I am building a chatbot but I got stuck fetching data from Firebase from Dialogflow.
I want a user to input his name, the bot to query Firebase and check if the name exists. If it does "user exists" else "user n not in the database".
[this is the code am using presently][1]
function SaveName(agent){
const number = agent.parameters.number; // when I input the number in Fialogflow
const docRef = db.collection('names').doc(sessionId);
return docRef.get()
.then(doc => {
if (!doc.exists) {
agent.add('No data found in the database!');
console.log(doc);
} else {
agent.add(doc.data().orders);
}
return Promise.resolve('Read complete');
}).catch(() => {
agent.add('Error reading entry from the Firestore database.');
agent.add('Please add a entry to the database first by saying, "Write <your phrase> to the database"');
});
}
This is possible by configuring the fulfillment for your agent.
Go to Intents, select your intent and then scroll down to the fulfillment section.
Click on Enable webhook call for this intent (see Custom webhook).
Create the webhook (you can enable the inline editor at Fulfillment - Inline Editor to do so).
In your webhook, query Firestore (see this answer).
Note that anyone will be able to know if a user exists or not in your db.
If you are building for the Google Assistant, have a look at account linking.
I was facing some similar issue where the fetched details were not being shown in the agent interaction. This happened with the template given here too. I then just removed the responses from the ReadFromFirestore Intent and was getting the expected fetched queries back. I again tried by putting the response back and calling the intent and the response was sort of alternating between the default response and the firestore query response.
Still, not keeping it worked for me and thought would help others too.
Related
I'm working on converting some existing bots to skills so that we can create a Skill Host bot that can call multiple "child bots". I used the skills-simple-bot-to-bot sample, and it kind of works. I can call the skill bot from the host and it works...for a single turn. But the host is creating a new conversation ID every turn, clearing the conversation state which makes my multi-turn waterfall dialogs not work. It is also causing any other values stored in conversation state to be cleared (e.g. I capture email address in conversation state so I don't have to reprompt it). The skill bot works fine when invoked directly (i.e. not through the skill host), so it's definitely something with the way the skill host works. I found this issue on GitHub which seems to be the same thing (that's for .NET but there are links to the other SDKs including js), though it seems to have been resolved years ago yet I'm still having the issue so I'm not sure it's the same.
So in short, how do I set up the skill host so that it doesn't generate new conversation IDs every turn?
Not sure this will help as it's almost exactly the same as the sample, but here's a snippet of the code where it appears conversation ID is being generated by the host. It seems maybe I need some way to NOT call createSkillConversationIdWithOptions, but if that's what is required I'm not sure what I need to do to generate the skillConversationId other than what is currently in the sample.
async sendToSkill(context, targetSkill) {
// NOTE: Always SaveChanges() before calling a skill so that any activity generated by the skill
// will have access to current accurate state.
await this.conversationState.saveChanges(context, true);
// Create a conversationId to interact with the skill and send the activity
const skillConversationId = await this.conversationIdFactory.createSkillConversationIdWithOptions({
fromBotOAuthScope: context.turnState.get(context.adapter.OAuthScopeKey),
fromBotId: this.botId,
activity: context.activity,
botFrameworkSkill: this.targetSkill
});
// route the activity to the skill
const response = await this.skillClient.postActivity(this.botId, targetSkill.appId, targetSkill.skillEndpoint, this.skillsConfig.skillHostEndpoint, skillConversationId, context.activity);
// Check response status
if (!(response.status >= 200 && response.status <= 299)) {
throw new Error(`[RootBot]: Error invoking the skill id: "${ targetSkill.id }" at "${ targetSkill.skillEndpoint }" (status is ${ response.status }). \r\n ${ JSON.stringify(response.body) }`);
}
}
i am using the microsoft bot sdk in combination with an restify server (In package.json: "botbuilder": "^4.11.0"). I start a waterfall dialog that triggers a long running API. I save the the conversation reference and the id of the sent message to create an reply after the API call is completed:
replyToId = (await stepContext.context.sendActivity({ attachments: [ac]})).id; (in Dialog)
this.conversationReference = TurnContext.getConversationReference(context.activity); (in bot.ts)
After the completion of the API call, I want to create a reply to the last message of the dialog:
await this.adapter.continueConversation(this.conversationReference, async turnContext => {
await turnContext.sendActivity(newMessage);
});
newMessage is the Activity-object that contains further information for the user about the result of the API call.
