I'm currently trying out the Sails framework and so far I'm quite impressed. One of the odd things I noticed however is that the server returns a 200 OK rather than a 304 Not Modified status code for all records (even when unchanged).
Is there a way to make Sails return a 304 for unmodified records? The reason I'm asking is that this seems to be the best practice and used by some big players like Google and Facebook.
The short answer is yes, you simply have to set the Last-Modified header on your responses.
"Sails is built on Express", which uses fresh (npmjs.org/package/fresh) to compare the request and response headers.
Simple example (based on Sails 0.10.0-rc5):
sails new test304response
cd test304response
sails generate api user -> generates User.js and UserController.js
edit api/models/User.js
module.exports = {
schema: true,
attributes: {
name: {
type: 'string',
required: true
}
}
};
edit api/controllers/UserController.js
module.exports = {
find: function (req, res, next) {
console.log('find:', req.fresh);
User.findOne(req.param('id'), function foundUser(err, user) {
// set the Last-Modified header to the updatedAt time stamp
// from the model
res.set('Last-Modified', user.updatedAt);
res.json(user);
});
},
};
sails lift
go to localhost:1337/user/create?name=Joe -> creates a new User
go to localhost:1337/user/1 -> queries the User with id 1
refresh localhost:1337/user/1 -> queries the same User, Last-Modified wasn't changed
The response has a 304 – Not Modified status (even in the Chrome DevTools, which actually do caching as long as you don't explicitly disable it in the settings).
Disclaimer: I just started to learn sails and also node, so I might have missed a simpler/cleaner solution. I'm also not completely certain, that setting Last-Modified is enough in all cases. However, I had the feeling that you are more interested to know if it is possible and not to get a best practice implementation.
Hope this helps. :)
Related
Following the installation of RestBase using standard config, I have a working version of summary API.
The problem that the caching mechanism seems strange to me.
The piece of code would decide whether to look at a table cache for fast response. But I cannot make it a server-cache depend on some time-constrain (max-age when the cache is written for example). It means that the decision to use cache or not entirely depend on clients.
Can someone explain the workflow of RestBase caching mechanism?
// Inside key.value.js
getRevision(hyper, req) {
//This one get the header from client request and decide to use cache
or not depend on the value. Does it mean server caching is non-existent?
if (mwUtil.isNoCacheRequest(req)) {
throw new HTTPError({ status: 404 });
}
//If should use cache, below run
const rp = req.params;
const storeReq = {
uri: new URI([rp.domain, 'sys', 'table', rp.bucket, '']),
body: {
table: rp.bucket,
attributes: {
key: rp.key
},
limit: 1
}
};
return hyper.get(storeReq).then(returnRevision(req));
}
Cache invalidation is done by the change propagation service, which is triggered on page edits and similar events. Cache control headers are probably set in the Varnish VCL logic. See here for a full Wikimedia infrastructure diagram - it is outdated but gives you the generic idea of how things are wired together.
I'm in the middle of upgrading our API from Sails v0.12 -> v1, which was prompted by the use of self-validating machines for controller actions. After finally getting through a ton of headache replacing deprecated code, I've landed in a rough spot...
With v0.12 (rather, with the older "req, res" controller style), one could use custom response handlers across the board. I've taken advantage of this, and have request logging at the end of all our response types (with some additional sugaring of data). This was done to log all requests in the database, so we can get insights into what our production servers are doing (because they are load-balanced, having a central place to view this is a must, and this was an easy route to take).
So now, my problem is moving forward with "Actions2" machine-style actions. How does one use these custom response types in these things? Are we being forced to repeat ourselves in our exists? I can't find any good documentation to help guide this process, nor can I find a consistent way to "hook" into the end of a response using machines as actions. I can't find any documentation on what kind of options machines can give to Sails.
#Nelson yes, I understand that, but at the time, that isn't what I wanted at all. I wanted all of the benefits of Actions2.
EDIT: While the original, crossed-out comment below does still work, the prefered way to use Actions2 and the custom responses folder paradigm, is to do something similar to the following in an Actions2 file:
module.exports = {
friendlyName: 'Human-friendly name of function',
description: 'Long description of function and what it does.',
inputs: {
userCommand: {
type: 'string',
required: true,
description: 'Long, human-readable description of the input'
}
},
exits: {
success: {
responseType: 'chatbotResponse'
}
},
fn: async function(inputs, exits){
// do some crazy stuff with the inputs, which has already been validated.
return exits.success('Woot');
}
}
This ultimately will route through the responses/chatbotResponse.js, which looks something similar to this:
module.exports = async function chatbotResponse(data){
let res = this.res,
req = this.req;
if (!data) {
data = 'Something didn\'t go as planned...';
}
// how to call a Node Machine style helper with named inputs
await sails.helpers.finalizeRequestLog.with({req: req, res: res, body: {plainString: data}});
return res.json(data);
};
ORIGINAL:
As it turns out, in the Actions2 function, you just need to add the env param async function(inputs, exists, env). The env will give you access to the req and res. So, if you have custom responses, that perform special tasks (like request logging), you can just use return await env.res.customResponse('Hurray, you made a successful call!');
I have a project that caters for individuals with poor internet connections in predominantly rural areas. I need to allow for users to download(or any other applicable means), or fill out details offline and then when they are ready and the internet connection is ready the data filled out offline should sync with the online database and give a report.
