I think someone must have run into this situation before. Basically I have a big "form" which is composed of multiple smaller "forms" inside. (In fact, they are not real forms, just sets of inputs that are grouped together to collect info for models).
This form is for a checkout page, which contains:
shipping address
shipping method
billing address
billing method
other additional info such as discounts code input, gift wrapping etc.
I would like to update the user filled info to the server as soon as they complete each part (for example, when they complete shipping address). However, I want to make it work seamlessly without the need for the users to click some kind of "update" button after filling each partial part. I wonder if there is some way to go around this?
You'll want to $watch the fields in question and act upon them (say save to db) when they are filled in. The issue you will run into is how to determine when a user has filled fields in. Things like onblur etc don't work very well in practice. I would recommend using what is called a debounce function which is basically a function that allows the user to pause for X amount of time without our code going "ok done! now let's.. ohh wait still typing..."
Here's an example that I use on my own cart - I want to automatically get shipping quotes once I have an address so I watch these fields, allow some pausing with my debounce function then call my server for quotes.
Here's some controller code:
// Debounce function to wait until user is done typing
function debounce(fn, delay) {
var timer = null;
return function() {
var context = this,
args = arguments;
clearTimeout(timer);
timer = setTimeout(function() {
fn.apply(context, args);
}, delay);
};
}
// Apply debounce to our shipping rate fetch method
var fetch = debounce(function() {
$scope.fetching = true;
cartService.updateShipping($scope.shipping, function(data) {
$scope.fetching = false;
$scope.quotes = data;
});
}, 1000);
// Watch the shipping fields - when enough done and user is done typing then get quote
$scope.$watch('shipping', function(newVal, oldVal) {
// I use this to play around with what fields I actually want before I do something
var fields = ['street', 'region', 'name', 'postal', 'country', 'city'];
var valid = true;
fields.forEach(function(field) {
if (!$scope.form[field].$valid) {
valid = false;
}
});
if (valid) fetch();
}, true);
My form fields are setup like this:
<input type="text" name="street ng-model="shipping.street" required>
<input type="text" name="name" ng-model="shipping.name" required>
Notice how I make them part of a "shipping" object - that allows me to watch the shipping fields independently of others such as billing.
Note that the above is for the extreme cases such as shipping fields. For simple things such as subscribing to a newsletter if they check a box then you don't need to use the above and can simply do an ng-click="spamMe();" call in your checkbox. That function (spamMe) would be in your controller and can then call your server etc...
var spamMe = function() {
// Grab the email field that might be at top - ideally check if it's filled in but you get the idea
var email = $scope.email;
$http.post('/api/spam', ....);
}
I'd apply a $scope.$watch on each of those variables to trigger a function that checks to see if all the fields for a given section are filled out, and if so, then submit it to the server as an ajax request.
Here's my attempt at writing this:
var shippingFields = ['address', 'city', 'state', 'zip'] // etc
function submitFieldsWhenComplete(section, fields) {
fieldValues = fields.forEach(function (field) {
return $scope[section][field]
});
if (fieldValues.every()) {
// We've got all the values, submit to the server
$http.post({
url: "/your/ajax/endpoint",
data: $scope.shipping
})
}
}
shippingFields.forEach(function(field) {
$scope.$watch(function() {
return $scope['shipping'][field]
}, function(val) {
submitFieldsWhenComplete('shipping', shippingFields);
});
});
Related
I need to enable/disable all validation rules in Angular form or subform under ng-form="myForm" based on a scope variable $scope.isValidationRequired. So, if isValidationRequired is false, none of the validations set for the designated group of fields will run, and the result will always be myForm.$valid==true, otherwise, the validation rules will run as usual.
I did a lot of research, and realized that this feature is not available out of the box with Angular. However, I found some add-ons or with some customization, it is possible.
For example, I can use the add-on angular-conditional-validation (github and demo) with custom directive enable-validation="isValidationRequired". This will be perfect, except that I cannot apply this feature for a group of fields under ng-form. I have to add this directive for each and every field where applicable.
The other solution is to use custom validation using Angular $validators pipeline. This requires some extra effort and I don't have time since the sprint is almost over and I have to give some results in a few days.
If you have any other suggestions please post an answer.