The problem is that newMessage is not displayed as an reply to the existing message but as a separate message, although newMessage.replyToId is set to this.replyToId:
Additional information: Both messages, the last of the dialog and the "reply" are adaptive cards, but it does not make a difference if I send just simple text, same behaviour.
Would be grateful for any help :)
Instead of using "replyToId", put the id of the message you want to reply to, at the end of the conversation reference. As an example, if your conversationReference has a conversationId of 19:ac....cf#thread.skype, change it to: 19:ac...cf#thread.skype;messageid=12345678, where 12345678 is what you are currently using for "replyToId"
I'm trying to make an app using DialogFlow which finds a specific object in a specific place.
This is a generic example.
The user would say something like "Where to I find Dog in Europe" and the app would reply with "Dog can be found in Europe via: breeding, finding it out in the wild or by buying it"
considering Dog as input1 and europe as input2
Ideally the app should be able to cross reference input1 and input2 to find the correct response. Can I implement a database like structure and do this?
You can't access a database from Dialogflow directly, but you can build your own fulfillment backend that can do anything you want. It communicates with Dialogflow via HTTP requests/responses in the Dialogflow Webhook format.
Here is an example fulfillment that reads data from Firebase database - https://github.com/actions-on-google/dialogflow-updates-nodejs
You can't access a database directly in Dialog flow, but you can build your own fulfillment back end. I have been using Airtable as a database and Integromat and Webhooks to query the database and parse the results back to Dialogflow. As a novice coder I found this to be the simnplest way.
KaySubb is right, you can make a fulfillment that reads data from a firebase database(or firestore).
You can do this turning on fulfillment at the bottom page of the intent page.
First go to https://console.firebase.google.com/ (login with google account) and you should be able to see your google cloud platform project.
To use firebase, you need to first install it. Get node.js as you need npm first. I'm not sure what OS you're on but go into command line or terminal and type.
npm install firebase --save
then type:
firebase login
this will authenticate your login and connect your project when you deploy.
Then use go to the directory you want to create your project in:
firebase init functions
Select your project and select javascript, install all dependencies
Now go to functions and open the index.js file. Here you can change you write code needed in js.
Write your functions and type:
firebase deploy
in the command line open in the file directory. When it completes, it will
give you a link. This as the webhook URL in dialogflow (it should start with
https://us-central). If you see only 1 link which says
console.firebase.google.com....... then open that link on a browser, click on
"functions" on the left side of the screen and get the link from there.
This should get you started with firebase, now you can link your project to firebase fulfillment. There is great firestore explanation here
https://www.youtube.com/watch?v=kdk6MhhI8oc
But I'll give you a brief explanation:
On the top of your index.js file you will need:
const functions = require('firebase-functions');
var admin = require("firebase-admin");
admin.initializeApp(functions.config().firebase);
var firestore = admin.firestore();
The basic code is here:
exports.webhook = functions.https.onRequest((request, response) => {
switch(request.body.result.action){
case 'saveData':
let params = request.body.result.parameters
firestore.collection('colName').doc('docName').add({
name:params.name
age:params.age
}).then(() => {
response.send({
speech:
`this is a response for "${params.name}".`
});
})
.catch((e => {
console.log('Error getting documents', e);
response.send({
speech:
`Sorry, something has gone wrong. Try again and if the problem persists, please report it.`
});
}))
break;
default:
}
})
I'll explain what it does:
You need the switch to decide which intent to do. request.body.result.action returns the action name (write this in dialogflow just above the parameters).
Once that is decided request.body.result.parameters give you the parameters from the intent. params.______ gives you the parameter.
