Peterbe.com

When you use a web framework like Tornado, which is single threaded with an event loop (like nodejs familiar with that), and you need persistency (ie. a database) there is one important questions you need to ask yourself:

Is the query fast enough that I don't need to do it asynchronously?

If it's going to be a really fast query (for example, selecting a small recordset by key (which is indexed)) it'll be quicker to just do it in a blocking fashion. It means less CPU work to jump between the events.

However, if the query is going to be potentially slow (like a complex and data intensive report) it's better to execute the query asynchronously, do something else and continue once the database gets back a result. If you don't all other requests to your web server might time out.

Another important question whenever you work with a database is:

Would it be a disaster if you intend to store something that ends up not getting stored on disk?

This question is related to the D in ACID and doesn't have anything specific to do with Tornado. However, the reason you're using Tornado is probably because it's much more performant that more convenient alternatives like Django. So, if performance is so important, is durable writes important too?

Let's cut to the chase... I wanted to see how different databases perform when integrating them in Tornado. But let's not just look at different databases, let's also evaluate different ways of using them; either blocking or non-blocking.

What the benchmark does is:

• On one single Python process...
• For each database engine...
• Create X records of something containing a string, a datetime, a list and a floating point number...
• Edit each of these records which will require a fetch and an update...
• Delete each of these records...

I can vary the number of records ("X") and sum the total wall clock time it takes for each database engine to complete all of these tasks. That way you get an insert, a select, an update and a delete. Realistically, it's likely you'll get a lot more selects than any of the other operations.

And the winner is:

pymongo!! Using the blocking version without doing safe writes.

Let me explain some of those engines

• pymongo is the blocking pure python engine
• with the redis, toredis and memcache a document ID is generated with uuid4, converted to JSON and stored as a key
• toredis is a redis wrapper for Tornado
• when it says (safe) on the engine it means to tell MongoDB to not respond until it has with some confidence written the data
• motor is an asynchronous MongoDB driver specifically for Tornado
• MySQL doesn't support arrays (unlike PostgreSQL) so instead the tags field is stored as text and transformed back and fro as JSON
• None of these database have been tuned for performance. They're all fresh out-of-the-box installs on OSX with homebrew
• None of these database have indexes apart from ElasticSearch where all things are indexes
• momoko is an awesome wrapper for psycopg2 which works asyncronously specifically with Tornado
• memcache is not persistant but I wanted to include it as a reference
• All JSON encoding and decoding is done using ultrajson which should work to memcache, redis, toredis and mysql's advantage.
• mongokit is a thin wrapper on pymongo that makes it feel more like an ORM
• A lot of these can be optimized by doing bulk operations but I don't think that's fair
• I don't yet have a way of measuring memory usage for each driver+engine but that's not really what this blog post is about
• I'd love to do more work on running these benchmarks on concurrent hits to the server. However, with blocking drivers what would happen is that each request (other than the first one) would have to sit there and wait so the user experience would be poor but it wouldn't be any faster in total time.
• I use the official elasticsearch driver but am curious to also add Tornado-es some day which will do asynchronous HTTP calls over to ES.

You can run the benchmark yourself

The code is here on github. The following steps should work:

$ virtualenv fastestdb
$ source fastestdb/bin/activate
$ git clone https://github.com/peterbe/fastestdb.git
$ cd fastestdb
$ pip install -r requirements.txt
$ python tornado_app.py

Then fire up http://localhost:8000/benchmark?how_many=10 and see if you can get it running.

Note: You might need to mess around with some of the hardcoded connection details in the file tornado_app.py.

Discussion

Before the lynch mob of HackerNews kill me for saying something positive about MongoDB; I'm perfectly aware of the discussions about large datasets and the complexities of managing them. Any flametroll comments about "web scale" will be deleted.

I think MongoDB does a really good job here. It's faster than Redis and Memcache but unlike those key-value stores, with MongoDB you can, if you need to, do actual queries (e.g. select all talks where the duration is greater than 0.5). MongoDB does its serialization between python and the database using a binary wrapper called BSON but mind you, the Redis and Memcache drivers also go to use a binary JSON encoding/decoder.

The conclusion is; be aware what you want to do with your data and what and where performance versus durability matters.

What's next

Some of those drivers will work on PyPy which I'm looking forward to testing. It should work with cffi like psycopg2cffi for example for PostgreSQL.

Also, an asynchronous version of elasticsearch should be interesting.