Use Case:
To clarify the need for this, I will mention the use-case. The end user can fill the form with invalid data and he can click Save button and in this case, the validation rules shouldn't be triggered. Only when the user clicks Validate and Save then the validation rules should be fired.
Solution:
See the final plunker code here.
UPDATE: as per comments below, the solution will cause the browser to hang if inner subforms are used under ng-form. More effort is needed to debug and resolver this issuer. If only one level is used, then it works fine.
UPDATE: The plunker here was updated with a more general solution. Now the code will work with a form that has sub-forms under ng-form. The function setAllInputsDirty() checks if the object is a $$parentForm to stop recursion. Also, the changeValidity() will check if the object is a form using $addControl then it will call itself to validate its child objects. So far, this function works fine, but it needs a bit of additional optimization.
One idea is to reset the errors in the digest loop if the validation flag is disabled. You can iterate through the form errors on change and set them to valid, one by one.
$scope.$watch(function() {
$scope.changeValidity();
}, true);
$scope.changeValidity = function() {
if ($scope.isValidationRequired === "false") {
for (var error in $scope.form.$error) {
while ($scope.form.$error[error]) {
$scope.form.$error[error][0].$setValidity(error, true);
}
}
}
}
Here is a plunkr: https://plnkr.co/edit/fH4vGVPa1MwljPFknYHZ
This is the updated answer that will prevent infinite loop and infinite recursion. Also, the code depends on a known root form which can be tweaked a bit to make it more general.
References: Pixelastic blog and Larry's answer
Plunker: https://plnkr.co/edit/ycPmYDSg6da10KdoNCiM?p=preview
UPDATE: code improvements to make it work for multiple errors for each field in each subform, and loop to ensure the errors are cleared on the subform level
var app = angular.module('plunker', []);
app.controller('MainCtrl', ["$scope", function($scope) {
$scope.isValidationRequired = true;
var rootForm = "form";
function setAllInputsDirty(scope) {
angular.forEach(scope, function(value, key) {
// We skip non-form and non-inputs
if (!value || value.$dirty === undefined) {
return;
}
// Recursively applying same method on all forms included in the form except the parent form
if (value.$addControl && key !== "$$parentForm") {
return setAllInputsDirty(value);
}
if (value.$validate){
value.$validate();
}
// Setting inputs to $dirty, but re-applying its content in itself
if (value.$setViewValue) {
//debugger;
return value.$setViewValue(value.$viewValue);
}
});
}
$scope.$watch(function() {
$scope.changeValidity();
}, true);
$scope.changeValidity = function(theForm) {
debugger;
//This will check if validation is truned off, it will
// clear all validation errors
if (!theForm) {
theForm = $scope[rootForm];
}
if ($scope.isValidationRequired === "false") {
for (var error in theForm.$error) {
errTypeArr = theForm.$error[error];
angular.forEach (errTypeArr, function(value, idx) {
var theObjName = value.$name;
var theObj = value;
if (theObj.$addControl) {
//This is a subform, so call the function recursively for each of the children
var isValid=false;
while (!isValid) {
$scope.changeValidity(theObj);
isValid = theObj.$valid;
}
} else {
while (theObj.$error[error]) {
theObj.$setValidity(error, true);
}
}
})
}
} else {
setAllInputsDirty($scope);
}
}
}]);
Simple question here.
I have this watch:
// Watch our model
$scope.$watch(function () {
// Watch our team name
return self.model.team.data.name;
}, function (name) {
console.log(name);
// if we have a name
if (name) {
// Store our model in the session
sessionStorage.designer = angular.toJson(self.model);
}
});
The team model is pull in from the database as a promise (hence the data) so when the watch first fires self.model.team has not been set so it is null.
How can I get my watch to either wait until it has been set or add a check into the return function of the watch?
Use a watch expression instead of a function. This will catch any errors with missing objects and return undefined.
// Watch our model
$scope.$watch('self.model.team.data.name', function (name) {
console.log(name);
// if we have a name
if (name) {
// Store our model in the session
sessionStorage.designer = angular.toJson(self.model);
}
});
There is no magic here - if one of the variables you are accessing could be null/undefined, then you cannot get its property if it's null/undefined. So, you have to guard against that:
$scope.$watch(
function(){
return (self.model.team && self.model.team.data.name) || undefined;
},
function(v){
// ...