I would definitely recommend reading the official documentation:
https://firebase.google.com/docs/firestore/quickstart
to help understand the data structure to help create the ideal database for you. Essentially a collection is a list and within that a doc is one entry. You can name them yourself of using the entries from param.
respond.send is what the bot will reply to the user, I've also shown how to use the parameters in the response.
.catch will just store any errors in the log, you can read the log in console.firebase.google.com.... open your project and click on function. There will be a place to read logs there. You can check any errors encountered over there.
default: will output whatever default response you wrote on dialogflow at the bottom of the intent.
Hope this helps,comment any questions. I have gone through a huge amount as concisely as I could. This will take some time to get used to and become good at, follow the docs and the youtube videos if you have a lot of trouble!
If you're having even more trouble, there is a slack that helps people that I can direct you to.
In the example bellow, is there a way to get the user id (uid) of the user who wrote to 'offers/{offerId}'? I tried to do as described here but it doesn't work in Firestore.
exports.onNewOffer = functions.firestore
.document('offers/{offerId}')
.onCreate(event => {
...
});
I was struggling on this for a while and finally contacted the firebase Support:
The event.auth.uid is undefined in the event object for firestore database triggers. (It works for the realtime Database Triggers)
When I console.log(event) I can’t find any auth in the output.
The official support answer:
Sorry the auth is not yet added in the Firestore SDK. We have it listed in the next features.
Keep an eye out on our release notes for any further updates.
I hope this saves someone a few hours.
UPDATE:
The issue has been closed and the feature will never be implemeted:
Hi there again everyone - another update. It has been decided that unfortunately native support for context.auth for Firestore triggers will not be implemented due to technical constraints. However, there is a different solution in the works that hopefully will satisfy your use case, but I cannot share details. On this forum we generally keep open only issues that can be solved inside the functions SDK itself - I've kept this one open since it seemed important and I wanted to provide some updates on the internal bugs tracking this work. Now that a decision has been reached, I'm going to close this out. Thanks again for everyone's patience and I'm sorry I don't have better news. Please use the workaround referenced in here.
Summary of how I solved this / a workable solution:
On client
Add logged in/current user's uid (e.g. as creatorId) to entity they're creating. Access this uid by storing the firebase.auth().onAuthStateChanged() User object in your app state.
In Firebase Firestore/Database
Add a Security Rule to create to validate that the client-supplied creatorId value is the same as the authenticated user's uid; Now you know the client isn't spoofing the creatorId and can trust this value elsewhere.
e.g.
match /entity/{entityId} {
allow create: if madeBySelf();
}
function madeBySelf() {
return request.auth.uid == request.resource.data.creatorId;
}
In Firebase Functions
Add an onCreate trigger to your created entity type to use the client-supplied, and now validated, creatorId to look up the creating user's profile info, and associate/append this info to the new entity doc.
This can be accomplished by:
Creating a users collection and individual user documents when new accounts are created, and populating the new user doc with app-useful fields (e.g. displayName). This is required because the fields exposed by the Firebase Authentication system are insufficient for consumer app uses (e.g., displayName and avatarURL are not exposed) so you can't just rely on looking up the creating user's info that way.
e.g. (using ES6)
import * as functions from 'firebase-functions'
import * as admin from 'firebase-admin'
const APP = admin.initializeApp()
export const createUserRecord = functions.auth.user()
.onCreate(async (userRecord, context) => {
const userDoc = {
id: userRecord.uid,
displayName: userRecord.displayName || "No Name",
avatarURL: userRecord.photoURL || '',
}
return APP.firestore().collection('users').doc(userRecord.uid).set(userDoc)
})
Now that you have a validated creatorId value, and useful user objects, add an onCreate trigger to your entity type (or all your created entities) to look up the creating user's info and append it to the created object.
export const addCreatorToDatabaseEntry = functions.firestore
.document('<your entity type here>/{entityId}')
.onCreate(async (snapshot, context) => {
const userDoc = await APP.firestore().collection('users').doc(snapshot.data().creatorId).get()
return snapshot.ref.set({ creator: userDoc.data() }, { merge: true })
})
This clearly leads to a lot of duplicated user info data throughout your system -- and there's a bit of clean up you can do ('creatorId` is duplicated on the created entity in the above implementation) -- but now it's super easy to show who created what throughout your app, and appears to be 'the Firebase way'.