UPDATE 1

Today I installed RethinkDB 2.0 and included it in the test.

It was added in this commit and improved in this one.

I've been talking to the core team at RethinkDB to try to fix this.

My colleague Axel Hecht showed me something I didn't know about sorting in Python.

In Python you can sort with a tuple. It's best illustrated with a simple example:

>>> items = [(1, 'B'), (1, 'A'), (2, 'A'), (0, 'B'), (0, 'a')]
>>> sorted(items)
[(0, 'B'), (0, 'a'), (1, 'A'), (1, 'B'), (2, 'A')]

By default the sort and the sorted built-in function notices that the items are tuples so it sorts on the first element first and on the second element second.

However, notice how you get (0, 'B') appearing before (0, 'a'). That's because upper case comes before lower case characters. However, suppose you wanted to apply some "humanization" on that and sort case insensitively. You might try:

>>> sorted(items, key=str.lower)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: descriptor 'lower' requires a 'str' object but received a 'tuple'

which is an error we deserve because this won't work for the first part of each tuple.

We could try to write a lambda function (e.g. sorted(items, key=lambda x: x.lower() if isinstance(x, str) else x)) but that's not going to work because you'll only ever get to apply that to the first item.

Without further ado, here's how you do it. A lambda function that returns a tuple:

>>> sorted(items, key=lambda x: (x[0], x[1].lower()))
[(0, 'a'), (0, 'B'), (1, 'A'), (1, 'B'), (2, 'A')]

And there you have it! Thanks for sharing Axel!

As a bonus item for people still reading...
I'm sure you know that you can reverse a sort order simply by passing in sorted(items, reverse=True, ...) but what if you want to have different directions depend on the key that you're sorting on.

Using the technique of a lambda function that returns a tuple, here's how we sort a slightly more advanced structure:

>>> peeps = [{'name': 'Bill', 'salary': 1000}, {'name': 'Bill', 'salary': 500}, {'name': 'Ted', 'salary': 500}]

And now, sort with a lambda function returning a tuple:

>>> sorted(peeps, key=lambda x: (x['name'], x['salary']))
[{'salary': 500, 'name': 'Bill'}, {'salary': 1000, 'name': 'Bill'}, {'salary': 500, 'name': 'Ted'}]

Makes sense, right? Bill comes before Ted and 500 comes before 1000. But how do you sort it like that on the name but reverse on the salary? Simple, negate it:

>>> sorted(peeps, key=lambda x: (x['name'], -x['salary']))
[{'salary': 1000, 'name': 'Bill'}, {'salary': 500, 'name': 'Bill'}, {'salary': 500, 'name': 'Ted'}]

UPDATE

Webucator has made a video explaining this blog post as a video.

Thanks Nat!

Thanks to Igor who emailed me and made me aware, you can't put pseudo classes in style attributes in HTML. I.e. this does not work:

<a href="#" style="color:pink :hover{color:red}">Sample Link</a>

See for yourself: Sample Link

Note how it does not become red when you hover over the link above.
This is what premailer used to do. Until yesterday.

BEFORE:

>>> from premailer import transform
>>> print transform('''
... <html>
... <style>
... a { color: pink }
... a:hover { color: red }
... </style>
... <a href="#">Sample Link</a>
... </html>
... ''')
<html><head><a href="#" style="{color:pink} :hover{color:red}">Sample Link</a></head></html>

AFTER:

>>> from premailer import transform
>>> print transform('''
... <html>
... <style>
... a { color: pink }
... a:hover { color: red }
... </style>
... <a href="#">Sample Link</a>
... </html>
... ''')
<html><head>
<style>a:hover {color:red}</style>
<a href="#" style="color:pink">Sample Link</a>
</head></html>

That's because the new default is exclude pseudo classes by default.

Thanks Igor for making me aware!

So I've now built my first real application using AngularJS. It's a fun side-project which my wife and I use to track what we spend money on. It's not a work project but it's also not another Todo list application. In fact, the application existed before as a typical jQuery app. So, I knew exactly what I needed to build but this time trying to avoid jQuery as much as I possibly could.

The first jQuery based version is here and although I'm hesitant to share this beginner-creation here's the AngularJS version

The following lists were some stumbling block and other things that stumped me. Hopefully by making this list it might help others who are also new to AngularJS and perhaps the Gods of AngularJS can see what confuses beginners like me.