});
The only "magic" is when you "$watch" for expressions, but the expressions need to be exposed on the scope. So, you could do:
$scope.model = self.model;
$scope.$watch("model.team.data.name", function(v){
// ...
});
But, really, you have to ask yourself why you need a $watch here to begin with. It seems to me that you are getting the team asynchronously once - it does not look like it will change except by maybe another async call. So, just handle that when you receive the data without the $watch:
someSvc.getTeam() // I made an assumption about a service that pulls the data from db
.then(function(team){
var name = team.data.name;
// if we have a name
if (name) {
// Store our model in the session
sessionStorage.designer = angular.toJson(self.model);
}
});
An unnecessary $watch is expensive - it is evaluated on every digest cycle, so, it's best to reduce the number of $watchers.
Im working on an extjs application. We're have a page that is for looking at a particular instance of an object and viewing and editing it's fields.
We're using refs to get hold of bits of view in the controller.
This was working fine, but I've been sharding the controller into smaller pieces to make it more managable and realised that we are relying on a race condition in our code.
The logic is as follows:
Initialise the controller
parse the url to extract the id of the object
put in a call to load the model with the given view.
in the load callback call the controller load method...
The controller load method creates some stores which fire off other requests for bits of information using this id. It then uses some of the refs to get hold of the view and then reconfigures them to use the stores when they load.
If you try and call the controller load method immediately (not in the callback) then it will fail - the ref methods return undefined.
Presumably this is because the view doesnt exist... However we aren't checking for that - we're just relying on the view being loaded by the time the server responds which seems like a recipe for disaster.
So how can we avoid this and be sure that a view is loaded before trying to use it.
I haven't tried rewriting the logic here yet but it looks like the afterrender event probably does what I want.
It seems like waiting for both the return of the store load and afterrender events should produce the correct result.
A nice little abstraction here might be something like this:
yourNamespace.createWaitRunner = function (completionCallback) {
var callback = completionCallback;
var completionRecord = [];
var elements = 0;
function maybeFinish() {
var done = completionRecord.every(function (element) {
return element === true
});
if (done)
completionCallback();
}
return {
getNotifier: function (func) {
func = func || function (){};
var index = elements++;
completionRecord[index] = false;
return function () {
func(arguments);
completionRecord[index] = true;
maybeFinish();
}
}
}
};
You'd use it like this:
//during init
//pass in the function to call when others are done
this.waiter = yourNamespace.createWaitRunner(controller.load);
//in controller
this.control({
'SomeView': {
afterrender: this.waiter.getNotifier
}
});
//when loading record(s)
Ext.ModelManager.getModel('SomeModel').load(id, {
success: this.waiter.getNotifier(function (record, request) {
//do some extra stuff if needs be
me.setRecord(record);
})
});
I haven't actually tried this out yet so it might not be 100% but I think the idea is sound
I have a project where I'm using BreezeJS to fetch data from my webserver. I'm using AngularJS with the ui-select2 module. Currently, I have it where when I load my page, breezejs makes a call to fetch the data that I dump into a scope variable. From there, select2 can easily make the reference to it and build accordingly.
If I want to ajaxify things, it gets really tricky. I want to have the ability to use select2's ajax or query support, but instead of using it to fetch data, I want to use breezejs to do it. So during a page load, nothing is loaded up until I start typing in X minimum characters before it makes an ajax fetch.
Constraints:
I do not want fetch data using select2's "ajax". I want BreezeJS to handle the service calls. When I use ajax, it makes an ajax call everytime I press a character in order to filter the results (and resemble autocomplete). I just want the list to load up once and use the native filtering after that.
Here is what I have so far:
breezejs - StateContext.JS
m.app.factory('StateContext', ['$http', function ($http) {
configureBreeze();
var dataService = new breeze.DataService({
serviceName: "/Map/api",
hasServerMetadata: false
});
var manager = new breeze.EntityManager({ dataService: dataService});
var datacontext = {
getAllStates: getAllStates
};
return datacontext;
function getAllStates() {
var query = breeze.EntityQuery
.from("States");
return manager.executeQuery(query);
}
function configureBreeze() {
breeze.config.initializeAdapterInstances({ dataService: "webApi" });
}
}]);
This works and returns my json object correctly.
