AWS |
AWS Big Data Blog |
Query your data streams interactively using Kinesis Data Analytics Studio and Python |
https://aws.amazon.com/blogs/big-data/query-your-data-streams-interactively-using-kinesis-data-analytics-studio-and-python/
|
Query your data streams interactively using Kinesis Data Analytics Studio and PythonAmazon Kinesis Data Analytics Studio makes it easy for customers to analyze streaming data in real time as well as build stream processing applications powered by Apache Flink using standard SQL Python and Scala Just a few clicks in the AWS Management console lets customers launch a serverless notebook to query data streams and get … |
2022-04-13 15:06:33 |
AWS |
AWS Machine Learning Blog |
Manage dialog to elicit Amazon Lex slots in Amazon Connect contact flows |
https://aws.amazon.com/blogs/machine-learning/manage-dialog-to-elicit-amazon-lex-slots-in-amazon-connect-contact-flows/
|
Manage dialog to elicit Amazon Lex slots in Amazon Connect contact flowsAmazon Lex can add powerful automation to contact center solutions so you can enable self service via interactive voice response IVR interactions or route calls to the appropriate agent based on caller input These capabilities can increase customer satisfaction by streamlining the user experience and improve containment rates in the contact center In both the self service … |
2022-04-13 15:09:58 |
python |
Pythonタグが付けられた新着投稿 - Qiita |
AWS Cloud9でPython基礎~ファイルの読み書き |
https://qiita.com/emiki/items/f93bddb4a1fd76b102b5
|
awscloud |
2022-04-14 00:46:39 |
python |
Pythonタグが付けられた新着投稿 - Qiita |
LSTMで分子生成入門 |
https://qiita.com/maskot1977/items/6e68b50287d80dbd4461
|
lstmlongshorttermmemory |
2022-04-14 00:03:55 |
js |
JavaScriptタグが付けられた新着投稿 - Qiita |
Prisma で Upsert Many はないのか |
https://qiita.com/dosukoi_man/items/ca3d7202a8762169edd0
|
upser |
2022-04-14 00:53:46 |
Linux |
Ubuntuタグが付けられた新着投稿 - Qiita |
60秒遅延なしでシャットダウンする方法【Ubuntu 22.04 Jammy Jellyfish】 |
https://qiita.com/relu/items/08175471f066053f39bd
|
managerlogoutpromptfalse |
2022-04-14 00:44:25 |
AWS |
AWSタグが付けられた新着投稿 - Qiita |
AWS Cloud9でPython基礎~ファイルの読み書き |
https://qiita.com/emiki/items/f93bddb4a1fd76b102b5
|
awscloud |
2022-04-14 00:46:39 |
GCP |
gcpタグが付けられた新着投稿 - Qiita |
【GCP】Cloud Buildトリガーで困った話メモ |
https://qiita.com/ricemountainer/items/173bafedfd44eefa1ac9
|
cloudbuild |
2022-04-14 00:34:58 |
海外TECH |
Ars Technica |
Microsoft’s tactics to win cloud battle lead to new antitrust scrutiny |
https://arstechnica.com/?p=1847730
|
amazon |
2022-04-13 15:48:40 |
海外TECH |
MakeUseOf |
How to Make a Free Temporary Website Within Seconds |
https://www.makeuseof.com/tag/pageeasy-free-temporary-web-page/
|
services |
2022-04-13 15:45:14 |
海外TECH |
MakeUseOf |
The 7 Best Retro Gaming Consoles You Can Buy |
https://www.makeuseof.com/tag/best-retro-gaming-consoles/
|
games |
2022-04-13 15:36:22 |
海外TECH |
MakeUseOf |
AI, Automation and Robots: 7 DIY Electronic Projects for the Kitchen |
https://www.makeuseof.com/ai-automation-and-robots-diy-electronic-projects-for-the-kitchen/
|
arduino |
2022-04-13 15:30:14 |
海外TECH |
MakeUseOf |
Why Do You Need to Reset Windows 11 to Install Build 22567? |
https://www.makeuseof.com/windows-11-build-22567-reset/
|
build |
2022-04-13 15:16:13 |
海外TECH |
MakeUseOf |
The Phenom Forge Brings Large-Format Functional Resin Printing to the Masses |
https://www.makeuseof.com/peopoly-phenom-forge-large-format-resin-printing/
|
The Phenom Forge Brings Large Format Functional Resin Printing to the MassesPeopoly s Phenom Forge resin printer incorporates improved MSLA optics and a new Nylon like resin to make large functional resin parts a reality |
2022-04-13 15:05:43 |
海外TECH |
DEV Community |
Beginner’s Guide to Diagnosing Audio Issues as Part of an Azure Serverless Media Workflow |
https://dev.to/bradenriggs/beginners-guide-to-diagnosing-audio-issues-as-part-of-an-azure-serverless-media-workflow-616
|
