Calling Django `reverse` in client-side Javascript - javascript

I'm using Django on Appengine. I'm using the django reverse() function everywhere, keeping everything as DRY as possible.
However, I'm having trouble applying this to my client-side javascript. There is a JS class that loads some data depending on a passed-in ID. Is there a standard way to not-hardcode the URL that this data should come from?
var rq = new Request.HTML({
'update':this.element,
}).get('/template/'+template_id+'/preview'); //The part that bothers me.

There is another method, which doesn't require exposing the entire url structure or ajax requests for resolving each url. While it's not really beautiful, it beats the others with simplicity:
var url = '{% url blog_view_post 999 %}'.replace (999, post_id);
(blog_view_post urls must not contain the magic 999 number themselves of course.)

Having just struggled with this, I came up with a slightly different solution.
In my case, I wanted an external JS script to invoke an AJAX call on a button click (after doing some other processing).
In the HTML, I used an HTML-5 custom attribute thus
<button ... id="test-button" data-ajax-target="{% url 'named-url' %}">
Then, in the javascript, simply did
$.post($("#test-button").attr("data-ajax-target"), ... );
Which meant Django's template system did all the reverse() logic for me.

The most reasonable solution seems to be passing a list of URLs in a JavaScript file, and having a JavaScript equivalent of reverse() available on the client. The only objection might be that the entire URL structure is exposed.
Here is such a function (from this question).

Good thing is to assume that all parameters from JavaScript to Django will be passed as request.GET or request.POST. You can do that in most cases, because you don't need nice formatted urls for JavaScript queries.
Then only problem is to pass url from Django to JavaScript. I have published library for that. Example code:
urls.py
def javascript_settings():
return {
'template_preview_url': reverse('template-preview'),
}
javascript
$.ajax({
type: 'POST',
url: configuration['my_rendering_app']['template_preview_url'],
data: { template: 'foo.html' },
});

Similar to Anatoly's answer, but a little more flexible. Put at the top of the page:
<script type="text/javascript">
window.myviewURL = '{% url myview foobar %}';
</script>
Then you can do something like
url = window.myviewURL.replace('foobar','my_id');
or whatever. If your url contains multiple variables just run the replace method multiple times.

I like Anatoly's idea, but I think using a specific integer is dangerous. I typically want to specify an say an object id, which are always required to be positive, so I just use negative integers as placeholders. This means adding -? to the the url definition, like so:
url(r'^events/(?P<event_id>-?\d+)/$', events.views.event_details),
Then I can get the reverse url in a template by writing
{% url 'events.views.event_details' event_id=-1 %}
And use replace in javascript to replace the placeholder -1, so that in the template I would write something like
<script type="text/javascript">
var actual_event_id = 123;
var url = "{% url 'events.views.event_details' event_id=-1 %}".replace('-1', actual_event_id);
</script>
This easily extends to multiple arguments too, and the mapping for a particular argument is visible directly in the template.

I've found a simple trick for this. If your url is a pattern like:
"xyz/(?P<stuff>.*)$"
and you want to reverse in the JS without actually providing stuff (deferring to the JS run time to provide this) - you can do the following:
Alter the view to give the parameter a default value - of none, and handle that by responding with an error if its not set:
views.py
def xzy(stuff=None):
if not stuff:
raise Http404
... < rest of the view code> ...
Alter the URL match to make the parameter optional: "xyz/(?P<stuff>.*)?$"
And in the template js code:
.ajax({
url: "{{ url views.xyz }}" + js_stuff,
... ...
})
The generated template should then have the URL without the parameter in the JS, and in the JS you can simply concatenate on the parameter(s).

Use this package: https://github.com/ierror/django-js-reverse
You'll have an object in your JS with all the urls defined in django. It's the best approach I found so far.
The only thing you need to do is add the generated js in the head of your base template and run a management command to update the generated js everytime you add a url

One of the solutions I came with is to generate urls on backend and pass them to browser somehow.
It may not be suitable in every case, but I have a table (populated with AJAX) and clicking on a row should take the user to the single entry from this table.
(I am using django-restframework and Datatables).
Each entry from AJAX has the url attached:
class MyObjectSerializer(serializers.ModelSerializer):
url = SerializerMethodField()
# other elements
def get_url(self, obj):
return reverse("get_my_object", args=(obj.id,))
on loading ajax each url is attached as data attribute to row:
var table = $('#my-table').DataTable({
createdRow: function ( row, data, index ) {
$(row).data("url", data["url"])
}
});
and on click we use this data attribute for url:
table.on( 'click', 'tbody tr', function () {
window.location.href = $(this).data("url");
} );

