We build a NN model in tensorflow similar to
https://github.com/GoogleCloudPlatform/cloudml-samples/tree/master/census/tf-keras
Then we exported it to GCP using the following gcloud ai-platform models create commands
and predict using gcloud ai-platform predict works fine
Now we want to do online prediction using javascript, seems to have several options:
using https://www.npmjs.com/package/#google-cloud/automl, but
Checking https://googleapis.dev/nodejs/automl/latest/v1beta1.PredictionServiceClient.html#predict
it seems that it can not be used for neural networks(?)
When we run the example code
https://www.npmjs.com/package/#google-cloud/automl
with our parameters we get 'Invalid resource ID' error
Using REST API, curl works fine but how do we set permanent authorization in the server ?
Any other suggestion ?
Amir
I solved the problem by building Python container that run on gcloud
(https://cloud.google.com/run/docs/quickstarts/build-and-deploy)
The JS call the container using REST API, the python container do the prediction and reply to results the JS
Related
I don't know where to start with this question!
Basically, I would like to build a control panel, web based, using node.js as the back end. This would run on my raspberry pi running Ubuntu Server.
The idea is that I can gain statistics (CPU, Temperature, Disk Space etc) and set up basic features like MongoDB database, hosting etc.
Now, this is obviously possible just like many web panels out there, however, is this possible with node.js.
I suppose the first question would be, can I start/stop services (reboot server, start MongoDB etc) via Node.Js, and secondly, can I get info from this to display in my web panel?
I tried Google but for the first time, I don't even know what question to ask :)
I found Node JS examples of running command line commands, however, when passing in simple commands, like "node -v" it crashed, so I am not sure this is the method used by other commercial server web panels.
Any advice would be great :)
Thanks
You should try this tutorial: https://stackabuse.com/executing-shell-commands-with-node-js/
From node doc for child_process:
https://nodejs.org/api/child_process.html
const { exec } = require('child_process');
exec('"/path/to/test file/test.sh" arg1 arg2');
// Double quotes are used so that the space in the path is not interpreted as
// a delimiter of multiple arguments.
exec('echo "The \\$HOME variable is $HOME"');
// The $HOME variable is escaped in the first instance, but not in the second.
I am new to elasticsearch and I am trying to build a movie search app. For this I plan to get data from kaggle and add to my elasticsearch which I have setup locally at localhost:9200. I see this in the localhost link:
name "bxiIZLL"
cluster_name "elasticsearch"
cluster_uuid "zc_JPmw4TQ2G5bvahEF6LQ"
version
number "5.6.14"
build_hash "f310fe9"
build_date "2018-12-05T21:20:16.416Z"
build_snapshot false
lucene_version "6.6.1"
tagline "You Know, for Search"enter code here
Now I need to add Kaggle data to this server. How can I do it?
I saw somewhere the curld -XPUT command. I am not sure how that can work with Kaggle.
A follow up question - if I want to publish my app later on, how can I host the elasticsearch ?
In order to upload a CSV file to elasticsearch:
download the file.
use logstash in order to read the file using file input
modify and transform the data as you need using logstash's CSV filter
output logstash to elasticsearch
For your follow up question - how can I host Elasticsearch - you can either run it by your own, in AWS EC2 for example, or use a managed service like Elastic cloud or AWS ES. good luck
I searched all over the web looking for the C# .NET classes and documentation to allow me to code functional equivalent of this Python method and I can't believe it simply doesn't exist but NuGet hasn't helped and the closest Google documentation I can find is here.
What does exist is classes and examples that show how to authenticate a .NET server or Installable with Google so that it can Access various APIs (Drive etc) but I just want the simple client token Authentication that is referred to for most other languages here
As you will see Python has the google.oauth2.id_token class with its verify_oauth2_token() method and NodeJS and PHP have their verifyIdToken() methods.