Beginner s Guide to Diagnosing Audio Issues as Part of an Azure Serverless Media WorkflowTranscribing media is a resource intensive process that is dependent on the quality of the audio and background noise meaning it can often produce inconsistent results as a product of the media quality Depending on the scale of the media you are transcribing this process can be both computationally and monetarily expensive being a costly endeavor especially if the results end up being inaccurate and noisy To alleviate some of the risks of transcribing audio that might be low quality we can instead turn our transcription tool into a workflow that gauges the quality of the audio and only transcribes it if the audio is high enough quality to produce accurate results Furthermore because media data can vary dramatically in size we can develop this workflow on the cloud to ensure our processing requirements are dynamically adjusted and avoid storing large quantities of data among other things Part Getting your environment ready for Serverless development To get started we first need to set up our development environment For users unfamiliar with Azure services it is worth noting that Azure has a pay as you go policy with some resources costing more than others so we recommend that you check out the pricing guide here before leveraging any of their services Follow the steps below to get the appropriate credentials and environment set up Create a free Dolby io account here Additionally you ll need to create a free Azure account With this free Azure account create a container and a storage blob for your media file here Also with your free Azure account create a cognitive services account here To get your Azure environment set up Microsoft has a handy guide for building your first project Make sure you are using Python and Azure Functions Core Tools x With the environment correctly set up you can now clone our sample project from GitHub The sample project presented at the Azure Serverless Conference is a basic example of an asynchronous and fully serverless media to transcription workflow To best understand how serverless media workflows work we will use this sample project as a template Part Why Azure Serverless Before we dive into the workflow let s briefly outline what Azure Serverless Functions are Serverless functions are a service provided by Azure that allows for event triggered code to be run without the need for personal servers or infrastructure Instead users can develop code such as a media workflow and allow it to run on the cloud in response to an event These events are referred to as Triggers which are the starting line for any Serverless process The most basic trigger to imagine is an HTTP trigger which would kickstart a Serverless event if a user navigated to a specific URL from their local machine Serverless functions are perfect for building a media workflow as they can be developed to function asynchronously meaning multiple jobs can be run at once with the associated costs scaling according to usage This means you can build a workflow that you only use once a month or a thousand times a day only paying for what you use Part Understanding media workflows with Azure Serverless Now that we understand the basics of Azure Serverless Functions let s take a look at the sample project we cloned from GitHub What does this workflow look like A depiction of the serverless media workflow available at As outlined above our workflow must begin with a trigger in this case it is a basic HTTP trigger There are two main paths the workflow follows dependent on the quality of the audio in the media provided In this example there are three main tools that the workflow relies on to ensure that the media will produce a sufficiently accurate transcription In the event that the media will not produce an accurate transcription the audio for the media is cleaned and returned to allow for manual review Dolby io Audio Diagnose A lightweight audio analysis tool that can return information relating to the audio such as its quality or audio defects Azure Cognitive Services Speech to Text A transcription tool that converts spoken audio to text Dolby io Enhance An audio enhancement tool that can help reduce noise and level speakers Before we test out the sample project let s briefly discuss how we connect each tool together in a serverless fashion Part Making things asynchronous One of the challenges of developing a Serverless media