Files
Sybil-2/ios/Packages/Sybil/Sources/Sybil/SybilAgentAttachments.swift
T

301 lines
18 KiB
Swift
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
import CryptoKit
import Foundation
import ImageIO
import UniformTypeIdentifiers
import UIKit
struct AgentAttachmentCapabilities: Codable, Equatable, Sendable {
var version: Int
var maxCount: Int
var maxTotalBytes: Int
var maxImageBytes: Int
var maxFileBytes: Int
var imageMimeTypes: [String]
var maxImagePixels: Int
var maxImageDimension: Int
var quotaBytes: Int64? = nil
var usedBytes: Int64? = nil
var remainingBytes: Int64? = nil
var maxConcurrentUploads: Int? = nil
var maxConcurrentDecodes: Int? = nil
var supported: Bool {
version == 1 && maxCount > 0 && maxTotalBytes > 0 && maxImageBytes > 0 && maxFileBytes > 0
&& maxImagePixels > 0 && maxImageDimension > 0
}
var countLimit: Int { min(maxCount, 8) }
var totalLimit: Int { min(maxTotalBytes, 32 * 1024 * 1024) }
var imageLimit: Int { min(maxImageBytes, 6 * 1024 * 1024) }
var fileLimit: Int { min(maxFileBytes, 8 * 1024 * 1024) }
var pixelLimit: Int { min(maxImagePixels, 50_000_000) }
var dimensionLimit: Int { min(maxImageDimension, 16_384) }
func validate(_ attachments: [AgentAttachment]) throws {
guard supported else { throw AgentAttachmentIssue.unavailable }
for attachment in attachments { try attachment.validate() }
guard attachments.count <= countLimit, attachments.reduce(0, { $0 + $1.sizeBytes }) <= totalLimit else {
throw AgentAttachmentIssue.local("Attachments exceed the advertised message limit.", code: "attachment_limit_exceeded")
}
for attachment in attachments {
guard attachment.sizeBytes <= (attachment.kind == .image ? imageLimit : fileLimit),
attachment.kind != .image || imageMimeTypes.contains(attachment.mimeType) else {
throw AgentAttachmentIssue.local("An attachment exceeds the current Agent limits or uses an unavailable image format.", code: "attachment_limit_exceeded")
}
}
}
}
struct AgentAssistantImageCapabilities: Decodable, Equatable, Sendable {
var version: Int
var supported: Bool { version == 1 }
}
struct AgentCapabilities: Decodable, Sendable {
var attachments: AgentAttachmentCapabilities?
var assistantImages: AgentAssistantImageCapabilities?
init(attachments: AgentAttachmentCapabilities? = nil, assistantImages: AgentAssistantImageCapabilities? = nil) {
self.attachments = attachments
self.assistantImages = assistantImages
}
private enum CodingKeys: String, CodingKey { case attachments, assistantImages }
init(from decoder: Decoder) throws {
let values = try decoder.container(keyedBy: CodingKeys.self)
// Unknown versions or malformed future capabilities must not break text chat.
attachments = try? values.decode(AgentAttachmentCapabilities.self, forKey: .attachments)
assistantImages = try? values.decode(AgentAssistantImageCapabilities.self, forKey: .assistantImages)
}
}
struct AgentAttachment: Codable, Hashable, Identifiable, Sendable {
enum Kind: String, Codable, Sendable { case image, text, file }
var id: String
var kind: Kind
var filename: String
var mimeType: String
var sizeBytes: Int
var sha256: String
var workspacePath: String? = nil
func validate() throws {
guard !id.isEmpty, id.utf8.count <= 1024, !id.unicodeScalars.contains(where: CharacterSet.controlCharacters.contains),
try AgentAttachmentPreparation.filename(filename) == filename,
sizeBytes >= 0, sizeBytes <= 8 * 1024 * 1024,
sha256.count == 64, sha256.utf8.allSatisfy({ (48...57).contains($0) || (97...102).contains($0) }),
try AgentAttachmentPreparation.mime(mimeType) == mimeType else {
throw AgentAttachmentIssue.local("The Agent returned an invalid attachment descriptor.", code: "invalid_upstream_response")
}
guard (kind == .image) == AgentAttachmentPreparation.nativeMIMEs.contains(mimeType),
kind != .image || sizeBytes <= 6 * 1024 * 1024 else {
throw AgentAttachmentIssue.local("Invalid image attachment descriptor.", code: "invalid_upstream_response")
}
if let workspacePath {
guard workspacePath.hasPrefix("/"), !workspacePath.unicodeScalars.contains(where: CharacterSet.controlCharacters.contains) else {
throw AgentAttachmentIssue.local("The Agent returned an invalid workspace path.", code: "invalid_upstream_response")
}
}
}
var metadataValue: JSONValue {
// All stored fields are JSON-compatible; this also keeps optional future fields separate from ChatAttachment.
