I'm using SMS studio do have a quick chat bot conversation with inbound SMS messages, and would like to forward the conversation log to email after it's complete. I've written a function that uses the SendGrid API to forward SMSes to email. It works independently - ie, if I configure the phone number to run the function immediately as a text comes in, it will email that single SMS input.
However, I'd like to add the function to the end of of Twilio Studio flow, so that it emails the entire log of the conversation to me, once it's over. Once I append the function to the end of the studio flow, it stops working, and I get a failure notice.
Here's the code in the function:
const got = require('got');
exports.handler = function(context, event, callback)
{
const requestBody = {
personalizations: [{ to: [{ email: context.TO_EMAIL_ADDRESS }] }],
from: { email: context.FROM_EMAIL_ADDRESS },
subject: `New SMS message from: ${event.From}`,
content: [
{
type: 'text/plain',
value: event.Body
}
]
}
got.post('https://api.sendgrid.com/v3/mail/send', {
headers: {
Authorization: `Bearer ${context.SENDGRID_API_KEY}`,
'Content-Type': 'application/json'
},
body: JSON.stringify(requestBody)
})
.then(response => {
let twiml = new Twilio.twiml.MessagingResponse();
callback(null, twiml);
})
.catch(err => {
callback(err);
});
};`
Here's the error the debugger returns, if I make this function the last step in a Twilio studio flow:
Error - 81017
Error on Twilio Function response
There was an error in the response back from a Twilio Function attached to the Studio flow.
Possible Causes
Your Function timed out before responding
Your Function returned an error response
Possible Solutions
Your Function must contain a callback.
Make sure you place the Function callback callback(err, response) is placed correctly in your Function code.
If you are using a JavaScript promise, make sure the callback is called in both success and catch blocks.
Your Function responded with an error.
I'm having a hard time figuring out what the error is. Does anyone have any tips?
Thanks!
Jarod from Twilio. I actually wrote an app very similar to this. Code looks good to me. Often when people encounter an error of this nature it is coming from the SendGrid promise. Here are two tips:
You can actually log errors and view responses from the function if you leave the function open while testing it. The logs will be at the bottom. These usually have more information.
Check that you are using an outbound email address that has been whitelisted on your SendGrid account Whitelabel - Sendgrid
Make sure you have added your environment variables to your Functions config. Twilio Functions Configuration - Console
Hope that helps! If not feel free to email support#twilio.com with more questions.
Related
I have found a few examples of context.getTwilioClient(), but I have not been able to locate any online documentation. It's probably right under my nose, but it is eluding me. In particular, I'm trying to get information about workers in a Task Router queue (i.e., how many workers are in the different statuses and how many workers are available), but having the documentation will help with future projects.
I found some documentation saying that the context.getTwilioClient(), "enables you to invoke the getTwilioClient method on the context object for a fully-initialized Twilio REST API client." (https://support.twilio.com/hc/en-us/articles/115007737928-Getting-Started-with-Twilio-Functions)
It then shows this example, but there is no implementation of "messages" when I attempt to run this code:
var client = context.getTwilioClient();
client.messages.create({
to: '+12025551212',
from: '+12065551212',
body: "hello world!"})
Thanks.
The messages property should be on the client. getTwilioClient returns the Twilio helper library for Node.js.
I just created a Function with your code, and it worked as expected, meaning that I got the SMS, however, the function did time out because the callback was never invoked. To end the function invocation and respond to the caller, make sure you always invoke the callback function, like this:
exports.handler = function(context, event, callback) {
var client = context.getTwilioClient();
client.messages.create({
to: '+1xxxxxxxxxx',
from: '+1xxxxxxxxxxx',
body: "hello world!"})
.then((message) => {
console.log('SMS successfully sent');
console.log(message.sid);
// Make sure to only call `callback` once everything is finished, and to pass
// null as the first parameter to signal successful execution.
return callback(null, `Success! Message SID: ${message.sid}`);
})
.catch((error) => {
console.error(error);
return callback(error);
});
};
You can learn more about the callback function here.