workflow is planning for asynchronous deployment meaning developing in a way that allows multiple instances of the workflow to run without them interfering or slowing each other down We already discussed triggers which are great for starting the workflow but what are some other useful tricks we can include to keep things running smoothly One useful trick is including callbacks in our API calls Because we are using REST APIs the actual processing is done on a different server In the case of the Dolby io APIs this processing is done on the Dolby io Servers whereas the Speech to Text processing is done on the Azure Cognitive Services server When we send the job to these servers we can include a callback parameter that signals where the API should send a request once the job is done Since we are using HTTP triggers we can specify that the callback directs to the trigger We don t want the workflow to begin at the start so when we callback to the HTTP trigger we include a job tag in the request destination An example of this process for diagnosis looks can be seen below import requests Python module for making http requestsurl Location to send the requestbody input presignedURL on complete url httpTriggerURL job diagnose success headers x job id headers x api key API KEY Content Type application json Accept application json x job id True response requests post url json body headers headers response raise for status Included in the body of the call is an on complete tag which includes the URL location of the trigger plus a tag informing the function that the next step it should take is evaluating the results of diagnose This process allows the function to be freed up as soon as it submits a diagnose job sitting idle until it has to start another diagnose job or until that diagnose job completes Structured like this the function is able to handle many requests sending media through the workflow The second trick for keeping things asynchronous and efficient is by not moving the media This means we don t have to change where files are stored rather we can keep them stored on the Azure cloud in the blob storage and pass directions to each of our tools so they can access where the files are stored This is done through the use of the pre signed URLs a useful concept that you can read about in more detail here Pre signed URLs allow the appropriate credentials and access to be passed in one simple URL which the Dolby io server or the Cognitive Services server can use to find the media An example of creating this URL can be seen below from datetime import datetime timedeltafrom azure storage blob import generate blob sas BlobSasPermissionsinput sas blob generate blob sas account name AZURE ACCOUNT NAME container name AZURE CONTAINER blob name AZURE BLOB INPUT NAME account key AZURE PRIMARY KEY Since we aren t editing the file read access is sufficient permission BlobSasPermissions read True expiry datetime utcnow timedelta hours SAS will expire in hoursinput url https AZURE ACC NAME blob core windows net AZURE CONTAINER AZURE BLOB INPUT NAME input sas blobThe generate blob sas function creates a Shared Access Signature which when formatted correctly into a URL creates a direct path to the media The three tools briefly mentioned earlier each are connected to create our workflow Dolby io Audio Diagnose A lightweight audio analysis tool that can return information relating to the audio such as its quality or audio defects Azure Cognitive Services Speech to Text A transcription tool that converts spoken audio to text Dolby io Enhance An audio enhancement tool that can help reduce noise and level speakers We start with diagnose which using the pre signed URL can access the file stored on the cloud to evaluate its audio quality It then returns via a callback a score out of with being low quality audio and being high quality audio If the score is at or above our threshold of we will transcribe it otherwise we will clean up the audio with enhance for manual review We settled on this threshold of after some testing as audio that scores above this threshold usually performs the best with Speech to Text Now that we understand the workflow