Tutorial: Criar um aplicativo iOS que inicia a Leitura Avançada com o conteúdo de uma foto (Swift)
A Leitura Avançada é uma ferramenta projetada de forma inclusiva que implementa técnicas comprovadas para melhorar a compreensão da leitura.
A API de Leitura da Visão de IA do Azure detecta o conteúdo de texto em uma imagem usando os modelos de reconhecimento mais recentes da Microsoft e converte o texto identificado em um fluxo de caracteres legível por computador.
Neste tutorial, você criará um aplicativo iOS do zero e integrará a API de Leitura e a Leitura Avançada usando o SDK de Leitura Avançada. Um exemplo de funcionamento completo deste tutorial está disponível no GitHub.
Pré-requisitos
- Uma assinatura do Azure. É possível criar uma gratuitamente.
- MacOS e Xcode.
- Um recurso de Leitura Avançada configurado para autenticação do Microsoft Entra. Siga estas instruções para a configuração.
- Uma assinatura do serviço de Visão de IA do Azure. Crie um recurso da Visão de IA do Azure no portal do Azure.
Criar um projeto do Xcode
Criar um projeto no Xcode.
Escolha Aplicativo de Modo de Exibição Único.
Obter o SDK CocoaPod
A maneira mais fácil de usar o SDK de Leitura Avançada é por meio do CocoaPods. Para instalar por meio do Cocoapods:
Siga o guia para instalar o Cocoapods.
Crie um Podfile executando
pod init
no diretório raiz do projeto Xcode.Adicione o CocoaPod ao seu Podfile adicionando
pod 'immersive-reader-sdk', :path => 'https://github.com/microsoft/immersive-reader-sdk/tree/master/iOS/immersive-reader-sdk'
. Seu Podfile deve ser semelhante ao seguinte, com o nome do destino substituindo picture-to-immersive-reader-swift:platform :ios, '9.0' target 'picture-to-immersive-reader-swift' do use_frameworks! # Pods for picture-to-immersive-reader-swift pod 'immersive-reader-sdk', :git => 'https://github.com/microsoft/immersive-reader-sdk.git' end
No terminal, no diretório do seu projeto Xcode, execute o comando
pod install
para instalar o pod do SDK de Leitura Avançada.Adicione
import immersive_reader_sdk
a todos os arquivos que precisam referenciar o SDK.Abra o projeto abrindo o arquivo
.xcworkspace
, e não o arquivo.xcodeproj
.
Adquirir um token de autenticação do Microsoft Entra
Você precisa de alguns valores da etapa de configuração de autenticação do Microsoft Entra na seção de pré-requisitos. Veja novamente o arquivo de texto que você salvou dessa sessão.
TenantId => Azure subscription TenantId
ClientId => Microsoft Entra ApplicationId
ClientSecret => Microsoft Entra Application Service Principal password
Subdomain => Immersive Reader resource subdomain (resource 'Name' if the resource was created in the Azure portal, or 'CustomSubDomain' option if the resource was created with Azure CLI PowerShell. Check the Azure portal for the subdomain on the Endpoint in the resource Overview page, for example, 'https://[SUBDOMAIN].cognitiveservices.azure.com/')
Na pasta do projeto principal, que contém o arquivo ViewController.swift, crie um arquivo de classe Swift chamado Constants.swift
. Substitua a classe pelo código a seguir, adicionando em seus valores quando aplicável. Mantenha esse arquivo como um arquivo local que só existe em seu computador e não faça commit desse arquivo no controle do código-fonte, pois ele contém segredos que não devem ser tornados públicos. Não recomendamos que você não mantenha segredos em seu aplicativo. Em vez disso, use um serviço de back-end para obter o token, em que os segredos podem ser mantidos fora do aplicativo e do dispositivo. O ponto de extremidade da API de back-end deve ser protegido por alguma forma de autenticação (por exemplo, OAuth) para impedir que usuários não autorizados obtenham tokens para usar em seu serviço de Leitura Avançada e cobrança; esse trabalho está além do escopo deste tutorial.
