nanobot

Nanobot is a massively overengineered approach to generating sample, test or demo data.

Quick Start:

// start a simple counter at 0, increment by adding one every step
var numseq = nanobot.Factory.Number.step(1,0,1);

// start a date sequence which skips three days between each value
var timeseq = nanobot.Factory.Date.forwardSeq(3, "day", new Date().getTime());

// create a set of 10 random strings, from which we will draw one member at a time
var vals = nanobot.Factory.String.fromSet(10);

// create a selector which will return a value from a set we supply manually
var ynm = nanobot.Factory.Selector.byIndex(['yes','no','maybe']);

// stuff our generator functions into our template object 
var template = {
    _id: numseq,
    status: ynm,
    name: { 
        first: vals, 
        last: vals 
    },
    created: function() { return "" + timeseq() }
};

// create a function which will execute our template
var maker = nanobot.Clones(template);

// make an army of clones to do our nefarious bidding
for(var i=0; i<100; ++i) {
  var obj = maker();
  console.log(obj);
}

Output

...
{ _id: 81,
  status: 'no',
  name: { first: 'ltlturrri', last: 'wrdaeliot' },
  created: 'Fri Sep 28 2012 18:33:32 GMT-0700 (PDT)' }
{ _id: 82,
  status: 'maybe',
  name: { first: 'wrdaeliot', last: 'wrdaeliot' },
  created: 'Mon Oct 01 2012 18:33:32 GMT-0700 (PDT)' }
...

Get Fancy

The template expansion can be applied inside of generator functions as well, because all overengineered systems need recursion! Here we make the ingredients array a variable size for our simulated smoothies. We could also create objects with variable field populations to simulate sparse attributes.

var fruitnames = ['strawberry','banana','mango','cherry','blueberry','kiwi','wheatgrass','soybeans'];
var fruit = nanobot.Factory.Selector.byIndex(fruitnames);
var names = nanobot.Factory.String.fromChars('bcdaeorst',6);
var arrsz = nanobot.Factory.Number.normalInt(4,3);
var amak = nanobot.Factory.Set.maker(arrsz, fruit);

var template = {
    name: names,
    ingredients: amak
};

var maker = nanobot.Clones(template);
maker();

Output

{ name: 'ooeee',
  ingredients: [ 'cherry', 'blueberry', 'wheatgrass', 'mango' ] }
{ name: 'deaora', 
  ingredients: [ 'wheatgrass' ] }
{ name: 'edcrd',
  ingredients: [ 'kiwi', 'cherry', 'blueberry' ] }
{ name: 'caear',
  ingredients: [ 'kiwi', 'cherry', 'blueberry', 'mango', 'soybeans' ] }

ObjWriter

ObjWriter is a convenience mechanism for creating a set of files, one per generated object. It will spread the generated files across a set of directories to avoid the dreaded 20-minute 'ls':

var maker = nanobot.Clones(template);
var w = new nanobot.ObjWriter(50000, maker);
w.generate();

Output

Done writing, output is in /tmp/objwriter.
Ok: 50020, Errored: 0 time: 35600ms

[nanobot] dan :-)ls /tmp/objwriter/
gen_0/ gen_1/ gen_2/ gen_3/ gen_4/ gen_5/
[nanobot] dan :-)ls /tmp/objwriter/gen_1/|head -n 10
obj_1.json
obj_10003.json
obj_10009.json
...

Correlated Values

Say you need some values which are derived in some way from other values, nanobot.Value.Propagator is your friend!

var timeseq = nanobot.Factory.Date.forwardSeq(3, "day", new Date().getTime());
var wrapper = nanobot.Factory.Value.propagator(timeseq);
var startDate = wrapper.current;
var oneWeekLater = function() { 
   var d = wrapper.current();
   return d.setDate(d.getDate() + 7);
};

var template = { start: startDate, end: oneWeekLater };
var maker = nanobot.Clones(template);

// make an army of clones to do our nefarious bidding
for(var i=0; i<100; ++i) {
  var obj = maker();
  wrapper.next();
  console.log(obj);
}

More Correlated Values

// create a set of 10 random strings, from which we will draw one member at a time
var vals = nanobot.Factory.String.fromSet(10);
// remember the last 10 values we've generated
var wrapper = nanobot.Factory.Value.propagator(vals, 10);

var template = { current: wrapper.next, previous: wrapper.prevN(1), wayOld: wrapper.prevN(7) };
...

Useful Datasets

nanobot.Factory.Lists includes a few functions which can generate useful lists of real-world values for things like people and products. The list values and/or weights are derived from public sources as indicated. Sample:

// straight-up census-weighted data:
var p =require('nanobot').Lists.person();
for(var i=0;i<100;++i) console.log(p);
....
Thomas Rosen
Doris Bates
Joe Merritt
Sharon Jones
Andre Martin

// more fun: combine only surnames!
var p =require('nanobot').Lists.person(true);
for(var i=0;i<100;++i) console.log(p);
....
Salerno Tremblay
Silva Shumaker
Stallings Hughes
Hannah Muscatello
Moore Givens
Hambrick Rocamora
Salgado Vitale
Glasgow Boone
Nichols Ackermann
Croom Gonsalves

// and some products:
var p =require('nanobot').Lists.person(true);
for(var i=0;i<100;++i) console.log(p);
....
Bioguard Master Calcium Hypochlorite
Aqua Kill Insecticide
Care Free Booost
Rid-a-rat Rat & Mouse Killer

You can use your own datasets by passing { value: '', weight: ''} pairs to nanobot.Selector.byWeight