If you still encounter this issue, can you tell use what Node Version you're using and which module dependencies and their versions?
You can find these details at Settings & More > Dependencies in your Twilio Functions Service.
We have a studio flow called "Google LA" that's triggered via Rest API. This flow has a Send and Wait for Reply so we hook this flow to "When a message comes in" so it will follow the rest of the flow when customer rates the service 1 to 5 stars. Now, within the Send and Wait for Reply, we want the customer's reply be forwarded to our main business phone number for tracking/recording purposes and so we can address their issues for rating us 1 to 3 stars. Here's our setup:
This is what we want:
Edited for philnash suggestion:
I created a function in Twilio with this code:
exports.handler = function(context, event, callback) {
const accountSid = context.ACCOUNT_SID;
const authToken = context.AUTH_TOKEN;
const client = require('twilio')(accountSid, authToken);
client.messages
.create({
body: widgets.negative1_3.inbound.Body,
from: '+12132779513',
to: '+12133885256'
})
.then(message => console.log(message.sid));
};
However, it did not send anything or the customer response. I renamed the negative1-3 widget to negative1_3 and published the studio flow.
I tried changing the body: 'Hello' to make sure that my function works, and yes. I received the 'Hello' sms to my verified caller ID phone number after it reaches the first_question -> check_response -> negative1_3.
Twilio developer evangelist here.
You don't necessarily need to forward the message here. You can make an API call to your own service with all the data you need from the message, so you can store and react to the information that way.
To do so you will want to add either an HTTP Request widget or a Run Function widget after the Send and Wait For Reply widget. Within those widgets, you can access the reply from the Send And Wait For Reply widget using liquid tags. You can see how to call on the variables in the docs for the Send and Wait For Reply widget. In the case of your widget, you should be able to get the body of the reply by referring to:
widgets.negative1-3.inbound.Body
(Although I am not sure how the name "negative1-3" will work, so you might try widgets["negative1-3"] instead, or rename the widget with underscores.)
Using the body of the inbound message, as well as the from number, you can send the data to your own application with the HTTP request widget or with a Run Function widget.
Edit
Your function can only access parameters that you set in the function widget config. You can then access those parameters in the event object. You also need to return once the message is sent successfully using the callback function. One other tip, you don't need to instantiate your own client, you can get it from the context. Like so:
exports.handler = function(context, event, callback) {
const client = context.getTwilioClient();
client.messages
.create({
body: event.Body,
from: '+12132779513',
to: '+12133885256'
})
.then(message => {
console.log(message.sid);
callback(null, "OK");
})
.catch(error => callback(error));
};
My company uses Twilio Flex as our phone system and I was recently tasked with setting up a feature that will let us edit a TwiML voice message that plays before our normal voice message. This TwiML message will be changed through a Twilio bot that I've published in our Microsoft Teams.
The reason for this is so that our support desk can add a short message in the lines of "We're currently experiencing issues with X" before our normal "Welcome to [Company] support" message.
If TwiML's can be edited using HTTP POST/PUT or Twilio's API this should be a trivial matter, but so far I've not been able to figure out how.
I couldn't find any reference to this in the API doc, so I decided that HTTP POST would be the way to go. Using this as a start off point, I'm able to retrieve my TwiML using HTTP GET:
https://support.twilio.com/hc/en-us/articles/223132187--Not-Authorized-error-when-trying-to-view-TwiML-Bin-URL
const axios = require('axios');
const crypto = require('crypto');
const accountSidFlex = process.env.accountSidFlex;
const authTokenFlex = process.env.authTokenFlex;
var URL = 'https://handler.twilio.com/twiml/EHXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX' + '?AccountSid=' + accountSidFlex
var twilioSig = crypto.createHmac('sha1', authTokenFlex).update(new Buffer(URL, 'utf-8')).digest('Base64')
var config = {
headers:{
'X-TWILIO-SIGNATURE': twilioSig
}
}
axios.get(
URL,config
).catch(error => console.log(error))
.then(response => {
console.log(response.data)
})
response.data shows the TwiML's current XML content.