1. AJAX doesn't work like jQuery

Similar to Backbone, I think, the default thing is to send the GET, POST request with the data the body blob. jQuery, by default, sends it as application/x-www-form-urlencoded. I like that because that's how most of my back ends work (e.g. request.GET.get('variable') in Django). I ended up pasting in this (code below) to get back what I'm familiar with:

module.config(function ($httpProvider) {
  $httpProvider.defaults.transformRequest = function(data){
    if (data === undefined) {
      return data;
    }
    return $.param(data);
  };
  $httpProvider.defaults.headers.post['Content-Type'] = ''
    + 'application/x-www-form-urlencoded; charset=UTF-8';
});

2. App/Module configuration confuses me

The whole concept of needing to define the code as an app or a module confused me. I think it all starts to make sense now. Basically, you don't need to think about "modules" until you start to split distinct things into separate files. To get started, you don't need it. At least not for simple applications that just have one chunk of business logic code.

Also, it's confusing why the the name of the app is important and why I even need a name.

3. How to do basic .show() and .hide() is handled by the data

In jQuery, you control the visibility of elements by working with the element based on data. In AngularJS you control the visibility by tying it to the data and then manipulate the data. It's hard to put your finger on it but I'm so used to looking at the data and then decide on elements' visibility. This is not an uncommon pattern in a jQuery app:

<p class="bench-press-question">
  <label>How much can you bench press?</label>
  <input name="bench_press_max">
</p>

if (data.user_info.gender == 'male') {
  $('.bench-press-question input').val(response.user_info.bench_press_max);
  $('.bench-press-question').show();
}

In AngularJS that would instead look something like this:

<p ng-show="male">
  <label>How much can you bench press?</label>
  <input name="bench_press_max" ng-model="bench_press_max">
</p>

if (data.user_info.gender == 'male') {
  $scope.male = true;
  $scope.bench_press_max = data.user_info.bench_press_max;
}

I know this can probably be expressed in some smarter way but what made me uneasy is that I mix stuff into the data to do visual things.

4. How do I use controllers that "manage" the whole page?

I like the ng-controller="MyController" thing because it makes it obvious where your "working environment" is as opposed to working with the whole document but what do I do if I need to tie data to lots of several places of the document?

To remedy this for myself I created a controller that manages, basically, the whole body. If I don't, I can't manage scope data that is scattered across totally different sections of the page.

I know it's a weak excuse but the code I ended up with has one massive controller for everything on the page. That can't be right.

5. setTimeout() doesn't quite work as you'd expect

If you do this in AngularJS it won't update as you'd expect.

<p class="status-message" ng-show="message">{{ message }}</p>

$scope.message = 'Changes saved!';
setTimout(function() {
  $scope.message = null;
}, 5 * 1000);

What you have to do, once you know it, is this:

function MyController($scope, $timeout) {
  ...
  $scope.message = 'Changes saved!'; 
  $timeout(function() {
    $scope.message = null;
  }, 5 * 1000);
}

It's not too bad but I couldn't see this until I had Googled some Stackoverflow questions.

6. Autocompleted password fields don't update the scope

Due to this bug when someone fills in a username and password form using autocomplete the password field isn't updating its data.

Let me explain; you have a username and password form. The user types in her username and her browser automatically now also fills in the password field and she's ready to submit. This simply does not work in AngularJS yet. So, if you have this code...:

<form>
<input name="username" ng-model="username" placeholder="Username">
<input type="password" name="password" ng-model="password" placeholder="Password">
<a class="button button-block" ng-click="submit()">Submit</a>
</form>

$scope.signin_submit = function() {
  $http.post('/signin', {username: $scope.username, password: $scope.password})
    .success(function(data) {
      console.log('Signed in!');
    };
  return false;
};

It simply doesn't work! I'll leave it to the reader to explore what available jQuery-helped hacks you can use.

7. Events for selection in a <select> tag is weird

This is one of those cases where readers might laugh at me but I just couldn't see how else to do it.
First, let me show you how I'd do it in jQuery:

$('select[name="choice"]').change(function() {
  if ($(this).val() == 'other') {
    // the <option value="other">Other...</option> option was chosen 
  }
});

Here's how I solved it in AngularJS:

$scope.$watch('choice', function(value) {
  if (value == 'other') {
    // the <option value="other">Other...</option> option was chosen 
  }
});

What's also strange is that there's nothing in the API documentation about $watch.

