - Parameter collection: a `Collection` like `Array`
/// - Returns: true or false
/// Since this works with `Collection` it works with `Array`, `ArraySlic`, etc..
func isSubcollection<U>(of collection:U) -> Bool
where U:Collection, U.Element == Self.Element, Self.Element:Equatable
let (a,b) = (collection,self)
algal / join_pdf.bash
Last active Feb 16, 2021
join PDFs on macOS from the command line with zero dependencies
View join_pdf.bash
if [ "$#" -lt 2 ]; then
echo "usage: $0 output_pdf_path input_path1..."
echo "This script joins the PDFs at input_path1, input_path2, etc."
echo "into one pdf at output_pdf_path"
echo "Known good: macOS 11.1, and below for many versions"
exit 1
algal / mdfindfile
Last active Nov 12, 2020
Easily search for files based their names or glob patterns, by wrapping Spotlight.
View mdfindfile
#!/usr/bin/env bash
# case-insensitive search for $1, maybe in dir $2
case "$#" in
0) echo "usage: $(basename $0) PATTERN [DIR]"
echo ""
echo " Lists paths of files matching PATTERN in DIR or below"
1) exec mdfind "kMDItemDisplayName == '$1'c"
*) exec mdfind "kMDItemDisplayName == '$1'c" -onlyin "$2"
algal /
Created Sep 9, 2020
Plot a confusion matrix in scikitlearn from data not from an estimator
# This uses scikit learn internals, since the sk public API requires you to pass
# in an estimator and sometimes you just want to pass in the some data you'd
# use to calculate a raw CM
def plot_cm(y_true,y_pred,labels):
from sklearn.metrics._plot.confusion_matrix import ConfusionMatrixDisplay
sample_weight = None
normalize = None
include_values = True
algal /
Created Aug 29, 2020
To use an SFUSD Gmail account to subscribe to a non-SFUSD Google Calendar

Subscribing to a Google Calendar from an SFUSD Google Account

SFUSD Google accounts can only send and receive emails from SFUSD teachers, so they cannot receive calendar invitations from normal gmail accounts, like the gmail accounts of their parents. So you cannot invite them to join a calendar as usual.

Here's a workaround.

In the normal gmail account:

  1. Create the calendar to share from a normal gmail account.
  2. Go to Calendar Settings / General
algal / excel_parseDates.vba
Last active Sep 5, 2020
please god why aren't these predefined?
View excel_parseDates.vba
' Expects a datetimestr in the format "YYYYMMDD" with - or / or no separator
' Parses independently of local region, unlike VBA.DateTime.DateValue()
' Known-good on Excel for Mac Version 16.4
Function parseYYYYMMDD(dateTimeStr As String) As Date
Dim normalized As String
normalized = VBA.Strings.Trim(dateTimeStr)
normalized = VBA.Strings.Replace(dateTimeStr, "/", "")
normalized = VBA.Strings.Replace(normalized, "-", "")
from typing import Dict, List, Tuple
def makeKeyDict(items:List[Tuple]) -> Dict:
"Gives a list of tuples (a,b), returns a dict mapping a to set of b"
d = {}
for (a,b) in items:
d[a] = set([b]).union(d.get(a,set()))
return d
def findKeysWithMultipleValues(keydict) -> List:
algal /
Created Jul 28, 2020
Read file paths, names, hashes into a data frame
from import * # to get L
import pandas as pd
def readMD5file(md5path:Path) -> pd.DataFrame:
Generate MD5 output file by doing a search like:
find /home/jupyter/data/foldersToAdd/ -iname '*jpg' -print0 | xargs -0 -n 100 md5sum >> /home/jupyter/data/foldersToAdd.md5.out