I'm using DialogFlow to create a Google Assistant Application. For fullfilment I'm using a custom app with NodeJS client library.
I noted that when DialogFlow's request get my application I can see all request, also the parameters object
const astronomyAssistant = functions.https.onRequest((request, response) => {
const app = new DialogflowApp({ request, response });
console.log(`Request headers: ${JSON.stringify(request.headers)}`);
console.log(`Request body: ${JSON.stringify(request.body)}`);
app.handleRequest(actionMap);
});
There some way to access the request object inside of a handle action? How can I access request object using app?
Workaround:
You can put all handle function that depends of request object inside of functions.https.onRequest callback.
For instance:
const astronomyAssistant = functions.https.onRequest((request, response) => {
const app = new DialogflowApp({ request, response });
console.log(`Request headers: ${JSON.stringify(request.headers)}`);
console.log(`Request body: ${JSON.stringify(request.body)}`);
const foo = app => {
console.log(request);
}
actionMap.set('input.foo', foo);
app.handleRequest(actionMap);
});
But, for sure this is not a good practice.
You can use app.getArgument("my-parameter") to quickly access any parameter you've defined within your actions.
Related
I'm currently developing a flutter app that sends some http requests to an external service, but actually, I keep some API keys in the app, and I want to secure them using cloud functions. How can I create a function that executes something like this? This is a Stripe request that I actually make from the app.
Future<Customer?> createCustomer(String userId) async {
final String url = 'https://api.stripe.com/v1/customers';
Map<String,String> headers = {
'Authorization': 'Bearer <API_KEY_HERE>',
'Content-Type': 'application/x-www-form-urlencoded'
};
var response = await _client.post(
Uri.parse(url),
headers: headers,
body: {'name': "test", "metadata[userId]": userId},
);
final decodedResult = json.decode(response.body);
log.info(decodedResult);
if (response.statusCode == 200) {
try {
final customer = Customer.fromMap(decodedResult);
currentCustomer = customer;
return customer;
} catch (e) {
log.error(e.toString());
return null;
}
} else {
return null;
}
}
You cloud very well, from a Cloud Function, call the Stripe REST API with, for example, the axios library.
But since we use the Admin SDK for Node.js to write Cloud Functions for Firebase, I would suggest to use the Stripe Node.js API (This link points to the customers.create method).
You can do that from any type of Cloud Function, e.g. a Callable one that you call from your app or a background triggered one, e.g. when a new customer doc is created in a Firestore collection.
The following search on SO will return many code examples: https://stackoverflow.com/search?q=Firebase+Cloud+Functions+stripe
I have feathers API with REST/axios. Configured like this:
const axios = require("axios");
const restClient = rest(serverPaths.serverPath.slice(0, -1));
restApi.configure(restClient.axios(axios.create({
headers: {accessToken: '--change---me'} // this works all the time
})));
The request:
const sr = await this.$restApi.service('users').find({}, {
headers: {
accesstoken: '-this is a new token-' // this does not work
}
});
However simply changing to get and it works:
const sr = await this.$restApi.service('users').get(1, {
headers: {
accesstoken: '-this is a new token-' // now it works
}
});
So for get I receive: accesstoken: --change---me, -this is a new token- so both of them, which is fine, I'll just delete the first one from config file.
But that's mean I do something wrong. Something simple and stupid but I just can't figure it out what.
Note: all regular API request with axios, not using feathers services, works without problem, I can send any header I want and it just work.
I want to trigger a Google Composer airflow dag using Appscript. Is there any way to do it via rest API or another way.
If it is possible please suggest the solution.
Airflow has an endpoint that allows to trigger a DAG through its REST API, however it’s not possible to access it directly, since within the Cloud Composer architecture, the Airflow web server is located under an App Engine flexible environment. By default, the Airflow web server is integrated with Identity-Aware Proxy (IAP) and authentication is required.
Based on that, I found an example in the Cloud Composer documentation, that guides you to trigger a DAG using Cloud Functions, although the code is in JavaScript I don’t think it’s possible to execute it by Google App Script.
On the other hand, a workaround is to follow the Triggering DAGs guide changing some settings as follows.
In the creation of the function instead of setting the trigger type as Cloud Storage set it as HTTP, and check the “Allow unauthenticated invocations” for test purpose. An URL will be displayed, the goal is that every time that URL is accessed the DAG is executed.
Modify the first part of the index.js file, since no data would be passed as parameters and also the makeIapPostRequest function to return the response of the API call.
exports.triggerDag = async (req, res) => { // Modification
// Fill in your Composer environment information here.