The offline form also needs the same validation as online, to ensure no time wastage.
What are the options I know that HTML 5 has an offline application ability. I would prefer an open source option, which will allow people with intermittent internet issues to continue filling out a form or series of forms even though internet has dropped, and the data sync when internet reconnects.
So what are the best options? Having the user requiring to download a large application is also not the best case, I would prefer a browser or small download solution. Maybe even a way of downloading a validatable form in some format for re-upload.
This is something I've been muddling through myself as some of the users of the site I am currently tasked with building have poor connections or would like to fill in forms away from a network for various reasons. Depending on your precise needs and your customer's browser compatibility, the solution I've decided to go with is to use the HTML5 cache capability you mention in your post.
The amount of data stored is not that great, and it will mean that the webpage you want them to fill in is available offline.
If you couple this with the localStorage interface you can keep all form submissions until they regain connection.
As an example of my current solution:
The cache.php file, to write the manifest
<?php
header("Content-Type: text/cache-manifest");
echo "CACHE MANIFEST\n";
$pages = array(
//an array of the pages you want cached for later
);
foreach($pages as $page) {
echo $page."\n";
}
$time = new datetime("now");
//this makes sure that the cache is different when the browser checks it
//otherwise the cache will not be rebuilt even if you change a cached page
echo "#Last Build Time: ".$time->format("d m Y H:i:s T");
You can then have a simple ajax script checking for connection
setInterval( function() {
$.ajax({
url: 'testconnection.php',
type: 'post',
data: { 'test' : 'true' },
error: function(XHR, textStatus, errorThrown) {
if(textStatus === 'timeout') {
//update a global var saying connection is down
noCon = true;
}
}
});
if(hasUnsavedData) {
//using the key/value pairs in localstorage, put together a data object and ajax it into the database
//once complete, return unsavedData to false to prevent refiring this until we have new data
//also using localStorage.removeItem(key) to clear out all localstorage info
}
}, 20000 /*medium gap between calls, do whatever works best for you here*/);
Then for your form submission script, use localstorage if that noCon variable is set to true
$(/*submit button*/).on("click", function(event) {
event.preventDefault();
if(noCon) {
//go through all inputs in some way and put to localstorage, your method is up to you
$("input").each( function() {
var key = $(this).attr("name"), val = $(this).val();
localStorage[key] = val;
});
hasUnsavedData = true;
//update a global variable to let the script above know to save information
} else {
//or if there's connection
$("form").submit();
//submit the form in some manner
}
});
I've not tested every script on this page, but they're written based on the skeleton of what my current solution is doing, minus a lot of error checking etc, so hopefully it will give you some ideas on how to approach this
Suggestions for improvements are welcomed
I have been reading a lot lately, and even more experimenting with web Development. There are some things that I simply cant understand, therefore any help is appreciated.
I am not trying to get my homework done for me. I have some holes in my knowledge, that I desire to fill. Please, help me out with your views :)
REST questions:
Reading documentation this is perfectly understandable (NODE.JS / Express) example:
EXAMPLE ONE (get):
app.get('/', function(req, res) {
res.send('please select a collection, e.g., /collections/messages')
})
My explanation: When the root of the server is hit, send thie following message
EXAMPLE TWO (get):
app.get('/collections/:collectionName/:id', function(req, res) {
req.collection.findOne({name: req.collection.id(req.params.id)},
function(e, result){
if (e) return next(e)
res.send(result)
})
})
My explanation: When the url in hit, take id from the URL (that is located in params.id) and make search based on it (that is MongoDB).
EXAMPLE THREE (post):
app.post('/collections/:collectionName', function(req, res) {
req.collection.insert(req.body, {}, function(e, results){
if (e) return next(e)
res.send(results)
})
})
My explanation: When the URL is hit, take the payload(JSON in this case) that is located in req.body, and insert it as a new document.
Questions:
Are example one and two both RESTfull?
I am now totally confused with params.id. I do understand that POST is transmitted in rew.body... what is params.id? Is it containing URL variables, such as :ID?
