weixin_39853155的博客

Hello

Im also facing the same issue while running basic example using Google API Recognition. Im getting below console error: A moment of silence, please... Traceback (most recent call last): File "D:\Python\lib\site-packages\speech_recognition__main__.py", line 8, in with m as source: r.adjust_for_ambient_noise(source) File "D:\Python\lib\site-packages\speech_recognition__init__.py", line 532, i n adjust_for_ambient_noise buffer = source.stream.read(source.CHUNK) File "D:\Python\lib\site-packages\speech_recognition__init__.py", line 190, i n read return self.pyaudio_stream.read(size, exception_on_overflow=False) File "D:\Python\lib\site-packages\pyaudio.py", line 608, in read return pa.read_stream(self._stream, num_frames, exception_on_overflow) OSError: [Errno -9988] Stream closed

My System is Win 7 Professional. My Python version is 3.6.3. My SpeechRecognition library version is 3.8.1. My PyAudio library version is 3.8.1. My Audio devices are(I tested with default audio device): Microphone with name "Microsoft Sound Mapper - Input" found for Microphone(devi ce_index=0) Microphone with name "Microphone (High Definition Aud" found for Microphone(dev ice_index=1) Microphone with name "Microsoft Sound Mapper - Output" found for Microphone(dev ice_index=2) Microphone with name "Speakers (High Definition Audio" found for Microphone(dev ice_index=3)

Can you please help me resole this issue?

各位大佬的讨论让我学习到了很多，原来总觉得要一次把副本全拷贝出来，没想到还可以单独拷一个，下次接着拷，在仅仅可能导致单次显示不完全的情况下减少了内存占用，学习了

伪命题，产品中谁没事执行list_thread? 甚至我都怀疑有多少人（除了睿赛德）会在产品中启用finsh组件？ 另一个，不说100%的人吧，99.9%的人执行list_thread只会发生在debug阶段。 so，问题不大。顶多是defect，而不是bug. 其实这个问题就和#2023类似的。

伪命题，产品中谁没事执行list_thread? 甚至我都怀疑有多少人（除了睿赛德）会在产品中启用finsh组件？ 另一个，不说100%的人吧，99.9%的人执行list_thread只会发生在debug阶段。 so，问题不大。顶多是defect，而不是bug. 其实这个问题就和#2023类似的。

其实这个就是在一个实际的控制系统上被发现的，有个远端监控系统调用ps等命令询问控制器状态，结果昨天更新完rtt上演了一出好戏

ok, we'll wait.

Thanks for your work!

Lag On init Windows 10, Chrome 63.0.3239.84

+1 very needed feature ;)

trace.txt

Using config /Users/xxx/temp/redex/default.config Using binary (default) Extracting apk... Detected dex mode Api21DexMode Unpacking dex files Detecting Application Modules Unpacking APK finished in 0.20 seconds Running redex-all on 1 dex files Running redex binary at /tmp/redex.ImKeXy/redex-all dup_class_whitelist: { Lcom/facebook/soloader/MergedSoMapping; Ljunit/framework/TestSuite; L$r8$java8methods$utility } Loading classes from dex from /tmp/redex.ImKeXy/tmpiuupw2xg.redex_dexen/dex0/classes.dex Load classes from dexes completed in 0.2 seconds Deobfuscating dex elements completed in 0.0 seconds Process keep for whyareyoukeeping completed in 0.0 seconds Process keep for classes and members completed in 0.0 seconds Process keep for assumenosideeffects completed in 0.0 seconds Processing proguard rules completed in 0.0 seconds No Optimizations Rules completed in 0.1 seconds Initializing reachable classes completed in 0.0 seconds Redex_frontend completed in 0.4 seconds API Level Checker completed in 0.0 seconds Evaluating ShortenSrcStringsPass... ShortenSrcStringsPass (eval) completed in 0.0 seconds Hasher completed in 0.0 seconds Running ShortenSrcStringsPass... src strings shortened 513, 11441 bytes saved Hasher completed in 0.0 seconds ShortenSrcStringsPass (run) completed in 0.0 seconds Running IRTypeChecker...

deleting the .m2 directory (thx ) fixed this for me.

thanks!!!!

Hey Justin,

I'd be happy to verify the issue for you. Just let me know when you have something ready and I'll test it right away.

Thanks!

Formtastic 2.0.0.rc3 indeed resolves the issue for me. I was doing some custom wizardry to allow a date-picker field type, which does not work "out of the box" with 2.0.0.rc3, so I'll have to tackle that before upgrading.

Thank you for the quick response and resolution (and for working on such a useful library)!

I also have the problem of VimR not finding init.vim.

:healthcheck shows:

but the file is definitely there and can be loaded manually with :source.