Configurar o aplicativo para ser executado sem um storyboard
Abra AppDelegate.swift e substitua o arquivo pelo código a seguir.
import UIKit
@UIApplicationMain
class AppDelegate: UIResponder, UIApplicationDelegate {
var window: UIWindow?
var navigationController: UINavigationController?
func application(_ application: UIApplication, didFinishLaunchingWithOptions launchOptions: [UIApplication.LaunchOptionsKey: Any]?) -> Bool {
// Override point for customization after application launch.
window = UIWindow(frame: UIScreen.main.bounds)
// Allow the app run without a storyboard
if let window = window {
let mainViewController = PictureLaunchViewController()
navigationController = UINavigationController(rootViewController: mainViewController)
window.rootViewController = navigationController
window.makeKeyAndVisible()
}
return true
}
func applicationWillResignActive(_ application: UIApplication) {
// Sent when the application is about to move from active to inactive state. This can occur for certain types of temporary interruptions (such as an incoming phone call or SMS message) or when the user quits the application and it begins the transition to the background state.
// Use this method to pause ongoing tasks, disable timers, and invalidate graphics rendering callbacks. Games should use this method to pause the game.
}
func applicationDidEnterBackground(_ application: UIApplication) {
// Use this method to release shared resources, save user data, invalidate timers, and store enough application state information to restore your application to its current state in case it is terminated later.
// If your application supports background execution, this method is called instead of applicationWillTerminate: when the user quits.
}
func applicationWillEnterForeground(_ application: UIApplication) {
// Called as part of the transition from the background to the active state; here you can undo many of the changes made on entering the background.
}
func applicationDidBecomeActive(_ application: UIApplication) {
// Restart any tasks that were paused (or not yet started) while the application was inactive. If the application was previously in the background, optionally refresh the user interface.
}
func applicationWillTerminate(_ application: UIApplication) {
// Called when the application is about to terminate. Save data if appropriate. See also applicationDidEnterBackground:.
}
}
Adicionar funcionalidade para tirar e carregar fotos
Renomeie ViewController.swift para PictureLaunchViewController.swift e substitua o arquivo pelo código a seguir.
import UIKit
import immersive_reader_sdk
class PictureLaunchViewController: UIViewController, UINavigationControllerDelegate, UIImagePickerControllerDelegate {
private var photoButton: UIButton!
private var cameraButton: UIButton!
private var titleText: UILabel!
private var bodyText: UILabel!
private var sampleContent: Content!
private var sampleChunk: Chunk!
private var sampleOptions: Options!
private var imagePicker: UIImagePickerController!
private var spinner: UIActivityIndicatorView!
private var activityIndicatorBackground: UIView!
private var textURL = "vision/v2.0/read/core/asyncBatchAnalyze";
override func viewDidLoad() {
super.viewDidLoad()
view.backgroundColor = .white
titleText = UILabel()
titleText.text = "Picture to Immersive Reader with OCR"
titleText.font = UIFont.boldSystemFont(ofSize: 32)
titleText.textAlignment = .center
titleText.lineBreakMode = .byWordWrapping
titleText.numberOfLines = 0
view.addSubview(titleText)
bodyText = UILabel()
bodyText.text = "Capture or upload a photo of handprinted text on a piece of paper, handwriting, typed text, text on a computer screen, writing on a white board and many more, and watch it be presented to you in the Immersive Reader!"
bodyText.font = UIFont.systemFont(ofSize: 18)
bodyText.lineBreakMode = .byWordWrapping
bodyText.numberOfLines = 0
let screenSize = self.view.frame.height
if screenSize <= 667 {
// Font size for smaller iPhones.
bodyText.font = bodyText.font.withSize(16)
} else if screenSize <= 812.0 {
// Font size for medium iPhones.