I always use strings as opposed to integers in configuring urls, i.e.
instead of something like
... r'^something/(?P<first_integer_parameter>\d+)/something_else/(?P<second_integer_parameter>\d+)/' ...
e.g: something/911/something_else/8/
I would replace 'd' for integers with 'w' for strings like so ...
... r'^something/(?P<first_integer_parameter>\w+)/something_else/(?P<second_integer_parameter>\w+)/' ...
Then, in javascript I can put strings as placeholders and the django template engine will not complain either:
...
var url = `{% url 'myapiname:urlname' 'xxz' 'xxy' %}?first_kwarg=${first_kwarg_value}&second_kwarg=${second_kwarg_value}`.replace('xxz',first_integer_paramater_value).replace('xxy', second_integer_parameter_value);
var x = new L.GeoJSON.AJAX(url, {
style: function(feature){
...
and the url will remain the same, i.e something/911/something_else/8/.
This way you avoid the integer parameters replacement issue as string placeholders (a,b,c,d,...z) are not expected in as parameters

Related

Django custom template tags in javascript

I have a custom template tag that returns suppose name of a student and roll number if passed as an argument id of the student.
#register.inclusion_tag('snippet/student_name.html')
def st_name_tag(profile, disp_roll=True, disp_name=True):
#some calculations
return {'full_name':student.name,
'roll':student.roll_number,
}
The template(included) consists of some Html file which is written in a single line(to avoid unterminated string literal error from js).
I simply want to call the st_name_tag from inside the JS function.
My JS looks like:
{% load profile_tag %}
<script type = "text/javascript">
eventclick : function(st){
var div = ('<div></div>');
var st_id = st.id;
if (st.status == 'pass'){
div.append('<p>Student Name:{% st_name_tag '+st_id+' %}</p>');
}
}
So far I tried the above method along with removing the + and '' signs from st_id varaible. That hasnt helped me at all. Help Please!
You are trying to render a template based on the interaction by user. The first happens on the server (server-side as it is often referred to), and the latter happens on the user's browser.
The order that these happen is first to render the template on server, send and present in browser, then user interacts with js. Because of this fact, as I mentioned in the comment, it is not possible to affect the template rendered within javascript.
I would recommend you to use ajax in order to accomplish this. Whenever an iteraction occurs, you asynchronously make a request to the server to present you with new data.

How to handle multiple AJAX behaviors in one HTTP request?

I am using jQuery. I have implemented a multipart web page where a list of links* are rendered and each link is periodically updated through AJAX HTTP requests. That is, on the page there are many links of which each one is "timer-triggered" through JavaScript so to perform a HTTP request to the URL pointed by the link itself and, on response success, to replace those links with the retrieved data (the updated links).
This implementation works but it is "performance less" in cases when the page contains many links: one AJAX request is executed per link resulting in many hits to the server. In order to solve that performance issue I thought to make the JavaScript code to execute a unique AJAX request that retrieves the whole set of links and then to replace DOM data.
However I do not know how to implement the "unique request" mostly due to the practice/technique that I have to use and since it is the first time I notice this kind of problem. What can I do? Should I implement a JavaScript handler for event-registration or what?
* In my case link elements are used (<a></a> HTML tags) but those can be anything associated with a URL.
Update after the jfriend00 answer
If the solution is to build a JSON array as jfriend00 describes in his answer then I should implement the page behavior so to update the JSON array dynamically. Since my HTML links are even rendered dynamically along with some JavaScript code then that JavaScript code could update the JSON array dynamically by "registering"/"unregistering" links. If this is a solution in my case, how can I implement it?
I render links as "partial templates" along with the JavaScript code which JavaScript makes those links to execute AJAX requests. HTML-JS code per each link (the mentioned "partial templates") looks like the following:
<script type="text/javascript">
(function() {
var link = $('#link_1')
...
}());
</script>
It seems like you can just send some JSON that is your array of links to request and then receive JSON back that is an object where each key is the requested link and the data is the server response for that particular link.
If the links you want to process look something like this:
<a class="myLink" href="xxx"></a>
It could look something like this:
function processLinks()
// assuming you can specify some CSS selector to select the links in your page that
// you want to target
// create an array of URLs for the ajax call
// and an index of arrays --> DOM objects so we know which DOM object goes
// with a given URL when processing the ajax results
var urlArray = [];
var urlIndex = {};
var urlArray = $(".templateLink").each(function() {
urlArray.push(this.href);
urlIndex[this.href] = this;
});
$.ajax({
url: "your ajax url here",
data: JSON.stringify(urlArray),
dataType: "json"
}).done(function(data) {
// assumes you get data back as {"url1": data1, "url2": data2, ...}
$.each(data, function(url, urlData) {
// get DOM object that goes with this URL
var domObj = urlIndex[url];
// apply urlData to domObj here
})
});
}
Updating my answer now that you've disclosed your "partial templates".
To process them all at once, change this type of structure which processes them one at a time:
<script>
(function() {
var link = $('#link_1')
...
}());
</script>
<a href="yyy" id="link_2></a>
<script>
(function() {
var link = $('#link_2')
...
}());
</script>
to this which finds them all in the DOM and process them all at once:
<script>
// process all the template links
$(document).ready(processLinks);
</script>