My attempts at a POST only gives the same output as the GET, while PUT gives 405 Method Not Allowed.
var URL = 'https://handler.twilio.com/twiml/EHXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX' + '?AccountSid=' + accountSidFlex
var twilioSig = crypto.createHmac('sha1', authTokenFlex).update(new Buffer(URL, 'utf-8')).digest('Base64')
var config = {
headers:{
'X-TWILIO-SIGNATURE': twilioSig,
'Content-Type': 'text/xml'
}
}
var xml =
'<?xml version="1.0" encoding="UTF-8"?>\
<Response><Play digits="www"/>\
<Say voice="alice">"We are currently experiencing X related issues". </Say>\
</Response>';
axios.post(
URL,xml,config
)
.catch(error => console.log(error))
.then(response => {
console.log(response.data)
})
Ideally I'd like to be able to change a specific TwiML using either HTTP methods or the Twilio-API, so that we can use it in out Studio Flow. We'd just keep it silent until we need to add something to it and revert back to silent once the issues have passed.
Any help would be appreciated!
You cannot currently change the contents of TwiML Bins, Studio Flows, or Twilio Functions programatically. I believe the key functionality you are looking for is a way to dynamically update the messaging (Say/Play Widget) in a Studio flow based on some condition.
One way is to use a Function Widget to retrieve a Twilio Sync document for the message, returning the message as JSON and have the Say/Play widget play that message. You can find the Twilio Sync REST API examples for Add, Modify, and Retrieve in the associated document.
You can retrieve the parsed response using variable syntax detailed here, https://www.twilio.com/docs/studio/widget-library#run-function.
I've been stalking around here and have gotten most of my answers as I make my way through this new tool, but I'm now stuck and need some direct advice.
The Gather function in Studio is not PCI compliant, so I have to shift my call to a Function and return the parsed data--I finally figured out how to do that one--however, I've found that I cannot call the web service housed within the single function and had to send the with event.Digits to another function to make the web service call to my token provider. This works, however it has led to a strange result: my token is read back as TTS and then the call is hung up. I have no TTS action in play. Below are my sets of code:
Initial function called from Studio:
const got = require('got');
exports.handler = function(context, event, callback) {
let twiml = new Twilio.twiml.VoiceResponse();
twiml.gather({
input: 'dtmf',
finishOnKey: '#',
timeout: 10,
action: 'paymenttest',
method: 'GET'
}).say('Enter CC');
console.log(twiml);
callback(null, twiml);
};
This successfully calls my function with the digits entered:
const got = require('got');
exports.handler = function(context, event, callback) {
let twiml = new Twilio.twiml.MessagingResponse();
const url ='my payment gateway' + event.Digits + '&EXPDATE=1220&CARDTYPE=VI';
got.get(url, {
headers: {
'content-Type': 'application/x-www-form-urlencoded'
}
}).then(function(response) {
// Check the response and ask your second question here
event.callback(null, response.body);
}).catch(function(error) {
// Boo, there was an error.
callback(error)
});
};
This successfully returns the token....but as mentioned prior...it's read back out to me instead of getting included in the data returned back to Studio.
Twilio developer evangelist here.
Right now Studio is not well setup for using TwiML from a Twilio Function and then continuing the flow. In your case when you return the token from your second Function Twilio is dealing with it as if you just returned text to a regular TwiML webhook. When this happens Twilio defaults to assuming you meant <Say> and reads out the text.
While the team work on redirecting calls back into Studio flows there is a workaround.
Instead of returning the token in the second Function, return some TwiML that includes a <Redirect> to your Studio flow's webhook URL with ?FlowEvent=audioComplete appended. You will also need to add a dummy Say/Play widget after your Function widget (it becomes the next part in the flow that can trigger an "audio complete" message, so exists to collect that and send on to the next widget).