bodyText.font = bodyText.font.withSize(18)
} else if screenSize <= 896 {
// Font size for larger iPhones.
bodyText.font = bodyText.font.withSize(20)
} else {
// Font size for iPads.
bodyText.font = bodyText.font.withSize(26)
}
view.addSubview(bodyText)
photoButton = UIButton()
photoButton.backgroundColor = .darkGray
photoButton.contentEdgeInsets = UIEdgeInsets(top: 10, left: 5, bottom: 10, right: 5)
photoButton.layer.cornerRadius = 5
photoButton.setTitleColor(.white, for: .normal)
photoButton.setTitle("Choose Photo from Library", for: .normal)
photoButton.titleLabel?.font = UIFont.systemFont(ofSize: 18, weight: .bold)
photoButton.addTarget(self, action: #selector(selectPhotoButton(sender:)), for: .touchUpInside)
view.addSubview(photoButton)
cameraButton = UIButton()
cameraButton.backgroundColor = .darkGray
cameraButton.contentEdgeInsets = UIEdgeInsets(top: 10, left: 5, bottom: 10, right: 5)
cameraButton.layer.cornerRadius = 5
cameraButton.setTitleColor(.white, for: .normal)
cameraButton.setTitle("Take Photo", for: .normal)
cameraButton.titleLabel?.font = UIFont.systemFont(ofSize: 18, weight: .bold)
cameraButton.addTarget(self, action: #selector(takePhotoButton(sender:)), for: .touchUpInside)
view.addSubview(cameraButton)
activityIndicatorBackground = UIView()
activityIndicatorBackground.backgroundColor = UIColor.black
activityIndicatorBackground.alpha = 0
view.addSubview(activityIndicatorBackground)
view.bringSubviewToFront(_: activityIndicatorBackground)
spinner = UIActivityIndicatorView(style: .whiteLarge)
view.addSubview(spinner)
let layoutGuide = view.safeAreaLayoutGuide
titleText.translatesAutoresizingMaskIntoConstraints = false
titleText.topAnchor.constraint(equalTo: layoutGuide.topAnchor, constant: 25).isActive = true
titleText.leadingAnchor.constraint(equalTo: layoutGuide.leadingAnchor, constant: 20).isActive = true
titleText.trailingAnchor.constraint(equalTo: layoutGuide.trailingAnchor, constant: -20).isActive = true
bodyText.translatesAutoresizingMaskIntoConstraints = false
bodyText.topAnchor.constraint(equalTo: titleText.bottomAnchor, constant: 35).isActive = true
bodyText.leadingAnchor.constraint(equalTo: layoutGuide.leadingAnchor, constant: 20).isActive = true
bodyText.trailingAnchor.constraint(equalTo: layoutGuide.trailingAnchor, constant: -20).isActive = true
cameraButton.translatesAutoresizingMaskIntoConstraints = false
if screenSize > 896 {
// Constraints for iPads.
cameraButton.heightAnchor.constraint(equalToConstant: 150).isActive = true
cameraButton.leadingAnchor.constraint(equalTo: layoutGuide.leadingAnchor, constant: 60).isActive = true
cameraButton.trailingAnchor.constraint(equalTo: layoutGuide.trailingAnchor, constant: -60).isActive = true
cameraButton.topAnchor.constraint(equalTo: bodyText.bottomAnchor, constant: 150).isActive = true
} else {
// Constraints for iPhones.
cameraButton.heightAnchor.constraint(equalToConstant: 100).isActive = true
cameraButton.leadingAnchor.constraint(equalTo: layoutGuide.leadingAnchor, constant: 30).isActive = true
cameraButton.trailingAnchor.constraint(equalTo: layoutGuide.trailingAnchor, constant: -30).isActive = true
cameraButton.topAnchor.constraint(equalTo: bodyText.bottomAnchor, constant: 100).isActive = true
}
cameraButton.bottomAnchor.constraint(equalTo: photoButton.topAnchor, constant: -40).isActive = true
photoButton.translatesAutoresizingMaskIntoConstraints = false
if screenSize > 896 {
// Constraints for iPads.
