dfm

This will change your life. Add this as a git pre-push hook.

How, you ask?

wget https://gist.github.com/dfm/5828028/raw/772270b65a804678db70012d3406f41c79cb53a5/pre-push
chmod +x pre-push
mv pre-push /path/to/your/favorite/repo/.git/hooks
cd /path/to/your/favorite/repo
git push

dfm

import flask

def create_app(config_filename=None, ...):
    app = flask.Flask(__name__)

    # Default settings.
    app.config.from_object("myapp.default_settings")

    # If a filename was provided, load the requested settings from a file.
    if config_filename is not None:

dfm

import json
import urllib2

url = "http://api-dev.sdss3.org/spectrumQuery"
payload = {"ra": ["227.69425d", "227.85043d", "236.84261d"],
           "dec": [26.94044, 26.40779, 1.38737]}
# or...
# payload = {"ra": ["227.69425d", "227.85043d", "236.84261d"],
#            "dec": [26.94044, 26.40779, 1.38737],
#            "radius": [3., 3., 3.]}

dfm

#!/usr/bin/env python
# -*- coding: utf-8 -*-

from __future__ import division, print_function
import emcee
import numpy as np
import matplotlib.pyplot as pl

np.random.seed(123)

dfm

#!/usr/bin/env python
# -*- coding: utf-8 -*-

import kplr
from urllib2 import HTTPError

# Choose a star.
client = kplr.API()
kic = client.star(10592770)
print("Kepler magnitude: {0}".format(kic.kic_kepmag))

dfm

#!/usr/bin/env python
# -*- coding: utf-8 -*-

import json
import gzip
import redis
import random
import requests
from StringIO import StringIO
from datetime import datetime, timedelta

dfm

import emcee
import numpy as np

sampler = emcee.EnsembleSampler(...)
pos, lnprob, rstate = sampler.run_mcmc(p0, 1000)

nindependent = 0
while nindependent < 500:
    pos, lnprob, rstate = sampler.run_mcmc(pos, 1000, rstate0=state, lnprob0=lnprob)
    try:

dfm

I'm doing an experiment in natural language processing (NLP) to study models for discovery of (personally) relevant scientific literature and I need your bibliographies! The immediate motivation for this experiment is a class that I'm taking at NYU this semester but I've been interested in this problem for a while and I expect this research to extend past this semester. The basic idea is to take the content of all the astro-ph articles on the arXiv and combine that with user data to generate recommendations of articles that readers might be interested in looking at. The model is based on work by Wang & Blei (2011) but generalized to use a more sophisticated author-topic (AT) model based on Rosen-Zvi, et al. (2010). This is relevant for astro-ph because the user dataset is neccessarily

dfm

#include <math.h>

double params[] = {1.0e-3, 1.0, 5.0};

function kernel (double t1, double t2)
{
    double r, omr,
           d = x1 - x2,
           chi2 = d * d,
           p2 = params[2] * params[2],

dfm