and the pieces that connect to make it work let s test out the sample project Part Testing out the workflow With the workflow set up we can now test it out one of two ways We recommend testing locally before deploying it with Serverless as when developing locally you can make simple and fast changes to your code without having to worry about uploading it to the Azure servers Additionally when developing locally you don t run the risk of exposing API keys When deploying to the Azure cloud make sure you review the Authentication Guide to be sure you are following best practices and protecting your keys Locally To test locally we need to create a HTTP tunnel to port forward to localhost There are many options available for this example I used a free ngrok account In my case my local function was deployed on so I initialized ngrok to port forward on LocalHost Once you have launched your HTTP tunnel its time to update the params json file Include the correct API keys along with the appropriate names for your Azure account and container want to search for the file in Additionally you can adjust the score threshold and the output suffix With our HTTP tunnel launched we also need to update the tunneling URL with the appropriate Forwarding address In Visual Studio Code press F to launch the project and navigate to https localhost YOUR SERVER api MediaProcessingWorkflow input file YOUR INPUT FILE You can monitor your console in Visual Studio for output logs and your container on the Azure portal for the transcribed results Serverless Deploying the functions to an Azure Server is a great choice once the code has been fully debugged and tested just remember to protect your API keys with the appropriate authentication strategy Adjust the params json file this time set the tunneling parameter to https YOUR FUNCTION APP NAME azurewebsites net api YOUR FUNCTION NAME Next deploy the function app through the Azure extension Once successfully deployed you can test your function by navigating to https YOUR FUNCTION APP NAME azurewebsites net api YOUR FUNCTION NAME input file YOUR INPUT FILE Part Building your own workflow With the function successfully deployed our transcription workflow tool is complete A depiction of the serverless media workflow available at To review our media follows a two step workflow From the Azure container the file is diagnosed to return an audio quality score If the audio scores below a quality threshold of we opt not to transcribe the audio and instead enhance the audio If the media scores at or above that threshold we then pass the file onto the next stage for transcription Once the transcription is complete the file is then deposited on the container This example serves as an introduction to building a Serverless media workflow on Azure By using the tips described above it is possible to add any number of extra steps into the workflow including APIs that check for profanity or APIs that translate foreign languages the possibilities are endless If you are interested in learning more about the workflow outlined above to check out the team s presentation at the Azure Serverless Conference where we go into more detail otherwise feel free to explore some of our other great projects over at the Dolby io Samples Github |
2022-04-13 15:14:48 |
海外TECH |
DEV Community |
Store your notes in github repository |
https://dev.to/vivekweb2013/store-your-notes-in-github-repository-437o
|
Store your notes in github repositoryWhat if you could store notes as markdown inside your github repository Would you a webapp that allows to store your notes in your private github repository If yes could you please share your feedback on my recent open source project |
2022-04-13 15:13:54 |
Apple |
AppleInsider - Frontpage News |
Apple's privacy features will cost Facebook $12.8 billion in 2022 |
https://appleinsider.com/articles/22/04/13/apples-privacy-features-will-cost-facebook-128-billion-in-2022?utm_medium=rss
|