// The project that holds your function
const PROJECT_ID = 'your-project-id';
// Navigate to your webserver's login page and get this from the URL
const CLIENT_ID = 'your-iap-client-id';
// This should be part of your webserver's URL:
// {tenant-project-id}.appspot.com
const WEBSERVER_ID = 'your-tenant-project-id';
// The name of the DAG you wish to trigger
const DAG_NAME = 'composer_sample_trigger_response_dag';
// Other constants
const WEBSERVER_URL = `https://${WEBSERVER_ID}.appspot.com/api/experimental/dags/${DAG_NAME}/dag_runs`;
const USER_AGENT = 'gcf-event-trigger';
const BODY = {conf: ‘’}; // Modification
// Make the request
try {
const iap = await authorizeIap(CLIENT_ID, PROJECT_ID, USER_AGENT);
const apiReponse = await makeIapPostRequest(WEBSERVER_URL, BODY, iap.idToken, USER_AGENT); // Modification
res.status(200).send('DAG_running!'); // Modification
} catch (err) {
console.error('Error authorizing IAP:', err.message);
throw new Error(err);
}
};
const makeIapPostRequest = async (url, body, idToken, userAgent) => {
const res = await fetch(url, {
method: 'POST',
headers: {
'User-Agent': userAgent,
Authorization: `Bearer ${idToken}`,
},
body: JSON.stringify(body),
});
if (!res.ok) {
const err = await res.text();
console.error('Error making IAP post request:', err.message);
throw new Error(err);
}
return {
apiRes: res.ok, // Modification
};
};
At this point, anything else has to be changed, so in your Script file execute the next instructions in order to trigger the DAG.
function myFunction() {
var response = UrlFetchApp.fetch("Cloud-function-URL");
Logger.log(response.getAllHeaders());
}
Finally, verify in the Airflow web interface if the DAG was triggered.
I'm using the account linking feature for Actions SDK and following the guide here (https://developers.google.com/assistant/identity/google-sign-in#start_the_authentication_flow)
It shows the initialization like this
const app = actionssdk({
// REPLACE THE PLACEHOLDER WITH THE CLIENT_ID OF YOUR ACTIONS PROJECT
clientId: CLIENT_ID,
});
But for my use case, I'll read the clientId from DB which is stored against the projectId of the project. I can extract the projectId only after the MAIN intent is triggered.
My question is, how can I set the clientId after initializing actionssdk?
This solution uses the new Actions SDK, but the principal is the same for the legacy SDK as well:
const {
conversation,
Canvas,
} = require('#assistant/conversation');
const functions = require('firebase-functions');
const wrapper = async (req, res) => {
// You can get any data you need here:
const myAsyncBootstrapData = await getData();
const app = conversation({debug: true, ...myAsyncBootstrapData});
app.handle('welcome', (conv) => {
conv.add('This is a demo.');
});
return app(req, res);
};
exports.ActionsOnGoogleFulfillment = functions.https.onRequest(wrapper);
functions.https.onRequest accepts any callable, including ones that return promises. If you need to block while loading configuration data asynchronously, you can do so by wrapping your definition in an async function.
I found a simple solution to this. I am adding it here for future references.
// handler.js
async function handleRequest(req, res) {
const clientId = // retrieve the clienId using your business logic
const app = actionssdk({
clientId: clientId
})
}
module.exports = handleRequest;
Instead of directly creating an instance of actionssdk, wrap it inside a function like this.
// index.js
const handler = require('./path/to/hander.js');
app.post('/webhook', handler);
Then when defining the webhook, use the wrapper function to handle the webhook requests
Trying to implement web-hook (with V2 dialogflow) running nodejs. Received response "MalformedResponse 'final_response' must be set.". Below is the code. To the end of POST (app.post) code block was expecting conv.close would send SimpleResponse. But that's not happening. Need help understand why this error is seen and probable direction to solve it.
Thanks
const express = require('express');
const {
dialogflow,
Image,
SimpleResponse,
} = require('actions-on-google')
const bodyParser = require('body-parser');
const request = require('request');
const https = require("https");
const app = express();
const Map = require('es6-map');
// Pretty JSON output for logs
const prettyjson = require('prettyjson');
const toSentence = require('underscore.string/toSentence');
app.use(bodyParser.json({type: 'application/json'}));
// http://expressjs.com/en/starter/static-files.html
app.use(express.static('public'));
// http://expressjs.com/en/starter/basic-routing.html
app.get("/", function (request, response) {
console.log("Received GET request..!!");
//response.sendFile(__dirname + '/views/index.html');
response.end("Response from my server..!!");
});
// Handle webhook requests
app.post('/', function(req, res, next) {
console.log("Received POST request..!!");
// Log the request headers and body, to aide in debugging. You'll be able to view the
// webhook requests coming from API.AI by clicking the Logs button the sidebar.
console.log('======Req HEADERS================================================');
logObject('Request headers: ', req.headers);
console.log('======Req BODY================================================');
logObject('Request body: ', req.body);
console.log('======Req END================================================');
// Instantiate a new API.AI assistant object.
const assistant = dialogflow({request: req, response: res});
// Declare constants for your action and parameter names
//const PRICE_ACTION = 'price'; // The action name from the API.AI intent
const PRICE_ACTION = 'revenue'; // The action name from the API.AI intent
var price = 0.0
// Create functions to handle intents here
function getPrice(assistant) {
console.log('** Handling action: ' + PRICE_ACTION);
let requestURL = 'https://blockchain.info/q/24hrprice';
request(requestURL, function(error, response) {
if(error) {
console.log("got an error: " + error);
next(error);
} else {
price = response.body;
logObject('the current bitcoin price: ' , price);
// Respond to the user with the current temperature.
