musically-ut

import numpy as np
from sklearn import datasets
from sklearn.semi_supervised import LabelPropagation, LabelSpreading
###for n_samples in [20, 200, 2000, 20000]:
###    X, y = datasets.make_classification(n_samples=n_samples, n_classes=3, n_informative=3)
for (X, y) in [datasets.load_iris(return_X_y=True)]:
    for model in [LabelPropagation(max_iter=1000),
                  #LabelSpreading(alpha=0.01),
                  #LabelSpreading(alpha=0.1),
                  #LabelSpreading(alpha=0.3)

musically-ut

#!/bin/bash

set -e
set -u
set -o pipefail

if [ -z $1 ]
then
	echo "Usage: $0 [user@]servername"
	exit 1

musically-ut

#!/usr/bin/env bash

set -e

for N in {1..5}
do
  for M in {0.1,1,2,5,10};
  do
    # Replace ./my_prog.py with your program to run it in parallel
    ./my_prog.py -N $N -M $M > output/stdout_N${N}_M${M}.txt &

musically-ut

4 dots puzzle

Initial position: There are 4 dots arranged as a square on a grid.

Move: The dots can be moved by reflecting them through any other three dots. So to move a point, draw a line segment from that dot to another dot, extend it to twice its length and move the first dot to the place where it terminates.

Question: Is it possible to transform the square into another square of a larger size?

---

musically-ut

#!/bin/sh

PAT='#\s*NO(.)?GIT'
git diff --cached --name-only | \
    xargs egrep -i --color --with-filename -n "$PAT" && echo "Found '"$PAT"' in code about to be committed. Rejecting commit." && exit 1

# Exit with error code zero if all went well.
exit 0

musically-ut

musically-ut

require 'graph';
require 'os';

function displayGraph(g, forward)
    local imgFile = os.tmpname() .. '.png'
    local dag = g.fg
    
    if forward == false then
        dag = g.bg
    end

musically-ut

par(mfrows=c(clustering$cluster,1))

for(j in 1:max(clustering$cluster))
{
  x <- 1:12
  d <- d[clustering$cluster == j,]
  plot.default(x = NA, xlim = c(1, 12), 
             ylim = c(min(y), max(y)), xlab = "Time", ylab = "Expression changes", 
             main = paste("Cluster", j), axes = FALSE)

musically-ut

function (eset, cl, mfrow = c(1, 1), colo, min.mem = 0, time.labels, 
          new.window = TRUE) 
{
  clusterindex <- cl[[3]]
  memship <- cl[[4]]
  memship[memship < min.mem] <- -1
  colorindex <- integer(dim(exprs(eset))[[1]])
  if (missing(colo)) {
    colo <- c("#FF8F00", "#FFA700", "#FFBF00", "#FFD700", 
              "#FFEF00", "#F7FF00", "#DFFF00", "#C7FF00", "#AFFF00", 

musically-ut

Scripts to convert datasets from unruly forms to SQLite

The script json2sqlite.py reads files which have one JSON block per new line to be inserted in a database. This has become a rather common dataformat which is easy to incrementally parse. The arguments to the scripts are self explanatory.

Example execution: python json2sqlite.py --gzip aggressive_dedup.json.gz amazon.sqlite reviews

For a more upto date scripts, see: Networks-Leraning/datasets2sqlite.