LANG="en_US.UTF-8"
LC_COLLATE="en_US.UTF-8"
LC_CTYPE="en_US.UTF-8"
LC_MESSAGES="en_US.UTF-8"
LC_MONETARY="en_US.UTF-8"
LC_NUMERIC="en_US.UTF-8"
LC_TIME="en_US.UTF-8"
LC_ALL=

Strangely, it works if I run VimR as root. The permissions seem fine however

-rwxrwxrwx  1 james  staff   111B Mar 29 10:24 /Users/james/.config/nvim/init.vim

I'm still seeing this on 0.9.3

-007 In our case it was because we were referencing a deprecated AWS provided policy. Check if the policyies your are attaching are AWS provided and deprecated. If so attach policy is doing very weird.

We have the same problem with aws_iam_role_policy_attachment. It does attach the policy to the role (verified in aws console) but tf fails.

When applying changes to module.ssm_patch_management.aws_iam_role_policy_attachment.attach_ssm_policy_to_role3, provider "aws" produced an unexpected new value for was present, but now absent.

Note that we actually attach 5 policies. It only complains on one.

Hi ronag,

I also faced the problem that the decoding performance is not well on some mobile devices (even slower than javascript h264 decoder). Is it possible to share your revised code?

Thanks!

Unfortunately graph=None to the FileWriter didn't solve the issue, running quite quickly out even with just a few models.

Yes this can be closed. I think the documentation created in #1241 is excellent and sufficient. I followed it exactly as shown and it worked. Thanks.

I solved my problem using the information in #1241. This new way of adding serialization support to POROs helped.

I'm inching closer. But now I can't get it to include my relationships...

My lame poro:

class Settings
  attr_accessor :id, :foo, :membership, :organization, :user

  def self.model_name
     ||= ActiveModel::Name.new(self)
  end

  def initialize(hash={})
     = hash
  end

  def cache_key
    "#{self.class.name.downcase}/#{self.id}-#{self.updated_at.strftime("%Y%m%d%H%M%S%9N")}"
  end

  def cache_key_with_digest
    "#{cache_key}/#{FILE_DIGEST}"
  end

  def updated_at
    [:updated_at] ||= DateTime.now.to_time
  end

  def read_attribute_for_serialization(name)
    if name == :id || name == 'id'
      id
    else
      [name]
    end
  end

  def id
    [:id] || ['id'] || object_id
  end

  def method_missing(meth, *args)
    if meth.to_s =~ /^(.*)=$/
      [$1.to_sym] = args[0]
    elsif .key?(meth)
      [meth]
    else
      super
    end
  end
end

My serializer:

class SettingsSerializer < ActiveModel::Serializer
  attributes :foo

  has_one :membership, serializer: MembershipSerializerWithOrganization
  has_one :organization, serializer: OrganizationSerializer
  has_one :user, serializer: UserSerializerWithMemberships

end

The result (note that the relationships are always null). Perhaps I no longer have an issue with poros anymore, but its this issue? #1180

{
    "data": {
        "attributes": {
            "foo": null
        },
        "id": "me",
        "relationships": {
            "membership": {
                "data": null
            },
            "organization": {
                "data": null
            },
            "user": {
                "data": null
            }
        },
        "type": "settings"
    }
}

Sorry for being so late following up, but yes once I put the include in the serializer options it worked. Thanks! Now I have to figure out why its not wrapping it using "data" and "includes"

I'm using grape and grape-active_model_serializers. My grape endpoint returns the PORO, and grape-active_model_serializers calls the serializer/adapter.

Ah, I didn't realize that the convention had changed from the include directive being in the serializer to being on the render. I'm not sure if grape-active_model_serializers will pass that include option along for me, but I can try and I'll get back to you.

Here's my model:

 ruby
module Poros
  class Settings
    # include ActiveModel::SerializerSupport
    attr_accessor :id, :membership, :organization, :user

    def initialize(attributes)
       = attributes[:id]
       = attributes[:membership]
       = attributes[:organization]
       = attributes[:user]
    end

    def self.model_name
       ||= ActiveModel::Name.new(self)
    end

    def initialize(hash={})
       = hash
    end

    def cache_key
      "#{self.class.name.downcase}/#{self.id}-#{self.updated_at.strftime("%Y%m%d%H%M%S%9N")}"
    end

    def cache_key_with_digest
      "#{cache_key}/#{FILE_DIGEST}"
    end

    def updated_at
      [:updated_at] ||= DateTime.now.to_time
    end

    def read_attribute_for_serialization(name)
      if name == :id || name == 'id'
        id
      else
        [name]
      end
    end

    def id
      [:id] || ['id'] || object_id
    end

    def method_missing(meth, *args)
      if meth.to_s =~ /^(.*)=$/
        [$1.to_sym] = args[0]
      elsif .key?(meth)
        [meth]
      else
        super
      end
    end
  end
end

Here's my serializer:

 ruby
class SettingsSerializer < ActiveModel::Serializer
  attributes :id
  has_one :membership, embed_in_root: true, serializer: MembershipSerializerWithOrganization
  has_one :organization, embed_in_root: true, serializer: OrganizationSerializer
  has_one :user
end

Here's a sample of the JSON I get as a response. You can see it doesn't look very JSONAPI-like. Additionally, the serializers for the related ActiveModel objects were not used. Those are simple to_json hashes.

