meg-codes

# -*- coding: utf-8 -*-
import pandas as pd
import LatLon as latlon
import re

from LatLon import string2latlon

# Requires: pandas, LatLon, py2.7

def makefloatlat(data):

meg-codes

#!/usr/bin/env python
"""
Python script to list all MySQL databases and allow them to be recreated.
This assumes you have server information and credentials stored in a ~/.my.cnf
file (ideally chmoded to 0600)

[client]
host=localhost
user=root
password=secretpasswordgoeshere

meg-codes

Installing QuantLib

Follow the instructions here to install Quantlib.

Make especially sure that you use the --with-intraday flag for QuantLib.

Install an up-to-date clang with OpenMP support

The default MacOS clang++ for High Sierra does not include support for OpenMP, so you'll need to install a version that does and then set R to use it for compiling libraries.

meg-codes

On the Adroit headnode

This requests an allocation on a compute node of Adroit for 15 minutes. Once granted you can see I'm moved to the compute node adroit-14. I also ask for two tasks (you can adjust based on whether you actually use the cores or not.)

[bhicks@adroit4 ~]$ salloc -N 1 --ntasks-per-node=2 -t 15:00
salloc: Granted job allocation 497989
[bhicks@adroit-14 ~]$ export XDG_RUNTIME_DIR=""

meg-codes

#!/usr/bin/env python

# Script to scrape all links from a site, compile counts of each link, status
# codes of access and output the results as a CSV
#
# There's absolutely no reason this shouldn't be pulled into an OOP paradigm
# per se, but I left it functionalized because that can be easier for multitasking.
#
# Requirements:
# requests, bs4

meg-codes

import pickle

import tensorflow as tf
from tensorflow import keras

import numpy as np

import matplotlib.pyplot as plt

# reload pickled data

meg-codes

import pickle
from tensorflow import keras

# Compute nodes don't have internet, so download in advance and pickle the
# image training data

# adapted from https://www.tensorflow.org/tutorials/keras/basic_classification
# see notes there for how to actually work things

fashion_mnist = keras.datasets.fashion_mnist

meg-codes

#!/bin/bash
#SBATCH --gres=gpu:1 
#one GPU, asking for only one CPU core since we're offloading to GPU
#SBATCH -t 00:05:00
#SBATCH --mail-type=begin
#SBATCH --mail-type=end
#SBATCH --mail-user=yourNETID@princeton.edu

module load anaconda3
# assumes you have a conda environment named tf with tensorflow and matplotlib, i.e.

meg-codes

const webpackConf = require('./webpack.config.js')
const merge = require('webpack-merge')

module.exports = function (config) {
  config.set({
    browsers: [ 'ChromeHeadless' ], //run in Chrome
    browserNoActivityTimeout: 60000,
    frameworks: [ 'mocha', 'sinon-chai' ], //use the mocha test framework
    files: [
      //'src/**/*.{jsx,js}',

meg-codes

// path and webpack
const path = require('path')
const webpack = require('webpack')

// webpack plugins
const BundleTracker = require('webpack-bundle-tracker')
const VueLoaderPlugin = require('vue-loader/lib/plugin')
const MiniCssExtractPlugin = require('mini-css-extract-plugin')

// configure dev mode because of absurdly misleading webpack documentation