Hope this helps. I've found Firebase to be super amazing in some ways, and make some normally easy things (like this) harder than they 'should' be; on balance though am a major fan.
The documentation states clearly that the context.auth param is only available in the Realtime Database.
This field is only populated for Realtime Database triggers and
Callable functions. For an unauthenticated user, this field is null.
For Firebase admin users and event types that do not provide user
information, this field does not exist.
Personally I realized that I had the userId already in the path of my data.
export const onCreate = functions.firestore.document('docs/{userId}/docs/{docId}')
.onCreate((snapshot, context) => {
const userId = context.params.userId;
Until this is added to firestore functions a workaround is to add the user_id as a field when creating a document then deleting after. You can then grab it in the function onCreate then after you use it for what you need it for, while still in the function, just delete the field from that document.
As already suggested above, the workaround will be to add a user_id field with the data itself and then read it on the server.
The drawback with this approach will be a security loophole. As we are not verifying the user id on the server, any other user will be able to impersonate other users by sending their id with the data.
For security critical applications, the solution for this will be to use security rules to verify that the correct user_id has been sent with the data
allow write: if resource.data.user_id == request.auth.uid;
You could add your data via Callable function, from which you can access current user id:
exports.addNewOffer = functions.https.onCall(async (data, context) => {
const uid = context.auth.uid
const writeResult = await admin.firestore().collection('offers').add({ ... })
...
return { id: writeResult.id, ... }
})
What is about snap._fieldsProto.uid.stringValue
Example:
exports.hello = functions.firestore.document('hello/{worldId}').onCreate((snap, context) => {
console.log(snap._fieldsProto.uid.stringValue)
});
This should do the trick:
exports.newMessage = functions.firestore
.document('groups/messages/{messageId}')
.onCreate((snap, context) => {
var fromId = snap.data().fromId;
You can get the current signed-in user tokenId by calling getIdToken() on the User: https://firebase.google.com/docs/reference/functions/functions.auth.UserInfo
I use this trick (thanks #Robban) to publish a Contentful entry through the API, without triggering the webhook.
However, I could not figure out how to UNpublish an entry through the API without triggering the webhook.
According to the Contentful documentation, to unpublish an entry through the API, it goes like this:
client.getSpace('<space_id>')
.then((space) => space.getEntry('<entry_id>'))
.then((entry) => entry.unpublish())
As <entry_id> is the only payload, how could I indicate to the webhook it should not proceed as usual, as it was an API call?
There is unfortunately, again, no difference between a call from the API directly or from the web app. The web app does exactly this call under the hood.
Further more, in the case of an unpublish the only thing your webhook would recieve is a deletion object which does not contain any fields. This means that the trick shown in the previous answer does not apply here.
The only way I can think of solving this would be to make another call to some data store (could be Contentful) and put the entry id and perhaps also some timestamp in there. Your webhook could then upon recieving an unpublish event query this datastore and see if processing should continue or if it seems the unpublish was made through the web app.
Basically something like this:
client.getSpace('<space_id>')
.then((space) => space.getEntry('<entry_id>'))
.then((entry) => {
otherService.SetUnpublishedThroughManualAPICall(entry.sys.id);
entry.unpublish();
})
Then in your webhook with some pseudo code:
function HandleUnpublish(object entry) {
if(OtherService.CheckIfManualUnpublish(entry.sys.id)){
//Do some processing...
}
}
You could opt to use a field in Contentful as your store for this. In that case you would just before unpublishing set this field. Something like this:
client.getSpace('<space_id>')
.then((space) => space.getEntry('<entry_id>'))
.then((entry) => {
entry.fields['en-US'].unpublishedTroughApi = true;
entry.update();
})
.then((entry) => entry.unpublish())
Then in your webhook you would have to fetch the entry again via the management API and inspect the field. Keep in mind that this would result in a number of extra API calls to Contentful.