Apple x s privacy features will cost Facebook billion in Almost months after Apple launched App Tracking Transparency a new analysis predicts its second year will still see major disruption to advertiser with Facebook YouTube and more collectively losing around billion Not everyone will have gotten their App Tracking Transparency options right first time Apple released its App Tracking Transparency ATT feature in iOS on April and it immediately had an impact on companies relying on advertising revenue By July it was estimated to be causing a to revenue drop for advertisers Read more |
2022-04-13 15:00:52 |
海外TECH |
Engadget |
Amazon accused of ramping up anti-union efforts ahead of another warehouse election |
https://www.engadget.com/amazon-anti-union-report-election-staten-island-153310661.html?src=rss
|
Amazon accused of ramping up anti union efforts ahead of another warehouse electionAmazon is said to have intensified its anti union efforts ahead of a union election at a warehouse later this month The Amazon Labor Union told Motherboard the company is mandating daily anti union meetings at LDJ a facility in Staten Island New York It s also said to have distributed anti union literature and disciplined a leader of the drive for organizing on the warehouse floor What s more ALU says Amazon has hired anti union consultants to pose as employees Workers at the warehouse which reportedly has around employees are scheduled to begin a union election on April th Amazon s anti union efforts ramped up in recent days according to the report The ALU recently won an election at a nearby facility JFK which became the first Amazon warehouse in the US to formally unionize Amazon plans to appeal the union s victory Amazon and the National Labor Relations Board in December reached a deal in December under which the company agreed to inform past and current warehouse workers in the US of their right to organize The terms of the agreement afforded workers more leeway to organize in break rooms which is said to have been a key factor in ALU s success at JFK However Amazon reportedly isn t sticking to those terms at LDJ The ALU said the company removed pro union literature from the break room and took down a pro union banner after the JFK election result became clear A lawyer representing ALU workers has filed unfair labor practice charges against Amazon for removing the banner and allegedly retaliating against a worker to stifle unionization efforts Engadget has contacted Amazon for comment Amazon has long been accused of cracking down on workers attempts to organize Last year alone it spent million on anti union consultants The company s also said to be working on a chat app for workers in which terms like quot union quot and quot pay raise quot are on a blocklist The NLRB said the company illegally interfered in a union election in Bessemer Alabama last year and called for a rerun However the Retail Wholesale and Department Store Union claimed Amazon interfered in the second election as well The result of that vote hinges on a court hearing over challenged ballots |
2022-04-13 15:33:10 |
海外TECH |
Engadget |
Houston Astros' stadium will be the first in MLB to use Amazon's 'Just Walk Out' tech |
https://www.engadget.com/houston-astros-amazon-just-walk-out-shopping-152141941.html?src=rss
|
Houston Astros x stadium will be the first in MLB to use Amazon x s x Just Walk Out x techAmazon has brought its checkout free quot Just Walk Out quot technology to airports grocery stores and other shops but now it s coming to a particularly useful place for sports fans the ballpark The Houston Astros have teamed up with Amazon to install Just Walk Out systems at two concession stores in Minute Maid Park Visit th Hole or Market and you can buy snacks or souvenirs between innings by inserting your credit card at the entry gate grabbing things off the shelf and leaving when you re done There will be staff to greet you and offer help as necessary and you ll still have to show ID if you re buying alcohol However you otherwise won t have to talk to a cashier or use a self checkout system As you might guess that could be extremely helpful given the crowds and lineups that frequently slow you down in stadiums The Astros stadium is the first in Major League Baseball to adopt Just Walk Out and they re using the same slightly modified system Amazon is offering to other retailers The tech uses computer vision and other forms of AI to track shoppers as they enter and take or put back items Amazon s own stores just rely on the company s Go mobile app instead of credit cards Amazon didn t mention whether or not other MLB teams would embrace the zero checkout offering but it won t be surprising if they and other sports leagues do Stadium operators depend heavily on both merch sales and a swift traffic flow to turn a profit ーthe cost of Just Walk Out could easily be worthwhile if increases the chances you ll buy an expensive hot dog or replica jersey |
