amontalenti

import re

from pymongo import MongoReplicaSetClient
from pymongo.read_preferences import ReadPreference

REPLICA_SET = "ptrack-mongo1" # or, parsely_articles for old CrawlDB

conn = MongoReplicaSetClient(read_preference=ReadPreference.SECONDARY, replicaSet=REPLICA_SET)

# optional collection filter for crawlDB in particular

amontalenti

database = "parsely_articles"
collections = range(10) # populate with collections similarly to before
script_cmds = []
for collection in collections:
    cmd =\
    "echo restoring {db}/{col}"\
    " && "\
    "lbzip2 --decompress {db}-{col}.bson.bz2"\
    " && "\
    "mongorestore --d {db} -c {col} {db}-{col}.bson".format(

amontalenti

#!/usr/bin/env python

#
# simple static Flask fileserving app
# with livereload integration
#
from flask import Flask

STATIC_FOLDER = "."

amontalenti

<script>
/*
 * $sf.ext.meta looks up metadata from parent window. Here's
 * the relevant documentation from the SafeFrame standard:
 * 
 * "Use to retrieve metadata about the SafeFrame position that 
 * was specified by the host. The host may specify additional
 * metadata about this 3rd party content. The host specifies
 * this metadata using the $sf.host.PosMeta class."
 *

amontalenti

function getCurrentDateString() {
    // JavaScript Date API does not return zero-padded date strings, so we need this utility
    var pad = function(number) {
      if (number < 10) {
        return '0' + number;
      }
      return number;
    };
    // get the current day and format it as YYYY-MM-DD-HH
    var date = new Date(),

amontalenti

"""urltools.py - parse and format web URLs.

HINT:

>>> "http://google.com".split("://")
["http", "google"]

>>> "google.com/hangout/parsely.com/am".split("/")
["google.com", "hangout", "parsely.com", "am"]

amontalenti

Hello again, growing Pythonistas!

Pleased and excited

So, first of all, I have to say that I was simultaneously pleased & impressed with the quality of submissions I got from you guys for our little Python take-home assignment. Pleased, because the problem seemed to be accessible enough that each of you could work on it with little background in Python beyond the bits you've been exposed to over the last few months. Impressed, because all of your submissions passed all of my test cases!

I was also happy with how different the submissions were. We Pythonistas often like to brag that, unlike in other languages, in Python there is "rarely more than one way to do it". However, this was a simple example, yet the solutions provided varied widely. (And indeed, in programming, no matter how simple the language, there's always more than one way to do it.) This therefore gives us a nice window into various understandings of Python code style, architecture, and idioms.

Patterns among code

amontalenti

from nltk import FreqDist
from nltk.corpus import stopwords
from nltk import wordpunct_tokenize

# my little NLP utility library
import nlp2

import pandas as pd

df = pd.read_csv("hook_letters.csv")

amontalenti

from sst.actions import *
import time
import json

from settings import DASH_USERNAME, DASH_PASSWORD, APIKEYS

envs = {
    "bri": "dash.parsely.com",
    "ue1a": "ue1a-dash-web1.cogtree.com",
}

amontalenti

import time
import random
import requests

words = [word.strip() for word in open("/usr/share/dict/words")]
names = ["John", "Bob", "Peter", "Joe", "Sarah", "Clare"]
sections = ["politics", "entertainment", "life", "sports"]
apikeys = ["arstechnica_com", "foxnews_com"]

def measured_query(apikey, q):