Java gets a little more complicated with: -
import com.google.api.client.googleapis.auth.oauth2.GoogleIdToken;
import com.google.api.client.googleapis.auth.oauth2.GoogleIdToken.Payload;
import com.google.api.client.googleapis.auth.oauth2.GoogleIdTokenVerifier;
GoogleIdTokenVerifier verifier = new GoogleIdTokenVerifier.Builder(transport, jsonFactory)
// Specify the CLIENT_ID of the app that accesses the backend:
.setAudience(Collections.singletonList(CLIENT_ID))
// Or, if multiple clients access the backend:
//.setAudience(Arrays.asList(CLIENT_ID_1, CLIENT_ID_2, CLIENT_ID_3))
.build();
// (Receive idTokenString by HTTPS POST)
GoogleIdToken idToken = verifier.verify(idTokenString);
But C# has diddly :-( Please advise!
If it helps I'm implementing a Javascript SSO client. (BTW have no interest in the YOLO library)
I finally found Google C# token verification code on the web. It says it is no longer actively maintained. Why is there no MuGet package? Why is Google making this so hard?
Or just call the Google Validator like Python does?
Also see previous answer Here
I am trying to learn how to process payments with Square, and am trying to run their examples from GitHub to get a feel regarding how to structure the payments application. I built the node example from here: https://github.com/square/connect-api-examples/tree/master/connect-examples/v2/node_payment using npm install and npm build to get the app up and running.
I am using "4532759734545858" for the card number, "123" for CVV, "0120" for expiration, and "94103" for the zip. I got the card number from here: https://docs.connect.squareup.com/articles/using-sandbox where it states that this is a good number to use for a Visa sandbox.
Also, I have updated the config.json with properties from my developer settings.
When trying to process a payment a get a DOM element that says "Card Declined" without further specifying the error. Is there something I can do to parse the error?
Based on the documentation at: https://docs.connect.squareup.com/articles/using-sandbox#generatingerrorstates it seems the amount_money field of the request is not being populated, but I am having trouble confirming.
Ideally I would like to get to a point where I can add a card as a hash value to my db and use it for recurring billing...
That "card declined" message is actually the error you get back from Square's APIs. You can play around with the error messaging in the app.js file and the `error.jade. Try error.catagory, code, detail.
Keep in mind that this is just a sample app, to show that you can use the APIs with node.js, you probably don't want to use this code in your production system.
Suppose I want to execute 'db.things.insert({colors : ["blue", "black"]})' in browser. I can execute it in the Mongodb -shell but not yet understanding how to execute it like: open up the Google Chrome Console, initialize the DB -object with some connection and execute the command. Does there exist some plugin? Sorry I am totally new to MongoDB, trying to test just things fast with Browser shell. How can I do the initialization like that?
Trial 0: perhaps with REST -interface?
I have enabled the REST with "$ echo 'rest=true' > /etc/mongodb.conf;
$ sudo restart mongodb", works in Ubuntu. More about rest
here, not sure
yet whether needed here but perhaps with some POST/REST -method I can
do the init.
Trial 1: Oreilly's book about MongoDB and 50 Tips (page 47)
The book has some example
> db = connect ("ny1a:27017/foo")
> db = connect ("ny1a:27017/admin")
so now
> db=connect("localhost:27017/test")
ReferenceError: connect is not defined
Yes because I need to source the connect -command, some further examples here, where can I get it?
P.s. I am studying this tutorial here.
You cannot simply access mongodb from the browser console. Your browser is a client, and there isn't (as far as I know) a javascript client-side library. Any javascript library you will find will most likely be for Node.js (server).
The mongo console is its own type of compiled client. The native language is javascript, but those commands only pertain to the actual mongo command shell. connect is a command for the mongo command shell.
Enabling REST starts a port on your mongod that will accept REST http communication. You get a browser page here: http://localhost:28017/
This is a very basic page displaying data, but you can further run queries yourself. See http://www.mongodb.org/display/DOCS/Http+Interface#HttpInterface-SimpleRESTInterface
Note, the REST interface is READ ONLY. You cannot insert.
That being said, you need a proper driver for your language of choice.