8. Controllers "dependency" injection is, by default, dependent on the controller's arguments

To have access to modules like $http and $timeout for example, in a controller, you put them in as arguments like this:

function MyController($scope, $http, $timeout) { 
  ...

It means that it's going to work equally if you do:

function MyController($scope, $timeout, $http) {  // order swapped
  ...

That's fine. Sort of. Except that this breaks minification so you have to do it this way:

var MyController = ['$scope', '$http', '$timeout', function($scope, $http, $timeout) {
  ...

Ugly! The first form depends on the interpreter inspecting the names of the arguments. The second form depends on the modules as strings.

The more correct way to do it is using the $inject. Like this:

MyController.$inject = ['$scope', '$http', '$timeout'];
function MyController($scope, $http, $timeout) {
  ...

Still ugly because it depends on them being strings. But why isn't this the one and only way to do it in the documentation? These days, no application is worth its salt if it isn't minify'able.

9. Is it "angular" or "angularjs"?

Googling and referring to it "angularjs" seems to yield better results.

This isn't a technical thing but rather something that's still in my head as I'm learning my way around.

In conclusion

I'm eager to write another blog post about how fun it has been to play with AngularJS. It's a fresh new way of doing things.

AngularJS code reminds me of the olden days when the HTML no longer looks like HTML but instead some document that contains half of the business logic spread all over the place. I think I haven't fully grasped this new way of doing things.

From hopping around example code and documentation I've seen some outrageously complicated HTML which I'm used to doing in Javascript instead. I appreciate that the HTML is after all part of the visual presentation and not the data handling but it still stumps me every time I see that what used to be one piece of functionality is now spread across two places (in the javascript controller and in the HTML directive).

I'm not givin up on AngularJS but I'll need to get a lot more comfortable with it before I use it in more serious applications.

This took me by surprise today!

If you run this unit test, it actually passes with flying colors:

import unittest


class BadAssError(TypeError):
    pass


def foo():
    raise BadAssError("d'oh")


class Test(unittest.TestCase):

    def test(self):
        self.assertRaises(BadAssError, foo)
        self.assertRaises(TypeError, foo)
        self.assertRaises(Exception, foo)


if __name__ == '__main__':
    unittest.main()

Basically, assertRaises doesn't just take the exception that is being raised and accepts it, it also takes any of the raised exceptions' parents.

I've only tested it with Python 2.6 and 2.7. And the same works equally with unittest2.

I don't really know how I feel about this. It did surprise me when I was changing one of the exceptions and expected the old tests to break but they didn't. I mean, if I want to write a test that really makes sure the exception really is BadAssError it means I can't use assertRaises().

Thanks to Theo Spears awesome effort premailer now support CSS specificity. What that means is that when linked and inline CSS blocks are transformed into tag style attributes the order is preserved as you'd expect.

When the browser applies CSS to elements it does it in a specific order. For example if you have this CSS:

p.tag { color: blue; }
p { color: red; }

and this HTML:

<p>Regular text</p>
<p class="tag">Special text</p>

the browser knows to draw the first paragraph tag in red and the second paragraph in red. It does that because p.tag is more specific that p.

Before, it would just blindly take each selector and set it as a style tag like this:

<p style="color:red">Regular text</p>
<p style="color:blue; color:red" class="tag">Special text</p>

which is not what you want.

The code in action is here.

Thanks Theo!

If you use django-fancy-cache you can either run with stats or without. With stats, you can get a number of how many times a cache key "hits" and how many times it "misses". Keeping stats incurs a small performance slowdown. But how much?

I created a simple page that either keeps stats or ignores it. I ran the benchmark over Nginx and Gunicorn with 4 workers. The cache server is a memcached running on the same host (my OSX 10.7 laptop).

With stats:

Average: 768.6 requests/second
Median: 773.5 requests/second
Standard deviation: 14.0

Without stats:

Average: 808.4 requests/second
Median: 816.4 requests/second
Standard deviation: 30.0

That means, roughly that running with stats incurs a 6% slower performance.

The stats is completely useless to your users. The stats tool is purely for your own curiousity and something you can switch on and off easily.

Note: This benchmark assumes that the memcached server is running on the same host as the Nginx and the Gunicorn server. If there was more network in between, obviously all the .incr() commands would cause more performance slowdown.

This personal blog site of mine uses django-fancy-cache and mincss.

What that means is that I can cache the whole output of every blog post for weeks and when I do that I can first preprocess the HTML and convert every external CSS into one inline STYLE block which will only reference selectors that are actually used.