photoButton.heightAnchor.constraint(equalToConstant: 150).isActive = true
photoButton.leadingAnchor.constraint(equalTo: layoutGuide.leadingAnchor, constant: 60).isActive = true
photoButton.trailingAnchor.constraint(equalTo: layoutGuide.trailingAnchor, constant: -60).isActive = true
} else {
// Constraints for iPhones.
photoButton.heightAnchor.constraint(equalToConstant: 100).isActive = true
photoButton.leadingAnchor.constraint(equalTo: layoutGuide.leadingAnchor, constant: 30).isActive = true
photoButton.trailingAnchor.constraint(equalTo: layoutGuide.trailingAnchor, constant: -30).isActive = true
}
spinner.translatesAutoresizingMaskIntoConstraints = false
spinner.centerXAnchor.constraint(equalTo: view.centerXAnchor).isActive = true
spinner.centerYAnchor.constraint(equalTo: view.centerYAnchor).isActive = true
activityIndicatorBackground.translatesAutoresizingMaskIntoConstraints = false
activityIndicatorBackground.topAnchor.constraint(equalTo: layoutGuide.topAnchor).isActive = true
activityIndicatorBackground.bottomAnchor.constraint(equalTo: layoutGuide.bottomAnchor).isActive = true
activityIndicatorBackground.leadingAnchor.constraint(equalTo: layoutGuide.leadingAnchor).isActive = true
activityIndicatorBackground.trailingAnchor.constraint(equalTo: layoutGuide.trailingAnchor).isActive = true
// Create content and options.
sampleChunk = Chunk(content: bodyText.text!, lang: nil, mimeType: nil)
sampleContent = Content(title: titleText.text!, chunks: [sampleChunk])
sampleOptions = Options(uiLang: nil, timeout: nil, uiZIndex: nil)
}
@IBAction func selectPhotoButton(sender: AnyObject) {
// Launch the photo picker.
imagePicker = UIImagePickerController()
imagePicker.delegate = self
self.imagePicker.sourceType = .photoLibrary
self.imagePicker.allowsEditing = true
self.present(self.imagePicker, animated: true, completion: nil)
self.photoButton.isEnabled = true
}
@IBAction func takePhotoButton(sender: AnyObject) {
if !UIImagePickerController.isSourceTypeAvailable(.camera) {
// If there is no camera on the device, disable the button
self.cameraButton.backgroundColor = .gray
self.cameraButton.isEnabled = true
} else {
// Launch the camera.
imagePicker = UIImagePickerController()
imagePicker.delegate = self
self.imagePicker.sourceType = .camera
self.present(self.imagePicker, animated: true, completion: nil)
self.cameraButton.isEnabled = true
}
}
func imagePickerController(_ picker: UIImagePickerController, didFinishPickingMediaWithInfo info: [UIImagePickerController.InfoKey : Any]) {
imagePicker.dismiss(animated: true, completion: nil)
photoButton.isEnabled = false
cameraButton.isEnabled = false
self.spinner.startAnimating()
activityIndicatorBackground.alpha = 0.6
// Retrieve the image.
let image = (info[.originalImage] as? UIImage)!