Here is how I call the service:
m.app.controller('LocationCtrl', ['$scope', 'StateContext', function ($scope, StateContext) {
$scope.getAllStates = function () {
StateContext.getAllStates().then(stateQuerySucceeded).fail(queryFailed);
}
$scope.getAllStates();
$scope.states = [];
function stateQuerySucceeded(data) {
data.results.forEach(function (item) {
$scope.states.push(item);
});
$scope.$apply();
console.log("Fetched States");
}
function queryFailed(error) {
console.log("Query failed");
}
$scope.select2StateOptions = {
placeholder: "Choose a State",
allowClear: true,
minimumInputLength: 2
};
}
and here is my html:
<div ng-app="m" id="ng-app">
...
...
<select ui-select2="select2StateOptions" ng-model="LocationModel.State">
<option value=""></option>
<option ng-repeat="state in states" value="{{state.id}}">{{state.name}}</option>
</select>
</div>
Currently the html select2 control loads up when the page loads. But I want to have it so when I type in more than 2 characters, I'll be able to make the call to $scope.getAllStates(); as an ajax call. BreezeJS already uses ajax natively when configuring the BreezeAdapter for webapi.
I was thinking about using select2's ajax, or query calls.. but I'd rather use breeze to fetch the data, since it makes querying extendable, and I don't want to violate my design pattern, or make the code harder to maintain, and I don't want the ajax calls to be made everytime I enter a new character into the textbox, I just want it to occur once.
Close attempt:
changed my html to:
<!-- Select2 needs to make this type="hidden" to use query or ajax, then it applies the UI skin afterwards -->
<input type="hidden" ui-select2="select2StateOptions" ng-model="LocationModel.State" /><br />
in my controller, changing select2StateOptions:
$scope.select2StateOptions = {
placeholder: "Choose a State",
allowClear: true,
minimumInputLength: 2,
query: function (query) {
debugger;
var data = StateContext.getAllStates().then(stateQuerySucceeded).fail(queryFailed);
}
};
Here's the problem. BreezeJS uses a Q library, which makes use of a thing called a "promise"; which is a promise that data will be returned after making the ajax call. The problem with this, the query function is expecting data to be populated, but the promise to call the "stateQuerySucceeded" function is made after returning from the query function.
So it hits the query function first. Then hits getAllStates(). Returns from the query (nothing is populated), then "stateQuerySucceeded" is called after that.
In otherwords, even though I have been able to fetch data, this is done too late.. select2's query function did not receive the data at the right time, and my html select is hanging on "Searching ... " with a search spinner.gif.
I don't really know this angular-ui-select2 control. I think the relevant part of the documentation is this example:
$("#e5").select2({
minimumInputLength: 2,
query: function (query) {
var data = {results: []}, i, j, s;
// simulate getting data from the server
for (i = 1; i < 5; i++) {
s = "";
for (j = 0; j < i; j++) {s = s + query.term;}
data.results.push({id: query.term + i, text: s});
}
query.callback(data);
}
});
I will leave aside the fact that you don't seem to be interested in using the two-or-more characters that the user enters in your query (maybe you just left that out). I'll proceed with what seems to me to be nonsense, namely, to fetch all states after the user types any two letters.
What I think you're missing is the role of the query.callback which is to tell "angular-ui-select2" when the data have arrived. I'm guessing you want to call query.callback in your success function.
$scope.select2StateOptions = {
placeholder: "Choose a State",
allowClear: true,
minimumInputLength: 2,
query: function (query) {
StateContext.getAllStates()
.then(querySucceeded).catch(queryFailed);
function querySucceeded(response) {
// give the {results:data-array} to the query callback
query.callback(response);
}
function queryFailed(error) {
// I don't know what you're supposed to do.
// maybe return nothing in the query callback?
// Tell the user SOMETHING and then
query.callback({results:[]});
}
}
};
As I said, I'm just guessing based on a quick reading of the documentation. Consider this answer a "hint" and please don't expect me to follow through and make this actually work.
I need to validate a form with a bunch of inputs in it. And, if an input is invalid, indicate visually in the form that a particular attribute is invalid. For this I need to validate each form element individually.