To see it in action, right-click and select "View Page Source". You'll see something like this:

/*
Stats about using github.com/peterbe/mincss
-------------------------------------------
Requests:         1 (now: 0)
Before:           81Kb
After:            11Kb
After (minified): 11Kb
Saving:           70Kb
*/
section{display:block}html{font-size:100%;-webkit-text-size-adjust:100%;-ms-tex...

The reason the saving is so huge, in my case, is because I'm using Twitter Bootstrap CSS framework which is awesome but as any framework, it will inevitably contain a bunch of stuff that I don't use. Some stuff I don't use on any page at all. Some stuff is used only on some pages and some other stuff is used only on some other pages.

What I gain by this, is faster page loads. What the browser does is that it, gets a URL, downloads all HTML, opens the HTML to look for referenced CSS (using the link tag) and downloads that too. Once all of that is downloaded, it starts to render the page. Approximately after that it starts to download all referenced Javascript and starts evaluating and executing that.

By not having to download the CSS the browser has one less thing to do. Only one request? Well, that request might be on a CDN (not a great idea actually) so even though it's just 1 request it will involve another DNS look-up.

Here's what the loading of the homepage looks like in Firefox from a US east coast IP.

Granted, a downloaded CSS file can be cached by the browser and used for other pages under the same domain. But, on my blog the bounce rate is about 90%. That doesn't necessarily mean that visitors leave as soon as they arrived, but it does mean that they generally just read one page and then leave. For those 10% of visitors who visit more than one page will have to download the same chunk of CSS more than once. But mind you, it's not always the same chunk of CSS because it's different for different pages. And the amount of CSS that is now in-line only adds about 2-3Kb on the HTML load when sent gzipped.

Getting to this point wasn't easy because I first had to develop mincss and django-fancy-cache and integrate it all. However, what this means is that you can have it done on your site too! All the code is Open Source and it's all Python and Django which are very popular tools.

** A Django cache_page on steroids**

Django ships with an awesome view decorator called cache_page which is awesome. But a bit basic too.

What it does is that it stores the whole view response in memcache and the key to it is the URL it was called with including any query string. All you have to do is specify the length of the cache timeout and it just works.
Now, it's got some shortcomings which django-fancy-cache upgrades. These "steroids" are:

1. Ability to override the key prefix with a callable.
2. Ability to remember every URL that was cached so you can do invalidation by a URL pattern.
3. Ability to modify the response before it's stored in the cache.
4. Ability to ignore certain query string parameters that don't actually affect the view but does yield a different cache key.
5. Ability to serve from cache but always do one last modification to the response.
6. Incrementing counter of every hit and miss to satisfy your statistical curiosity needs.

The documentation is here:
https://django-fancy-cache.readthedocs.org/

You can see it in a real world implementation by seeing how it's used on my blog here. You basically use it like this::

from fancy_cache import cache_page

@cache_page(60 * 60)
def myview(request):
    ...
    return render(request, 'template.html', stuff)

What I'm doing with it here on my blog is that I make the full use of caching on each blog post but as soon as a new comment is posted, I wipe the cache by basically creating a new key prefix. That means that pages are never cache stale but the views never have to generate the same content more than once.

I'm also using django-fancy-cache to do some optimizations on the output before it's stored in cache.

Remember mincss from last month? Well, despite it's rather crazy version number has only really had one major release. And it's never really been optimized.

So I took some metrics and was able to find out where all the time is spent. It's basically in this:

for body in bodies:
    for each in CSSSelector(selector)(body):
        return True

That in itself, on its own, is very fast. Just a couple of milliseconds. But the problem was that it happens so god damn often!

So, in version 0.8 it now, by default, first make a list (actually, a set) of every ID and every CLASS name in every node of every HTML document. Then, using this it gingerly tries to avoid having to use CSSSelector(selector) if the selector is quite simple. For example, if the selector is #container form td:last-child and if there is no node with id container then why bother.
It equally applies the same logic to classes.

And now, what you've all been waiting for; the results:

On a big document (20Kb) like my home page...

1. BEFORE: 4.7 seconds

2. AFTER: 0.85 seconds

(I ran it a bunch of times and averaged the times which had very little deviation)

So in the first round of optimization it suddenly becomes 500% faster. Pretty cool!

I've made it possible to switch this off just because I haven't yet tested it on equally many sites. All the unit tests pass of course.