// Retrieve the byte array from image.
let imageByteArray = image.jpegData(compressionQuality: 1.0)
// Call the getTextFromImage function passing in the image the user takes or chooses.
getTextFromImage(subscriptionKey: Constants.computerVisionSubscriptionKey, getTextUrl: Constants.computerVisionEndPoint + textURL, pngImage: imageByteArray!, onSuccess: { cognitiveText in
print("cognitive text is: \(cognitiveText)")
DispatchQueue.main.async {
self.photoButton.isEnabled = true
self.cameraButton.isEnabled = true
}
// Create content and options with the text from the image.
let sampleImageChunk = Chunk(content: cognitiveText, lang: nil, mimeType: nil)
let sampleImageContent = Content(title: "Text from image", chunks: [sampleImageChunk])
let sampleImageOptions = Options(uiLang: nil, timeout: nil, uiZIndex: nil)
// Callback to get token for Immersive Reader.
self.getToken(onSuccess: {cognitiveToken in
DispatchQueue.main.async {
launchImmersiveReader(navController: self.navigationController!, token: cognitiveToken, subdomain: Constants.subdomain, content: sampleImageContent, options: sampleImageOptions, onSuccess: {
self.spinner.stopAnimating()
self.activityIndicatorBackground.alpha = 0
self.photoButton.isEnabled = true
self.cameraButton.isEnabled = true
}, onFailure: { error in
print("An error occurred launching the Immersive Reader: \(error)")
self.spinner.stopAnimating()
self.activityIndicatorBackground.alpha = 0
self.photoButton.isEnabled = true
self.cameraButton.isEnabled = true
})
}
}, onFailure: { error in
DispatchQueue.main.async {
self.photoButton.isEnabled = true
self.cameraButton.isEnabled = true
}
print("An error occurred retrieving the token: \(error)")
})
}, onFailure: { error in
DispatchQueue.main.async {
self.photoButton.isEnabled = true
self.cameraButton.isEnabled = true
}
})
}
/// Retrieves the token for the Immersive Reader using Microsoft Entra authentication
///
/// - Parameters:
/// -onSuccess: A closure that gets called when the token is successfully received using Microsoft Entra authentication.
/// -theToken: The token for the Immersive Reader received using Microsoft Entra authentication.
/// -onFailure: A closure that gets called when the token fails to be obtained from the Microsoft Entra authentication.
/// -theError: The error that occurred when the token fails to be obtained from the Microsoft Entra authentication.
func getToken(onSuccess: @escaping (_ theToken: String) -> Void, onFailure: @escaping ( _ theError: String) -> Void) {
let tokenForm = "grant_type=client_credentials&resource=https://cognitiveservices.azure.com/&client_id=" + Constants.clientId + "&client_secret=" + Constants.clientSecret
let tokenUrl = "https://login.windows.net/" + Constants.tenantId + "/oauth2/token"
var responseTokenString: String = "0"
let url = URL(string: tokenUrl)!
var request = URLRequest(url: url)
request.httpBody = tokenForm.data(using: .utf8)
request.httpMethod = "POST"
let task = URLSession.shared.dataTask(with: request) { data, response, error in
guard let data = data,
let response = response as? HTTPURLResponse,
// Check for networking errors.
error == nil else {
print("error", error ?? "Unknown error")
onFailure("Error")
return
}
// Check for http errors.
guard (200 ... 299) ~= response.statusCode else {
print("statusCode should be 2xx, but is \(response.statusCode)")
print("response = \(response)")
onFailure(String(response.statusCode))
return
}
let responseString = String(data: data, encoding: .utf8)
print("responseString = \(String(describing: responseString!))")
let jsonResponse = try? JSONSerialization.jsonObject(with: data, options: [])
guard let jsonDictonary = jsonResponse as? [String: Any] else {
onFailure("Error parsing JSON response.")
return
}
guard let responseToken = jsonDictonary["access_token"] as? String else {
onFailure("Error retrieving token from JSON response.")
return
}
responseTokenString = responseToken
onSuccess(responseTokenString)
}
task.resume()
}
/// Returns the text string after it has been extracted from an Image input.
///
/// - Parameters:
/// -subscriptionKey: The Azure subscription key.
/// -pngImage: Image data in PNG format.
