Tabletop.js takes a Google Spreadsheet and makes it easily accessible through JavaScript. With zero dependencies!
Tabletop.js easily integrates Google Spreadsheets with Backbone.js, Handlebars, and anything else that is hip and cool. It will also help you make new friends and play jazz piano.
function init() {
Tabletop.init( { key: '0AmYzu_s7QHsmdDNZUzRlYldnWTZCLXdrMXlYQzVxSFE',
callback: function(data, tabletop) { console.log(data) },
simpleSheet: true } )
}
Will give you
[ { name: "Carrot", category: "Vegetable", healthiness: "Adequate" },
{ name: "Pork Shoulder", category: "Meat", healthiness: "Questionable" },
{ name: "Bubblegum", category: "Candy", healthiness: "Super High"} ]
Yes, that easy.
You might have seen some instructions on http://builtbybalance.com/Tabletop/, but please ignore them, because they're super super out of date. Probably don't break anything, but they sure ain't current (not my domain, can't take them down). These docs here are the most up-to-date, so treat them as the gospel truth!
Important note! Google's been throwing up weird logins walls every now and again to spreadsheet content, you might want to check out a caching mechanism like Flatware to help you avoid those sorts of situations.
The first step is to get your data out into a form Tabletop can digest
Take a Google Spreadsheet. Give it some column headers, give it some content.
Name Category Healthiness
Carrot Vegetable Adequate
Pork Shoulder Meat Questionable
Bubblegum Candy Super High
In Google Docs, then go up to the File menu and pick Publish to the web. Fiddle with whatever you want, then click Start publishing. A URL will appear, something like https://docs.google.com/spreadsheet/pub?hl=en_US&hl=en_US&key=0AmYzu_s7QHsmdDNZUzRlYldnWTZCLXdrMXlYQzVxSFE&output=html
Copy that! In theory you're interested in the part between key= and & but you can use the whole thing if you want.
Now you're going to feed your spreadsheet into Tabletop
Include Tabletop in your HTML, then try the following, substituting your URL for public_spreadsheet_url
<script type="text/javascript">
window.onload = function() { init() };
var public_spreadsheet_url = 'https://docs.google.com/spreadsheet/pub?hl=en_US&hl=en_US&key=0AmYzu_s7QHsmdDNZUzRlYldnWTZCLXdrMXlYQzVxSFE&output=html';
function init() {
Tabletop.init( { key: public_spreadsheet_url,
callback: showInfo,
simpleSheet: true } )
}
function showInfo(data, tabletop) {
alert("Successfully processed!")
console.log(data);
}
</script>
Open up your console and check out the data that you got. All of those rows were turned right into objects! See how easy that was?
Please don't holdwindow.onload against me, you're free to use $(document).ready and all of that jQuery jazz.
Check out the reference and the examples, but basically you're set. The only thing to think about right now is if you want to deal with multiple sheets you can get rid of simpleSheet: true (more on that later).
You might also be interested in the publishing/republishing/publish-as-it-changes aspects of Google Spreadsheets, but you'll need to google that for any specifics.
The simplest Tabletop initialization works like this
var tabletop = Tabletop.init( { key: public_spreadsheet_url, callback: showInfo } )
With a function living somewhere else called showInfo.
You pass in either key as the actual spreadsheet key, or just the full published-spreadsheet URL. It calls showInfo when done, passing an array of models. Options in general are
key is the key of the published spreadsheet or the URL of the published spreadsheet.
callback is the callback for when the data has been successfully pulled. It will be passed an object containing the models found in the spreadsheet (worksheets => models), and the tabletop instance. Each of these models contains the rows on that worksheet (see Tabletop.Model). If simpleSheet is turned on it simply receives an array of rows of the first worksheet.
simpleSheet can be true or false (default false). It assumes you have one table and you don't care what it's called, so it sends the callback an array of rows instead of a list of models. Peek at the examples for more info.
parseNumbers can be true or false (default false). If true, Tabletop will automatically parse any numbers for you so they don't run around as strings.
postProcess is a function that processes each row after it has been created. Use this to rename columns, compute attributes, etc. See the TimelineSetter example below.
wanted is an array of sheets you'd like to pull. If you have 20 sheets in a public spreadsheet you might as well only pull what you need to access. See the example in simple/multiple.html. Defaults to all.
endpoint is the protocol and domain you'd like to query for your spreadsheet. Defaults to https://spreadsheets.google.com.
singleton assigned the instantiated Tabletop object to Tabletop.singleton, implemented to simplify caching and proxying of requests. Defaults to false.
simple_url, if true, changes all requests to KEY and KEY-SHEET_ID. Defaults to false.
proxy allows you to easily use spreadsheets not located on Google Spreadsheet servers. Setting proxy: "http://www.proxy.com" is equivalent to setting { simple_url: true, singleton: true, endpoint: "http://www.proxy.com" }. Flatware might provide better documentation.
wait prevents tabletop from pulling the Google spreadsheet until you're ready. Used in the backbone.js example.
query sends a structured query along with the spreadsheet request, so you can ask for rows with age > 55 and the like. Right now it's passed with every request, though, so if you're using multiple tables you'll end up in Problem City. It should work great with simpleSheet situations, though.
debug returns on debug mode, which gives you plenty of messaging about what's going on under the hood.
parameterize changes the src of all injected scripts. Instead of src, src is URI encoded and appended to parameterize, e.g. set it to http://example.herokuapp.com/?url=. Mostly for gs-proxy.
callbackContext sets the this for your callback. It's the tabletop object by default.