I have one model & one view representing the entire form. Now when I update an attribute:
this.model.set('name', this.$name.val())
the validate method on the model will be called.
But, in that method I am validating all the attributes, so when setting the attribute above, all others are also validated, and if any one is invalid, an error is returned. This means that even if my 'name' attribute is valid, I get errors for others.
So, how do I validate just one attribute?
I think that it is not possible to just validate one attribute via the validate() method. One solution is to not use the validate method, and instead validate every attribute on 'change' event. But then this would make a lot of change handlers. Is it the correct approach? What else can I do?
I also think that this points to a bigger issue in backbone:
Whenever you use model.set() to set an attribute on the model, your validation method is run and all attributes are validated. This seems counterintuitive as you just want that single attribute to be validated.
Validate is used to keep your model in a valid state, it won't let you set an invalid value unless you pass a silent:true option.
You could either set all your attributes in one go:
var M=Backbone.Model.extend({
defaults:{
name:"",
count:0
},
validate: function(attrs) {
var invalid=[];
if (attrs.name==="") invalid.push("name");
if (attrs.count===0) invalid.push("count");
if (invalid.length>0) return invalid;
}
});
var obj=new M();
obj.on("error",function(model,err) {
console.log(err);
});
obj.set({
name:"name",
count:1
});
or validate them one by one before setting them
var M=Backbone.Model.extend({
defaults:{
name:"",
count:0
},
validate: function(attrs) {
var invalid=[];
if ( (_.has(attrs,"name"))&&(attrs.name==="") )
invalid.push("name");
if ( (_.has(attrs,"count"))&&(attrs.count===0) )
invalid.push("count");
if (invalid.length>0) return invalid;
}
});
var obj=new M();
obj.on("error",function(model,err) {
console.log(err);
});
if (!obj.validate({name:"name"}))
obj.set({name:"name"},{silent:true});
I recently created a small Backbone.js plugin, Backbone.validateAll, that will allow you to validate only the Model attributes that are currently being saved/set by passing a validateAll option.
https://github.com/gfranko/Backbone.validateAll
That is not the issue of Backbone, it doesn't force you to write validation in some way. There is no point in validation of all attributes persisted in the model, cause normally your model doesn't contain invalid attributes, cause set() doesn't change the model if validation fails, unless you pass silent option, but that is another story. However if you choose this way, validation just always pass for not changed attributes because of the point mentioned above.
You may freely choose another way: validate only attributes that are to be set (passed as an argument to validate()).
You can also overload your model's set function with your own custom function to pass silent: true to avoid triggering validation.
set: function (key, value, options) {
options || (options = {});
options = _.extend(options, { silent: true });
return Backbone.Model.prototype.set.call(this, key, value, options);
}
This basically passes {silent:true} in options and calls the Backbone.Model set function with {silent: true}.
In this way, you won't have to pass {silent: true} as options everywhere, where you call
this.model.set('propertyName',val, {silent:true})
For validations you can also use the Backbone.Validation plugin
https://github.com/thedersen/backbone.validation
I had to make a modification to the backbone.validation.js file, but it made this task much easier for me. I added the snippet below to the validate function.
validate: function(attrs, setOptions){
var model = this,
opt = _.extend({}, options, setOptions);
if(!attrs){
return model.validate.call(model, _.extend(getValidatedAttrs(model), model.toJSON()));
}
///////////BEGIN NEW CODE SNIPPET/////////////
if (typeof attrs === 'string') {
var attrHolder = attrs;
attrs = [];
attrs[attrHolder] = model.get(attrHolder);
}
///////////END NEW CODE SNIPPET///////////////
var result = validateObject(view, model, model.validation, attrs, opt);
model._isValid = result.isValid;
_.defer(function() {
model.trigger('validated', model._isValid, model, result.invalidAttrs);
model.trigger('validated:' + (model._isValid ? 'valid' : 'invalid'), model, result.invalidAttrs);
});
if (!opt.forceUpdate && result.errorMessages.length > 0) {
return result.errorMessages;
}
}
I could then call validation on a single attribute like so
this.model.set(attributeName, attributeValue, { silent: true });
this.model.validate(attributeName);