Files
firebase-ios-sdk/GeminiLanguageModel/Sources/GeminiAPIDataModels/GenerateContent/ResponseFormatConfig.swift
T
Andrew Heard 3796a55fdf feat(ai): add GeminiLanguageModel (#16601)
Implemented `GeminiLanguageModel`, providing an adapter for Google's
Gemini models conforming to Apple's `FoundationModels.LanguageModel`
protocol. This PR includes text to text generation capabilities; other
modalities and Gemini features will be handled in follow-up PRs.

**Note:** The `GeminiLanguageModel` implementation is guarded by the
environment variable `GEMINI_LANGUAGE_MODEL`. By default none of the
`GeminiLanguageModel` implementation is included.

Key additions:
- `GeminiLanguageModel`: Implements `LanguageModel` protocol with
  reasoning capabilities, configuring execution via `ModelResource`
  and `EndpointConfiguration`.
- `GeminiLanguageModelExecutor`: Converts prompt instructions and
  history into Gemini streaming requests, yielding responses as
  `LanguageModelSession.ResponseStream`.
- `GeminiTranscriptTranslator`: Translates Apple `Transcript` entries
  and segments into Gemini `Content` structures while preserving thought
  signatures and validating supported transcript features.
- `GeminiLanguageModel.Error` and `GeminiErrorMapper`: Translates Gemini
  API, HTTP, and network errors into standard `LanguageModelError` cases
  (`rateLimited`, `timeout`, `guardrailViolation`, `refusal`, and
  `unsupportedLanguageOrLocale`) or `GeminiLanguageModel.Error` cases
  (`serviceUnavailable`, `networkFailure`, `modelNotFound`, `apiError`).
  The `APIError` payload is designed for seamless forward compatibility
  with both GenerateContent and the upcoming Interactions API.
- Testing infrastructure: Adds `ConditionTrait.requireFoundationModels`
  to conditionally run Foundation Models tests only on supported
  platforms (macOS 27+, iOS 27+, watchOS 27+, visionOS 27+), along with
  shared test constants, mock fixtures, and comprehensive unit and
  integration test suites.
2026-09-04 18:05:45 -04:00

110 lines
4.0 KiB
Swift

// Copyright 2026 Google LLC
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
import Foundation
/// Configuration for the response output format.
/// Handles single object serialization for Gemini Developer API and array-wrapped serialization for Gemini Enterprise Agent Platform (Vertex AI).
package struct ResponseFormatConfig: Codable, Sendable, Equatable, Hashable {
/// Optional. Text output format configuration.
package var text: TextResponseFormat?
/// Optional. Audio output format configuration.
package var audio: AudioResponseFormat?
/// Optional. Image output format configuration.
package var image: ImageResponseFormat?
/// Optional. Video output format configuration.
package var video: VideoResponseFormat?
/// Creates a new `ResponseFormatConfig`.
package init(
text: TextResponseFormat? = nil,
audio: AudioResponseFormat? = nil,
image: ImageResponseFormat? = nil,
video: VideoResponseFormat? = nil
) {
self.text = text
self.audio = audio
self.image = image
self.video = video
}
package init(from decoder: any Decoder) throws {
let container = try decoder.singleValueContainer()
if let array = try? container.decode([ResponseFormatConfig].self), let first = array.first {
self.text = first.text
self.audio = first.audio
self.image = first.image
self.video = first.video
} else {
let objectContainer = try decoder.container(keyedBy: CodingKeys.self)
self.text = try objectContainer.decodeIfPresent(TextResponseFormat.self, forKey: .text)
self.audio = try objectContainer.decodeIfPresent(AudioResponseFormat.self, forKey: .audio)
self.image = try objectContainer.decodeIfPresent(ImageResponseFormat.self, forKey: .image)
self.video = try objectContainer.decodeIfPresent(VideoResponseFormat.self, forKey: .video)
}
}
package func encode(to encoder: any Encoder) throws {
if let useArrayFormat = encoder.userInfo[ResponseFormatConfig.useArrayFormatKey] as? Bool,
useArrayFormat
{
var container = encoder.unkeyedContainer()
let helper = EncodingHelper(text: text, audio: audio, image: image, video: video)
try container.encode(helper)
} else {
var container = encoder.container(keyedBy: CodingKeys.self)
try container.encodeIfPresent(text, forKey: .text)
try container.encodeIfPresent(audio, forKey: .audio)
try container.encodeIfPresent(image, forKey: .image)
try container.encodeIfPresent(video, forKey: .video)
}
}
private enum CodingKeys: String, CodingKey {
case text
case audio
case image
case video
}
private struct EncodingHelper: Encodable {
let text: TextResponseFormat?
let audio: AudioResponseFormat?
let image: ImageResponseFormat?
let video: VideoResponseFormat?
enum CodingKeys: String, CodingKey {
case text
case audio
case image
case video
}
func encode(to encoder: any Encoder) throws {
var container = encoder.container(keyedBy: CodingKeys.self)
try container.encodeIfPresent(text, forKey: .text)
try container.encodeIfPresent(audio, forKey: .audio)
try container.encodeIfPresent(image, forKey: .image)
try container.encodeIfPresent(video, forKey: .video)
}
}
/// Key to put in encoder.userInfo to control whether to encode this type as an array of objects
/// (Gemini Enterprise Agent Platform / Vertex AI) or as a single flat object (Gemini Developer API).
package static let useArrayFormatKey = CodingUserInfoKey(rawValue: "useArrayFormatKey")!
}