My explanations... are they correct?
Example three is also REST, regardless of the fact that POST is used?
Example three, '/collections/:collectionName. Why is the ':collectionName' passed in URL, I could have placed it in req.body as a parameter (along with new data) and take it from there? What is the benefit of doing it?
Thank you
An API must be using HATEOAS to be RESTful. On first example, if / is the entry point of your API, the response should contain links for the available collections, not a human readable string like that. That's definitely not RESTful.
Exactly.
They're OK, except that there's nothing in the third example implying it's a JSON body. It should check for a Content-Type header sent by the client.
REST isn't dependent on HTTP. As long as you're using the HTTP methods as they were standardized, it's fine. POST is the method to use for any action that isn't standardized, so it's fine to use POST for anything, if there isn't a method specific for that. For instance, it's not correct to use POST for retrieval, but it's fine to use it for creating a new resource if you don't have the full representation.
POST means the data body is subordinated to the resource at the target URI. If collectionName were in the POST body, this would mean you were POSTing to /collections, which would make more sense to create a new collection, not a new item of a collection.
I've implemented a REST/CRUD backend by following this article as an example: http://coenraets.org/blog/2012/10/creating-a-rest-api-using-node-js-express-and-mongodb/ . I have MongoDB running locally, I'm not using MongoLabs.
I've followed the Google tutorial that uses ngResource and a Factory pattern and I have query (GET all items), get an item (GET), create an item (POST), and delete an item (DELETE) working. I'm having difficulty implementing PUT the way the backend API wants it -- a PUT to a URL that includes the id (.../foo/) and also includes the updated data.
I have this bit of code to define my services:
angular.module('realmenServices', ['ngResource']).
factory('RealMen', function($resource){
return $resource('http://localhost\\:3000/realmen/:entryId', {}, {
query: {method:'GET', params:{entryId:''}, isArray:true},
post: {method:'POST'},
update: {method:'PUT'},
remove: {method:'DELETE'}
});
I call the method from this controller code:
$scope.change = function() {
RealMen.update({entryId: $scope.entryId}, function() {
$location.path('/');
});
}
but when I call the update function, the URL does not include the ID value: it's only "/realmen", not "/realmen/ID".
I've tried various solutions involving adding a "RealMen.prototype.update", but still cannot get the entryId to show up on the URL. (It also looks like I'll have to build the JSON holding just the DB field values myself -- the POST operation does it for me automatically when creating a new entry, but there doesn't seem to be a data structure that only contains the field values when I'm viewing/editing a single entry).
Is there an example client app that uses all four verbs in the expected RESTful way?
I've also seen references to Restangular and another solution that overrides $save so that it can issue either a POST or PUT (http://kirkbushell.me/angular-js-using-ng-resource-in-a-more-restful-manner/). This technology seems to be changing so rapidly that there doesn't seem to be a good reference solution that folks can use as an example.
I'm the creator of Restangular.
You can take a look at this CRUD example to see how you can PUT/POST/GET elements without all that URL configuration and $resource configuration that you need to do. Besides it, you can then use nested resources without any configuration :).
Check out this plunkr example:
http://plnkr.co/edit/d6yDka?p=preview
You could also see the README and check the documentation here https://github.com/mgonto/restangular
If you need some feature that's not there, just create an issue. I usually add features asked within a week, as I also use this library for all my AngularJS projects :)
Hope it helps!
Because your update uses PUT method, {entryId: $scope.entryId} is considered as data, to tell angular generate from the PUT data, you need to add params: {entryId: '#entryId'} when you define your update, which means
return $resource('http://localhost\\:3000/realmen/:entryId', {}, {
query: {method:'GET', params:{entryId:''}, isArray:true},
post: {method:'POST'},
update: {method:'PUT', params: {entryId: '#entryId'}},
remove: {method:'DELETE'}
});
Fix: Was missing a closing curly brace on the update line.
You can implement this way
$resource('http://localhost\\:3000/realmen/:entryId', {entryId: '#entryId'}, {
UPDATE: {method: 'PUT', url: 'http://localhost\\:3000/realmen/:entryId' },
ACTION: {method: 'PUT', url: 'http://localhost\\:3000/realmen/:entryId/action' }
})
RealMen.query() //GET /realmen/
RealMen.save({entryId: 1},{post data}) // POST /realmen/1
RealMen.delete({entryId: 1}) //DELETE /realmen/1
//any optional method
RealMen.UPDATE({entryId:1}, {post data}) // PUT /realmen/1
//query string
RealMen.query({name:'john'}) //GET /realmen?name=john
Documentation:
https://docs.angularjs.org/api/ngResource/service/$resource
Hope it helps