The only thing that we haven't handled in thie workaround is sending the token to the flow. I don't believe we can do this via this redirect workaround, so instead I'd recommend storing the token in your own database or something like a Twilio Sync object. This way you can use it outside of Studio however you like. If you need it within the Studio flow then you can create one more Function that returns the token as JSON, and that will be stored within the flow variables then.
If you would prefer to use <Pay>, as this would be a lot easier, I also recommend requesting the pay connector you need.
I think philnash answer here got old, even when it still works.
Right now, you should have to call the first function using TwiML Redirect node.
In the second function, you should have to add a redirect to the webhook of the flow adding ?FlowEvent=return&foo=bar (where foo=bar should be changed by the info you really want to return).
I've set up a lambda function and created some GET and POST methods inside the API Gateway which seem to work fine when testing them inside the web application.
I am then trying to call the functions inside an iOS application which is set up using the mobile hub. The functions also work inside the testing facility via the mobile hub perfectly fine, however when I actually test the functions inside the app I get:
"message" : "Internal server error"
I know the error is not much to work from, but I can't figure out a way to get a more detailed error description.
Any ideas?
This may happen because your Lambda function is not set to return a HTTP status code.
Changing from
exports.handler = (event, context, callback) => {
callback(null, 'Hello from Lambda');
};
to
exports.handler = (event, context, callback) => {
callback(null, { statusCode: 200, body: 'Hello from Lambda' });
};
Should fix the issue.
The JSON.stringify() solved my issue. The response.body needs to be in String format and not as JSON. I hope this helps.
exports.sendRes = (body, status = 200) => {
var response = {
statusCode: status,
headers: {
"Content-Type": "application/json"
},
body: JSON.stringify(body)
};
return response;
};
I had the same issue with the following code:
exports.handler = async event => {
console.log("hello world");
return {
statusCode: 200,
body: event
};
};
To fix all I had to do was JSON.stringify() the body.
exports.handler = async event => {
console.log("hello world");
return {
statusCode: 200,
body: JSON.stringify(event), // <-- here
};
};
I had this problem until I click in "Deploy API" under the "Actions" button.
The other possible reason could be the payload/request/response limits on API Gateway (10MB) and/or Lambda (6MB)
None of the above answers worked for me. I was having a permission issue. Below is how I solved it.
Context
This is my lambda function:
exports.handler = function(event, context, callback) {
callback(null, {
statusCode: '200',
body: JSON.stringify({ 'message': 'hello world' }),
headers: {
'Content-Type': 'application/json',
},
});
};
I used terraform to provision api gateway and lambda. I used the example code provided by this blog post.
Diagnosis
In the lambda console I ran a test event on my lambda. As my lambda was super basic I used the hello world test template, named, and saved it. The test return success.
I checked cloudwatch logs, but couldn't find anything of use. I'm new to AWS so wasn't sure if I had to set anything up.
In the api gateway console I ran a test event. I just added Content-Type:application/json to the headers of the event and ran the test. For whatever weird reason the test results returned on the right side of the browser so had to scroll to the right to see them.
I got this result: Execution failed due to configuration error: Invalid permissions on Lambda function
SOLUTION
I checked the basic terraform example for api gateway and lambda integration here and noticed I was missing the aws_lambda_permission resource. This is needed to give permission to api gateway to invoke the lambda function.
For those that aren't using terraform here is a link to the aws docs on how to create the appropriate permissions.
please try to
Give execute lambda permission API Gateway
tick checkbox : Use Lambda Proxy integration
Handle null pointer for query string, headers & body.
I solved the issue by adding "isBase64Encoded": False/True to my lambda response
results = {
"statusCode": 200,
"headers": {"Content-Type": "application/json"},
"body": json.dumps(res),
"isBase64Encoded": False
}
In my case, the issue resolved while adding the integration Response and redeploying API