microblink/blinkid-ios v8000.0.0

on GitHub

What's new

• SDK Modularization & Enhanced Flexibility - We’ve transformed our core architecture. Features previously known as "fallback modes" are now independent SDK modules.

  • Tailored Implementation: Users can now toggle specific modules such as Capture, Barcode, MRZ, or VIZ scanners to build a workflow that fits their exact needs.
  • Total Control: This modular approach offers unprecedented flexibility, allowing you to use BlinkID as a comprehensive recognition tool or a specialized scanner, all while leveraging our latest feature updates.

• Expanded Barcode Support - When operating in Barcode Mode, BlinkID is no longer limited to PDF417.

  • Universal Scanning: We’ve added support for QR codes, Code 128, Code 39, ITF, EAN, UPC, and DataMatrix.
  • Versatility: This expansion allows the SDK to be used across a wider variety of document types beyond standard IDs.

• Streamlined Extraction & Anonymization - We are putting more power in the hands of our integrators by removing restrictive custom rules.

  • Document Rules: Custom document rules have been removed. Customers can now implement their own logic based on result completeness to determine how a document is processed.
  • Anonymization: We’ve transitioned away from fixed anonymization rules. You can now programmatically decide which fields to mask and how to handle them based on the specific data returned.

• Intelligent Timeout Mechanism - To balance performance with accuracy, we’ve introduced a new timeout feature

  • Non-Blocking Scans: Ensures a smooth user experience by preventing "stuck" scan states.
  • Step-by-Step Decisions: After each extraction step, you can decide whether the current data is sufficient or if the SDK should continue processing to gather more information.
  • Inactivity Timer: A new inactivity timer ends the scanning session after a configurable period (default 10 seconds) during which no changes in scanning events are observed.
  • Per-Step Timer: The existing per-step timer continues to apply independently to each scanning step (first side, second side, and barcode), and works alongside the new inactivity timer.

• Improved Barcode Reading Logic - Refined the internal logic for barcode detection and decoding, resulting in more reliable extraction across supported symbologies and document types.

• New Initializers - BlinkIDUXView now exposes a simplified initializer that takes only the analyzer and a scan-completion closure. UX configuration and frame-processing callbacks are provided via SwiftUI environment values (uxSettings, frameProcessResultCallback), keeping the call site clean while remaining fully customizable.

  public struct BlinkIDUXView: View {
      @Environment(\.uxSettings) private var uxSettings
      @Environment(\.frameProcessResultCallback) private var frameProcessResultCallback
  
      private let analyzer: any CameraFrameAnalyzer<CameraFrame, UIEvent>
      private let onScanCompleted: (BlinkIDResultState) -> Void
  
      public init(
          analyzer: any CameraFrameAnalyzer<CameraFrame, UIEvent>,
          onScanCompleted: @escaping (BlinkIDResultState) -> Void
      ) {
          self.analyzer = analyzer
          self.onScanCompleted = onScanCompleted
          self.externalViewModel = nil
      }
  }

• New UI for Extraction Modes - Added dedicated UI flows tailored to each extraction mode:

  • barcodeOnly: streamlined UI focused exclusively on barcode capture.
  • documentWithBarcode: UI optimized for documents where both the visual document and the barcode are captured.
  • fullDocument: UI for complete document extraction across all supported sides.

• Document-Specific Improvements

  • EU Residence Permits: For documents with redundant fields on both the front and back, BlinkID now intelligently merges these into the top-level results for better data consistency.
  • Egypt Driver’s License: Enhanced data extraction to now include Date of Birth (DOB) as an additional field.
  • Zimbabwe ID: Improved data extraction to support and return both alphabetic and numeric characters within the document number field.

Bug fixes

• Date Conversion Accuracy: Resolved an issue where Islamic-to-Gregorian date conversions could occasionally differ by +/- 1 day. These conversions are now precise and consistent.

API changes

BlinkIDUXView API Redesign

Overview

The scanning result observation API has been redesigned from a Combine-based publisher pattern to a SwiftUI-native closure and modifier approach. This removes the need for managing AnyCancellable subscriptions manually and makes integration significantly more concise.

What Changed

BlinkIDUXModel + Combine publisher -> BlinkIDUXView with result closure

The previous API required instantiating a BlinkIDUXModel, subscribing to its $result publisher via .sink, and storing the cancellable. The new API passes the result handler directly as a trailing closure on BlinkIDUXView, which receives the scanningResultState inline.

Frame processing result handling via .onFrameProcessResult

Advance-to-next-step logic is now expressed through the new .onFrameProcessResult modifier. The closure receives a handle with access to processResult?.resultCompleteness, and you call handle.advanceToNextStep() directly — no separate publisher or delegate needed.

Migration

Before:

let scanningUxModel = BlinkIDUXModel(analyzer: analyzer)
scanningUxModel.$result
    .sink { [weak self] scanningResultState in
        if let scanningResultState {
            if let scanningResult = scanningResultState.scanningResult {
            } else {
            }
        }
    }
    .store(in: &cancellables)

After:

BlinkIDUXView(analyzer: analyzer) { scanningResultState in
    if let scanningResult = scanningResultState.scanningResult {
    } else {
    }
}

Buffer

MBSampleBufferWrapper has been removed — pass your CMSampleBuffer directly to CameraFrame when handling camera processing yourself.