Adapters: add Meet, Zoom, Teams (coloured border) + adapter registry

Front-loads the remaining visual adapters per the Signal→Meet→Zoom priority.
All three reuse GridCallAnalyzer's coloured-border (saturated) detection path
and share the new bottom-left name anchor:

- GridCallAnalyzer: generalise nameAtBottom:Bool into a NameAnchor enum
  (.bottomCenter for Signal's centered footer, .bottomLeft for Meet/Zoom/Teams
  where the name hugs the tile's bottom-left corner, .center for completeness).
  tileRect estimates the tile up-and-right of a bottom-left name.
- MeetAdapter (Google-blue ring, browser-hosted), ZoomAdapter (green/yellow
  border), TeamsAdapter (violet ring): coloured-border on, white-border off,
  bottom-left names. Geometry constants are first-pass pending real fixtures.
- AdapterRegistry.adapter(for:) maps CallDetector.DetectedApp → AppAdapter so
  VisualObserver can be constructed when live visual capture is wired in;
  unmapped apps degrade to audio-only.

Synthetic 4-tile tests: Meet picks each blue-bordered speaker with no
adjacent-tile bleed, Zoom picks the green-bordered speaker, and Signal's
white-only detector correctly ignores a coloured border. 18/18 XCTest pass.
This commit is contained in:
Grant Gilliam
2026-06-06 09:57:53 -05:00
parent a56b47143c
commit c347acbd97
7 changed files with 230 additions and 9 deletions
@@ -0,0 +1,43 @@
import Foundation
import CoreVideo
/// Zoom adapter (native app: us.zoom.xos).
///
/// Zoom's active-speaker cue is a **coloured border** around the speaking tile
/// by default a green/yellow outline (configurable in Zoom settings). The
/// participant **name sits in the tile's bottom-LEFT corner**, so the tile is
/// estimated extending up and to the right of the name.
///
/// Gotchas to calibrate against real fixtures later:
/// - **Speaker view** shows one big tile; the active speaker fills it (no useful
/// per-tile border) handle by attributing speech to the large tile.
/// - The self-view tile and screen-share change the layout.
///
/// Detection *logic* is validated on synthetic frames; geometry constants are a
/// first pass pending real Zoom screenshots.
struct ZoomAdapter: AppAdapter {
static let bundleIDs = ["us.zoom.xos"]
let adapterVersion = "zoom-0.1.0"
let preferredFPS = 3
private let analyzer: GridCallAnalyzer
init() {
var config = GridCallAnalyzer.Config()
config.nameAnchor = .bottomLeft
config.detectColoredBorder = true // green/yellow speaking border
config.detectWhiteBorder = false
config.tileExpandX = 3.0
config.tileExpandY = 5.0
self.analyzer = GridCallAnalyzer(config: config)
}
func analyze(frame: CVPixelBuffer, at t: TimeInterval) -> [SpeakerObservation] {
analyzer.analyze(pixelBuffer: frame, at: t)
}
// Exposed for fixture/synthetic tests.
func analyze(cgImage: CGImage, at t: TimeInterval) -> [SpeakerObservation] {
analyzer.analyze(cgImage: cgImage, at: t)
}
}