//assistant.tell("The demo price is " + price);
}
});
}
getPrice(assistant);
var reponseText = 'The demo price is ' + price;
// Leave conversation with SimpleResponse
assistant.intent(PRICE_ACTION, conv => {
conv.close(new SimpleResponse({
speech: responseText,
displayText: responseText,
}));
});
}); //End of app.post
// Handle errors.
app.use(function (err, req, res, next) {
console.error(err.stack);
res.status(500).send('Oppss... could not check the price');
})
// Pretty print objects for logging.
function logObject(message, object, options) {
console.log(message);
console.log(prettyjson.render(object, options));
}
// Listen for requests.
let server = app.listen(process.env.PORT || 3000, function () {
console.log('Your app is listening on ' + JSON.stringify(server.address()));
});
In general, The "final_response" must be set error is because you didn't send anything back. You have a lot going on in your code, and while you're on the right track, there are a few things in the code that could be causing this error.
First - in the code, it looks like you are confused about how to send a response. You have both a call to conv.close() and the commented out assistant.tell(). The conv.close() or conv.ask() methods are the way to send a reply using this version of the library. The tell() method was used by a previous version and is no longer supported.
Next, your code looks like it is only setting up the assistant object when the routing function is called. While this can be done, it is not the usual way to do it. Typically you'll create the assistant object and setup the Intent handlers (using assistant.intent()) as part of the program initialization. This is a rough equivalent to setting up the express app and the routes for it before the request itself comes in.
The portion that sets up the Assistant and then hooks it into a route might look something like this:
const assistant = dialogflow();
app.post('/', assistant);
If you really wanted to examine the request and response objects first, you might do this as something like
const assistant = dialogflow();
app.post('/', function( req, res ){
console.log(JSON.stringify(req.body,null,1));
assistant( req, res );
});
Related to this appears to be that you're trying to execute code in the route handler and then trying to call the intent handler. Again, this might be possible, but isn't the suggested way to use the library. (And I haven't tried to debug your code to see if there are problems in how you're doing it to see if you're doing it validly.) More typical would be to call getPrice() from inside the Intent handler instead of trying to call it from inside the route handler.
But this leads to another problem. The getPrice() function calls request(), which is an asynchronous call. Async calls are one of the biggest problems that causes an empty response. If you are using an async call, you must return a Promise. The easiest way to use a Promise with request() is to use the request-promise-native package instead.
So that block of code might look something (very roughly) like this:
const rp = require('request-promise-native');
function getPrice(){
return rp.get(url)
.then( body => {
// In this case, the body is the value we want, so we'll just return it.
// But normally we have to get some part of the body returned
return body;
});
}
assistant.intent(PRICE_ACTION, conv => {
return getPrice()
.then( price => {
let msg = `The price is ${price}`;
conv.close( new SimpleResponse({
speech: msg,
displayText: msg
});
});
});
The important thing to note about both getPrice() and the intent handler are that they both return a Promise.
Finally, there are some odd aspects in your code. Lines such as res.status(500).send('Oppss... could not check the price'); probably won't do what you think they will do. It won't, for example, send a message to be spoken. Instead, the Assistant will just close the connection and say that something went wrong.
Many thanks to #Prisoner. Below is the V2 working solution based on above comments. Same has been verified on nodejs webhook (without firebase). V1 version of the code was referenced from https://glitch.com/~aog-template-1
Happy coding..!!
// init project pkgs
const express = require('express');
const rp = require('request-promise-native');
const {
dialogflow,
Image,
SimpleResponse,
} = require('actions-on-google')
const bodyParser = require('body-parser');
const request = require('request');
const app = express().use(bodyParser.json());
// Instantiate a new API.AI assistant object.
const assistant = dialogflow();
// Handle webhook requests
app.post('/', function(req, res, next) {
console.log("Received POST request..!!");
console.log('======Req HEADERS============================================');
console.log('Request headers: ', req.headers);
console.log('======Req BODY===============================================');
console.log('Request body: ', req.body);
console.log('======Req END================================================');
assistant(req, res);
});
// Declare constants for your action and parameter names
const PRICE_ACTION = 'revenue'; // The action name from the API.AI intent
var price = 0.0
// Invoke http request to obtain blockchain price
function getPrice(){
console.log('getPrice is invoked');
var url = 'https://blockchain.info/q/24hrprice';
return rp.get(url)
.then( body => {
// In this case, the body is the value we want, so we'll just return it.
// But normally we have to get some part of the body returned
console.log('The demo price is ' + body);
return body;
});
}
// Handle AoG assistant intent
assistant.intent(PRICE_ACTION, conv => {
console.log('intent is triggered');
return getPrice()
.then(price => {
let msg = 'The demo price is ' + price;
conv.close( new SimpleResponse({
speech: msg,
}));
});
});
// Listen for requests.
let server = app.listen(process.env.PORT || 3000, function () {
console.log('Your app is listening on ' + JSON.stringify(server.address()));
});