/// - Returns: a string of text representing the
func getTextFromImage(subscriptionKey: String, getTextUrl: String, pngImage: Data, onSuccess: @escaping (_ theToken: String) -> Void, onFailure: @escaping ( _ theError: String) -> Void) {
let url = URL(string: getTextUrl)!
var request = URLRequest(url: url)
request.setValue(subscriptionKey, forHTTPHeaderField: "Ocp-Apim-Subscription-Key")
request.setValue("application/octet-stream", forHTTPHeaderField: "Content-Type")
// Two REST API calls are required to extract text. The first call is to submit the image for processing, and the next call is to retrieve the text found in the image.
// Set the body to the image in byte array format.
request.httpBody = pngImage
request.httpMethod = "POST"
let task = URLSession.shared.dataTask(with: request) { data, response, error in
guard let data = data,
let response = response as? HTTPURLResponse,
// Check for networking errors.
error == nil else {
print("error", error ?? "Unknown error")
onFailure("Error")
return
}
// Check for http errors.
guard (200 ... 299) ~= response.statusCode else {
print("statusCode should be 2xx, but is \(response.statusCode)")
print("response = \(response)")
onFailure(String(response.statusCode))
return
}
let responseString = String(data: data, encoding: .utf8)
print("responseString = \(String(describing: responseString!))")
// Send the second call to the API. The first API call returns operationLocation which stores the URI for the second REST API call.
let operationLocation = response.allHeaderFields["Operation-Location"] as? String
if (operationLocation == nil) {
print("Error retrieving operation location")
return
}
// Wait 10 seconds for text recognition to be available as suggested by the Text API documentation.
print("Text submitted. Waiting 10 seconds to retrieve the recognized text.")
sleep(10)
// HTTP GET request with the operationLocation url to retrieve the text.
let getTextUrl = URL(string: operationLocation!)!
var getTextRequest = URLRequest(url: getTextUrl)
getTextRequest.setValue(subscriptionKey, forHTTPHeaderField: "Ocp-Apim-Subscription-Key")
getTextRequest.httpMethod = "GET"
// Send the GET request to retrieve the text.
let taskGetText = URLSession.shared.dataTask(with: getTextRequest) { data, response, error in
guard let data = data,
let response = response as? HTTPURLResponse,
// Check for networking errors.
error == nil else {
print("error", error ?? "Unknown error")
onFailure("Error")
return
}
// Check for http errors.
guard (200 ... 299) ~= response.statusCode else {
print("statusCode should be 2xx, but is \(response.statusCode)")
print("response = \(response)")
onFailure(String(response.statusCode))
return
}
// Decode the JSON data into an object.
let customDecoding = try! JSONDecoder().decode(TextApiResponse.self, from: data)
// Loop through the lines to get all lines of text and concatenate them together.
var textFromImage = ""
for textLine in customDecoding.recognitionResults[0].lines {
textFromImage = textFromImage + textLine.text + " "
}
onSuccess(textFromImage)
}
taskGetText.resume()
}
task.resume()
}
// Structs used for decoding the Text API JSON response.
struct TextApiResponse: Codable {
let status: String
let recognitionResults: [RecognitionResult]
}
struct RecognitionResult: Codable {
let page: Int
let clockwiseOrientation: Double
let width, height: Int
let unit: String
let lines: [Line]
}
struct Line: Codable {
let boundingBox: [Int]
let text: String
let words: [Word]
}
struct Word: Codable {
let boundingBox: [Int]
let text: String
let confidence: String?
}
}
Compilar e executar o aplicativo
Defina o esquema de arquivo no Xcode selecionando um simulador ou um destino do dispositivo.
No Xcode, pressione Ctrl+R ou selecione o botão reproduzir para executar o projeto. O aplicativo deve ser iniciado no simulador ou dispositivo especificado.
Em seu aplicativo, você deverá ver:
Tire ou carregue uma foto de texto pressionando o botão Tirar Foto ou o botãoEscolher Foto da Biblioteca. Em seguida, a Leitura Avançada é inicializada e exibe o texto da foto.