weixin_39782752的博客

Hi, I'm interested in using this component but I also need reordering. Any info on this pull request ? Thanks! Paul

see above

Hello, here is api of Nuance. don't know how to combine to your code because it is python api.

import os
import requests
class Nuance_api():
    """Nuance ASR Service API"""
    def __init__(self, appid, appkey, n_id):
        self.appid = appid
        self.appkey = appkey
        self.id = n_id
    def recogenize(self, language, voicefile):
        nuance_url = 'https://dictation.nuancemobility.net/NMDPAsrCmdServlet/dictation'
        n_params = {'appId':self.appid, 'appKey':self.appkey, 'id':self.id}
        voice_file = {'file': open(voicefile, 'rb')}
        n_headers = {'Accept-Language':language, 'Content-Type':'audio/amr;codec=amr;bit=4.75;rate=8000', 'Accept':'text/plain', 'Accept-Topic':'Dictation'}
        r = requests.post(url=nuance_url, params=n_params, headers=n_headers, files=voice_file, verify=False)
        if r.status_code == 200 :
            return r.text

My requirement is that, support multiple languages, large quota as much as possible. Guessing lots of people have the same. The problem of AT&T and wit.ai is less languages support. I do not yet try how IBM service is. For app dev on Android or iPhone, there is also another service which is ispeech.org can be used for free. I will have a look at IBM and Amazon Alexa to see if they fulfill my needs. And I noticed that your API is created in binary, that will cause me stop integrate Nuance API into it.

BTW, Nuance has better API support, however I am looking to find a right way to use it since the doc and support give conflict information about request URL. If you have interesting, suggest to have a look at Nuance API here: http://nuancemobiledeveloper.com/public/index.php?task=home If I find a way to use it. Maybe I can try to combine it into your API.

sorry for late reply, I just saw the comment. And thank you for help. I got the speech API key, and unfortunately it is 50req/day limitation. Anyway, thank you so much to develop so charming API.

Well, I think it works and guessing it just support English voice.~~~

+1

I was confused by this, would be nice to see this merged!

same problem

$ g++ --version g++ (GCC) 4.1.0 (SUSE Linux)

$ make --version GNU Make 3.80

Reference：http://blog.chinaunix.net/uid-20682147-id-5078298.html

Just confirming, it now works on my dash based, and FreeBSD systems. Great job. Thanks!

okay thank you. I'm just new in contributing open source projects and have to learn more about testing systems

what's wrong with build?

Hi, No worries. Actually I don't use Bing. I prefer Google, but Bing Images is actually not half bad.

Thanks for the quick replies. I solved it on my end, so the ball is in your court :)

On Mon, Sep 28, 2015 at 1:06 PM, Ruined1 - notifications.com < github.keepmy.d2f8a97cfc.notifications#github.com.0sg.net> wrote:

https://github.com/stisev, i truly don't know a single person who uses Bing, that includes my clients, and i fix computers for a couple major corporations. Most people use Google and those who don't know how to choose are usually using Yahoo as their default. In all my career I have yet to find a single person who intentionally uses Bing, not dating there aren't people who use it, but let's not over exaggerate... That being said, I'm on a mobile right now, I will review the issue at a later time when a PC is available...

— Reply to this email directly or view it on GitHub https://github.com/10se1ucgo/DisableWinTracking/issues/93#issuecomment-143860056 .

Thanks for the feedback - Looks good from our point of view, had been stable for us since rc2.

In production using <%= javascript_packs_with_chunks_tag 'application', 'data-turbolinks-track': 'reload' %> without the vendor will also add the vendor js file.

Maybe I am unclear on the instructions for when you need to add different js files to the packs tag.

We've been using rc2 on our staging site for a couple weeks without issues.

Upgrading to rc5 and using javascript_packs_with_chunks_tag. we get a manifest error stating rails can't find vendor.js in the manifest.

Webpacker can't find vendors in /Users/rowan/www/runn-app/public/packs/manifest.json.

We were using

and updated to

Our webpack dev config is:

  mode: "development", // Gives us sane defaults
  output: {
    pathinfo: false,
    filename: "[name].js",
    chunkFilename: "[name].chunk.js"
  },
  optimization: {
    splitChunks: {
      cacheGroups: {
        commons: {
          test: /[\\/]node_modules[\\/]/,
          name: "vendors",
          chunks: "all"
        }
      }
    }
  },

And the output of json manifest (images etc removed)

{
  "application.css": "/packs/application-6789eef5.css",
  "application.css.map": "/packs/application-6789eef5.css.map",
  "application.js": "/packs/application.js",
  "application.js.map": "/packs/application.js.map",
  "entrypoints": {
    "application": {
      "js": [
        "/packs/vendors.chunk.js",
        "/packs/application.js"
      ],
      "js.map": [
        "/packs/vendors.chunk.js.map",
        "/packs/application.js.map"
      ],
      "css": [
        "/packs/application-6789eef5.css"
      ],
      "css.map": [
        "/packs/application-6789eef5.css.map"
      ]
    }
  },
  "vendors.js": "/packs/vendors.chunk.js",
  "vendors.js.map": "/packs/vendors.chunk.js.map",
}

Leaving it the original way using javascript_pack_tag works fine.

And

Also appears to work, bringing in vendor. But I am not sure if its just some weird caching issue or it is actually adding the vendor tag as well.

(this is all done in dev mode with rails s and webpack-dev-server

The other way to do that:

ActiveModel::Serializer.config.adapter = ActiveModelSerializers::Adapter::JsonApi
ActiveModel::Serializer.config.key_transform = :underscore

This is how I fix this issue

`
data = ActiveModelSerializers::SerializableResource.new(,
            serializer: CustomerSerializer,
            key_transform: :underscore)
render json: data.serializable_hash

`

Right now am just using curl: curl -i -X POST -d "_type=clicks_per_day&facebook=1" 'http://localhost:4242/events' I've not changed the default TTL for Redis. The event is definitely picked up by FnordMetric as the events_received and events_processed count increases and the queue is 0. It's just that it need one more event to get the previous events count updated in the WebUI.

It's not really a big issue for me, but I think it would be nice to have anyway, so I've opened an issue about it.

Don't know if this warrants a new issue, but a removeHeader method would be useful, I needed to remove the Accept header and using the overwriter at best nets you an empty Accept header.

I ended up injecting into the httpclient, but still, more control over headers wouldn't be bad.

thanks !