{
    "attributes": {
        "id": "me",
        "membership": {
            "created_at": "2015-07-14T14:21:48.559Z",
            "id": "a86b662e-d484-4692-a68b-94a92b3a3968",
            "organization_id": "ca8cf9b0-47fc-4937-99e4-8a7eaa2a42f8",
            "role": "superadmin",
            "updated_at": "2015-07-14T14:21:48.559Z",
            "user_id": "f673f449-681a-4945-9545-9528c2840bab"
        },
        "organization": {
            "country_code": "CA",
            "created_at": "2015-07-14T14:21:48.390Z",
            "id": "ca8cf9b0-47fc-4937-99e4-8a7eaa2a42f8",
            "lft": 1,
            "municipality": "Toronto",
            "name": "Prestige Worldwide",
            "parent_id": null,
            "postal_code": "H0H0H0",
            "properties": null,
            "rgt": 6,
            "status": "active",
            "street_address_1": "80 North St",
            "street_address_2": null,
            "subdivision_code": "ON",
            "updated_at": "2015-07-14T20:12:01.057Z"
        },
        "user": {
            "created_at": "2015-07-14T14:21:48.534Z",
            "email": "superadmin.com",
            "id": "f673f449-681a-4945-9545-9528c2840bab",
            "status": "active",
            "updated_at": "2015-07-14T14:21:48.534Z"
        }
    }
}

Hi I actually tried adding these methods to my poro, but I had some difficulty because my poro is just a container for 3 activemodel objects. I can't remember the exact result I got, but it didn't seem to serialize those AM objects properly.

Ah yes to_json alone works. I thought I tried that at one point but had trouble. I can submit as a PR. Where would you like it in the namespace?

Oh and BTW, my workaround for this problem at the moment was to subclass the JsonApi adapter to simply skip links for now...

module API
  class AmsJsonapiAdapter < ActiveModel::Serializer::Adapter::JsonApi
    def links_for(serializer, options)
      {}
    end
  end
end

I was using a fork of grape-active_model_serializers https://github.com/Thanx/grape-active_model_serializers.git, but once I started using the jsonapi adapter, I built my own very simple grape json formatter:

module API
  module AmsFormatter
    class << self
      def call(resource, env)
        serializer_options = {}
        serializer_options.merge!(env['ams_serializer_options']) if env['ams_serializer_options']
        # For some reason serializable_hash doesn't include meta in the response, so I have to use as_json
        ActiveModel::SerializableResource.new(resource, serializer_options).as_json.to_json
      end
    end
  end
end

I also had to create this little grape helper to pass my serializer options to the formatter via the env:

        def render(resources, ams_serializer_options = {})
          env['ams_serializer_options'] = ams_serializer_options
          resources
        end

So now I can implement a grape endpoint like this and put metadata into the jsonapi response:

    formatter :json, API::AmsFormatter

     get ':id' do
       render(Resource.find(params[:id]), meta: { foo: 'bar' })
     end

Looking at the code, I can't even see how this would work for ActionController. The SerializableResource filters the adapter options, so that they must be one of [:include, :fields, :adapter, :meta, :meta_key, :links]. Which means there's no way the adapter options could contain a :context entry.

added rubocop corrections to pull request

:umbrella: The latest upstream changes (presumably #4625) made this pull request unmergeable. Please resolve the merge conflicts.

改了 batch_first 参数，暂时转换没有报错 以后有问题再补充

你好，请问是改torch中的batch_first参数么？ nn.LSTM(nIn, nHidden, bidirectional=True,batch_first = True)

你好，请问你改batch_first =True成功了么，我改完还是报同样的错

that looks solid! perfecto.

Is it beyond the scope? It already creates a map element that is its own HTML page. It wouldn't be a horrible option and would be super useful for titles. As I'm programmatically building multiple maps, I don't want to hand write any of the files.

One could presumably import the map file into a more robust page. I'm serving them as their own pages individually.

Ex: http://coletownsend.github.io/ben-cole-project/

Woo thanks!

Ah sorry man! It works fine when I spin up a local server, but just looking at it locally (i.e. as a file in safari) it doesn't work.

:umbrella: The latest upstream changes (presumably #6531) made this pull request unmergeable. Please resolve the merge conflicts.

:umbrella: The latest upstream changes (presumably #5114) made this pull request unmergeable. Please resolve the merge conflicts.

Getting this on Android 7.0, Nexus 6p.

:hourglass: Testing commit 865ec52 with merge 9ca3afe...

:pushpin: Commit 865ec52 has been approved by indirect

:hourglass: Testing commit 1afd8df with merge 8129ffb...

:hourglass: Testing commit 1afd8df with merge a534931...

:pushpin: Commit 1afd8df has been approved by indirect

:bulb: This pull request was already approved, no need to approve it again.

• This pull request previously failed. You should add more commits to fix the bug, or use retry to trigger a build again.
• There's another pull request that is currently being tested, blocking this pull request: #6570

:hourglass: Testing commit 1afd8df with merge ff56a6d...