Once you've initialized a tabletop object you can access its good parts.
.sheets() are the Tabletop.Models that were populated, one per worksheet. You access a sheet by its name.
.sheets(name) is how you access a specific sheet. Say I have a worksheet called Cats I Know, I'll access it via tabletop.sheets("Cats I Know")
.model_names are the names of the models [read: sheets] that Tabletop knows about
.data() returns the rows of the first model if you're in simpleSheet mode. It's the same as .sheets() otherwise. This is just a little sugar.
.fetch() manually initializes a pulling of the data
.addWanted(name) adds a sheet to the list that are updated with .fetch
Tabletop.Model is pretty boring, let's be honest.
.name is the name of the worksheet it came from (the tab at the bottom of the spreadsheet)
.column_names gives you the names of the columns in that table
.all() returns an array of the rows of the table, in the style of [ { name: "Tom", age: 5}, { name: "Liz", age: 12 } ]
.toArray() returns the rows of the table with numbered indices instead of named ones [ [ "Tom", 5] , [ "Liz", 12 ] ]
Imagine it's a read-only, JavaScript CMS that you can edit through Google Docs. It's like Christmas up in here.
Super easy. Just feed the models to Handlebars and you're all set to build the templates.
I've put together a Backbone.tabletopSync driver for Backbone collections and models. It's read-only, but you can't really complain if you're powering your Backbone app through Google Spreadsheets.
Source is, of course, in /src, and you can check it out in action in /examples/backbone/
Tabletop was originally built to work with ProPublica's TimelineSetter, a JS+Ruby library that creates timelines. You need some specifically-formatted JSON which is created by a Ruby script from a CSV, which means your workflow is usually spreadsheet -> CSV -> Ruby -> JSON -> JS.
With Tabletop, though, you get to hook right into a Google Spreadsheet for all of your info! You just need to massage your data a little bit, thanks to Google's API messing with column names and you needing a timestamp.
You can see this in the examples directory, but here are the important parts.
<script type="text/javascript">
var public_spreadsheet_url = "https://docs.google.com/spreadsheet/pub?hl=en_US&hl=en_US&key=0AmYzu_s7QHsmdHk2akhfdG5iTk96OUt6UjJJX1dYMFE&single=true&gid=0&output=html"
$(document).ready( function() {
/*
Need to post-process elements because Google Spreadsheets abbreviates
'display_date' column as 'displaydate' and you need to parse the date
into the 'timestamp' field for TimelineSetter
simpleSheet: true sends array of rows to callback, so you don't need to
do the whole Tabletop.sheets('Sheet1').all() thing.
*/
Tabletop.init( { key: public_spreadsheet_url,
callback: drawTimeline,
simpleSheet: true,
postProcess: function(element) {
element["display_date"] = element["displaydate"];
element["timestamp"] = Date.parse( element["date"] );
}
})
})
function drawTimeline(data, tabletop) {
var currentTimeline = TimelineSetter.Timeline.boot(
data,
{"container":"#timeline","interval":""}
);
}
</script>
See the postProcess call? That's called on every row after Tabletop.Model gets done working on it. It allows you to rename columns or edit data points without having to messily do it outside of Tabletop. I'm sure it has uses outside of TimelineSetter, too.
A sample lives in /examples/timeline_setter/
Yeah, Google Spreadsheets can sometimes be slow or sometimes be overwhelmed or maybe one day Google will just up and disappear on us. So Tabletop.js now supports fetching your data from elsewhere, using options like endpoint and proxy.
proxy is the fun one, in that it rewrites your requests to be simpler-looking and plays nicely with the app & example I put together.
If you don't mind running around with Heroku and AWS, Flatware is an app I built that uploads the spreadsheet JSON response to S3.
gs-proxy is another option that also uses Heroku. You'll set parameterize to something like http://example.herokuapp.com/?url= and off you go!
You can point proxy at anything you'd like as long as it has KEY and KEY-SHEET_ID files sitting in a directory. Feel free to host it on your own server! You can use /caching/local.rb if you want a pretty easy solution for generating the flat files.
Empty tables are trouble. We can't get column names from them (c'mon, Google!), so don't be too confused when a table with 0 rows is coming back with an empty .column_names or your code starts throwing weird errors when processing the results.
Empty rows are trouble. If you have a row that's completely empty, Google doesn't return any rows after the empty row. As a result, you need to make sure every line in your spreadsheet has data in it.
Turn on debugging by passing debug: true when you initialize Tabletop. Check out the console, I try to keep my error messages chatty and informative. Or just email me at [email protected], I'm happy to help!
The more examples the better, right? Feel free to fork or contact me if you have a good example of something you've done.
A contextual video player with popcorn.js by @maboa
The WNYC mayoral tracker uses Tabletop along with Backbone.js
A Facebook-esque timeline from Balance Media (with a git repo)
Mapsheet creates super easy, customizable maps.
If you aren't really feeling Tabletop, you should give Dataset a shot. It's "a JavaScript client-side data transformation and management library," which means it does a hell of a lot more than our dear Tabletop.
Jonathan Soma, who would rather be cooking than coding. Inspired by the relentless demands of John Keefe of WNYC.
Thanks to Scott Seaward for implementing multi-instance Tabletop.
Alan Palazzolo hooked the world up with gs-proxy and added support for it into Tabletop via parameterize
Max Ogden was kind enough to lend Tabletop nodejs support as part of an Open News code sprint