How to get the action name (or view name) in Grails from Javascript or JQuery

I read here, that in the GSP we could write:
${params.action}
From the GSP, we could call a Javascript function with ${params.action} as parameter (see here).
Is there any other way to get the action name without passing parameters from the GSP view?
I understand the desire to reduce application code happening in the views, but to allow the proper components to handle what they handle (separation of concerns), it is the best way to have the controller pass the action name to the view via a model.
The client-side script otherwise has only the url and a loose sense of the route table compared to the server-side which must know exactly the route table.
If you still feel it best to use client-side script, window.location provides the url. To parse the url, javascript's string.split using '/' as the delimiter should suffice, the following snippet should help:
var url = window.location;
var urlParts = url.split('/');
var controller = urlParts[3];
var action = urlParts[4];

How to override variable parameter loaded from another script

I have a script that loads the code dynamically. It is kind of a search engine. When I press a search button, the action gets triggered and a new page opens with many parameters.
I want to override one of the parameters generated with the script in the new URL. JS code is quite big and hard to read, but I have found the important part in the Firebug DOM editor.
This is the pattern of the URL generated when you perform the search:
http://www.example.com/...?ParameterOne=123&ParameterTwo=Two&ThisParameter=Sth&ParameterFour=Four...
What I want to edit is "ThisParameter" and change its value. This is the part edited in the DOM that does what I want:
Foobar = {
_options: [],
...
var options = {"ParameterOne":123,"ParameterTwo":"Two","ThisParameter":"ABC","ParameterFour":Four,...}
...
And this is the output of "ThisParameter" when you choose "Copy path" in Firebug's DOM tab:
_options[0].ThisParameter
I am wondering it this is possible at all. What makes me think that it is, is the fact that I can change this parameter in Firebug and it works perfectly. So, if Firebug can edit it, there should be a way to influence it with another script.
Looking forward to any suggestions, thank you in advance!
Since you cannot edit the dynamic script you have the following options:
You have to try to give the script the correct input and hope it uses your value.
Add a script to the results page which will read the url and arguments, change it and redirect, as we discussed here. (If you put everything in functions it should not conflict with the dynamic script if the functions are uniquely named.)
You could try adding something like this jQuery code to the page with the search button:
$('input[name=search_button_name]').click(function(e) {
e.preventDefault();
var form_search = $('#search_form_id');
$('<input>').attr({
type: 'hidden',
name: 'ThisParameter',
value: 'SomethingElse'
}).appendTo(form_search);
f.submit();
});
You can override any js function and method, or wrap you code around it. The easiest thing would be to look at the code you get and once it gets loaded, you re-declare a method with your own functionality.
I you are trying to replace a parameter in a specific jquery request, you can even wrap around the jquerys ajax method:
var jquery_ajax = $.ajax
$.ajax = function(options){
// parse only a specific occurence
if(options.url.indexOf("example.com") > -1) {
// change the url/params object - depending on where the parameter is
options.params.ThisParameter = "My Custom value"
}
// call the original jquery ajax function
jquery_ajax(options);
}
But it would be a lot cleaner to override the method that builds the ajax request rather than the ajax request itself.
I would investigate further on the scope of the variable options (var options), is it global? i.e. if you type 'options' in the Firebug console, does it display its properties?
If so, you could then access it via your own script and change is value, e.g.
options.ThisParameter = 'my-own-value';
You might hook your script to the click event of the search button.
I hope this helps, it could be more specific maybe if you have some sample code somewhere.

Extracting part of current page's URL in javascript

In ASP.NET MVC 3 site, I open the following uri
http://myserver/incidents/supercompany/register
That page needs to make use of ajax JSON retrieval mechanism with the help of jQuery:
$.getJSON('/Incidents/[[COMPANY NAME GOES HERE]]/SearchPeople/' + request.term, function (data)
I am new to javascript. How can I obtain supercompany part of my current url to build the
/Incidents/supercompany/SearchPeople/ string?
Assuming your URLs follow a logical structure, you can do this
var URLparts = window.location.toString().substr(7).split('/');
And you can now access URLparts[1] to get the company name. Obviously, you need to be able to know that that will ALWAYS be there, but it's a solution if you can guarantee that (eg. with htaccess).

Categories

Resources