2022-04-13 15:21:41 |
Linux |
OMG! Ubuntu! |
Amberol is Fab No-Frills Audio Player for GNOME Desktops |
https://www.omgubuntu.co.uk/2022/04/amberol-simple-gtk-music-player
|
Amberol is Fab No Frills Audio Player for GNOME DesktopsLooking for a simple focused no frills music player that fits in beautifully on the modern GNOME desktop Check out Amberol “Amberol aspires to be as small unintrusive and simple as possible It does not manage your music collection it does not let you manage playlists smart or otherwise it does not let you edit the metadata for your songs it does not show you lyrics for your songs or the Wikipedia page for your bands states Amberol s Gitlab page That s to say all this audio player does is play music and …Well that s pretty much it You do get a This post Amberol is Fab No Frills Audio Player for GNOME Desktops is from OMG Ubuntu Do not reproduce elsewhere without permission |
2022-04-13 15:32:17 |
海外科学 |
NYT > Science |
McKinsey Opened a Door in Its Firewall Between Pharma Clients and Regulators |
https://www.nytimes.com/2022/04/13/business/mckinsey-purdue-fda-records.html
|
McKinsey Opened a Door in Its Firewall Between Pharma Clients and RegulatorsThe firm let consultants advise both drugmakers and their government overseers internal records show “Who we know and what we know was part of their pitch |
2022-04-13 15:40:16 |
金融 |
RSS FILE - 日本証券業協会 |
株主コミュニティの統計情報・取扱状況 |
https://www.jsda.or.jp/shiryoshitsu/toukei/kabucommunity/index.html
|
株主コミュニティ |
2022-04-13 15:30:00 |
金融 |
RSS FILE - 日本証券業協会 |
動画で見る日証協の活動 |
https://www.jsda.or.jp/about/gaiyou/movie/index.html
|
日証協 |
2022-04-13 15:30:00 |
金融 |
金融庁ホームページ |
令和4年3月に開催された業界団体との意見交換会において金融庁が提起した主な論点を公表しました。 |
https://www.fsa.go.jp/common/ronten/index.html
|
意見交換会 |
2022-04-13 17:00:00 |
金融 |
金融庁ホームページ |
店頭デリバティブ取引情報について公表しました。(令和3年(2021年)3月末) |
https://www.fsa.go.jp/news/r3/shouken/20220413/20220413.html
|
取引情報 |
2022-04-13 17:00:00 |
金融 |
金融庁ホームページ |
「共同データプラットフォームの構築に向けた基礎調査」報告書を公表しました。 |
https://www.fsa.go.jp/common/about/research/20220413/20220413.html
|
調査 |
2022-04-13 17:00:00 |
金融 |
金融庁ホームページ |
「ソーシャルプロジェクトのインパクト指標等の検討に関する関係府省庁会議」(第2回)議事次第を公表しました。 |
https://www.fsa.go.jp/singi/social_impact/siryou/20220412.html
|
関係 |
2022-04-13 17:00:00 |
ニュース |
@日本経済新聞 電子版 |
ロシア外貨準備、1月に人民元がドル逆転 侵攻に備えか
https://t.co/jlnRZX95Jd |
https://twitter.com/nikkei/statuses/1514263062681104386
|
外貨準備 |
2022-04-13 15:23:42 |
ニュース |
BBC News - Home |
Wayne Couzens requests jury trial for indecent exposure charges |
https://www.bbc.co.uk/news/uk-england-london-61082238?at_medium=RSS&at_campaign=KARANGA
|
chargesthe |
2022-04-13 15:50:53 |
ニュース |
BBC News - Home |
Mutations across species reveal clues to ageing |
https://www.bbc.co.uk/news/health-61045950?at_medium=RSS&at_campaign=KARANGA
|
humans |
2022-04-13 15:01:04 |
ニュース |
BBC News - Home |
Matt Hancock: Not enough evidence to prosecute CCTV leak suspects |
https://www.bbc.co.uk/news/uk-politics-61097109?at_medium=RSS&at_campaign=KARANGA
|
minister |
2022-04-13 15:27:54 |
ニュース |
BBC News - Home |
War in Ukraine: The moment a puppy is rescued from the rubble |
https://www.bbc.co.uk/news/world-europe-61099211?at_medium=RSS&at_campaign=KARANGA
|
hands |
2022-04-13 15:42:19 |
北海道 |
北海道新聞 |
熊本、シンボル復旧も仮設95人 地震から14日で6年 |
https://www.hokkaido-np.co.jp/article/669247/
|
犠牲 |
2022-04-14 00:29:00 |
北海道 |
北海道新聞 |
カナダ中銀、大幅利上げ 22年ぶり、インフレ対応で |
https://www.hokkaido-np.co.jp/article/669246/
|
政策金利 |
2022-04-14 00:28:00 |
北海道 |
北海道新聞 |
ロシア債務不履行20日に再協議 ドル建て国債、業界団体 |
https://www.hokkaido-np.co.jp/article/669240/
|
債務不履行 |
2022-04-14 00:07:00 |
コメント
コメントを投稿