(try? JSONDecoder().decode(JSONValue.self, from: JSONEncoder().encode(self))) ?? .null
}
}
extension Message {
var agentAttachments: [AgentAttachment] {
guard role == .user || role == .assistant,
let info = metadata?.objectValue,
let entries = info["agentAttachments"]?.arrayValue else { return [] }
if role == .assistant {
guard info["kind"]?.stringValue == "message", info["status"]?.stringValue == "complete",
entries.count <= 8 else { return [] }
}
let attachments: [AgentAttachment] = entries.compactMap { entry in
// Inspect the original JSON before Codable can discard unknown fields. Output
// descriptors are deliberately path/URL-free; legacy user uploads stay compatible.
if role == .assistant {
let safeFields: Set<String> = ["id", "kind", "filename", "mimeType", "sizeBytes", "sha256"]
guard let fields = entry.objectValue, Set(fields.keys) == safeFields,
fields["kind"]?.stringValue == "image" else { return nil }
}
guard let data = try? JSONEncoder().encode(entry),
let attachment = try? JSONDecoder().decode(AgentAttachment.self, from: data),
(try? attachment.validate()) != nil else { return nil }
return attachment
}
if role == .assistant {
// Reject the entire output set, including partially invalid arrays. The
// count and per-image bounds above also bound this aggregate calculation.
guard attachments.count == entries.count,
Set(attachments.map(\.id)).count == attachments.count,
attachments.reduce(0, { $0 + $1.sizeBytes }) <= 32 * 1024 * 1024 else { return [] }
}
return attachments
}
}
struct AgentSubmission: Encodable, Sendable {
struct Reference: Encodable, Sendable { let id: String }
let requestId: String
let content: String
let attachments: [Reference]?
init(content: String, requestID: String, attachments: [AgentAttachment] = []) throws {
guard !requestID.isEmpty, requestID.utf16.count <= 256, content.utf16.count <= 40_000,
!content.trimmingCharacters(in: .whitespacesAndNewlines).isEmpty || !attachments.isEmpty,
attachments.count <= 8, Set(attachments.map(\.id)).count == attachments.count else {
throw AgentAttachmentIssue.local("Enter a message of at most 40,000 characters or attach files.", code: "invalid_attachment")
}
for attachment in attachments { try attachment.validate() }
guard attachments.reduce(0, { $0 + $1.sizeBytes }) <= 32 * 1024 * 1024 else {
throw AgentAttachmentIssue.local("Attachments exceed the total message size limit.", code: "attachment_limit_exceeded")
}
self.content = content; requestId = requestID
self.attachments = attachments.isEmpty ? nil : attachments.map { Reference(id: $0.id) }
}
var attachmentIDs: [String] { attachments?.map(\.id) ?? [] }
}
struct AgentAttachmentIssue: Error, LocalizedError, Codable, Equatable, Sendable {
var message: String
var code: String
var statusCode: Int?
// Preserve additive quota information without inventing field names.
var details: [String: JSONValue] = [:]
var retryAfter: String? = nil
var errorDescription: String? {
var result = "\(message) [\(code)]"
if code == "attachment_quota_exceeded" {
let numbers = ["requiredBytes", "remainingBytes", "quotaBytes"].compactMap { key -> String? in
guard let number = details[key]?.numberValue, number.isFinite, number >= 0, number < Double(Int64.max) else { return nil }
return "\(key): \(ByteCountFormatter.string(fromByteCount: Int64(number), countStyle: .file))"
}
if !numbers.isEmpty { result += " " + numbers.joined(separator: ", ") }
}
if let retryAfter { result += " Retry-After: \(retryAfter)." }
return result
}
static func local(_ message: String, code: String = "invalid_attachment") -> Self { Self(message: message, code: code) }
static let unavailable = Self.local("Attachments are unavailable on this Agent. Reconnect to check support, or remove the files to send text.", code: "attachments_unavailable")
static func wrapping(_ error: any Error) -> Self {
error as? Self ?? .local(error.localizedDescription, code: "attachment_transfer_failed")
}
static func response(_ data: Data, status: Int, retryAfter: String? = nil) -> Self {
let fields = (try? JSONDecoder().decode(JSONValue.self, from: data))?.objectValue ?? [:]
return Self(message: fields["message"]?.stringValue ?? fields["error"]?.stringValue ?? "Attachment request failed (HTTP \(status)).",
code: fields["code"]?.stringValue ?? "attachment_request_failed", statusCode: status, details: fields, retryAfter: retryAfter)
}
}
struct AgentPreparedAttachment: Sendable {
var filename: String
var mimeType: String
var data: Data
var sha256: String { SHA256.hash(data: data).map { String(format: "%02x", $0) }.joined() }
}
enum AgentAttachmentPreparation {
static func offMain<T: Sendable>(_ operation: @escaping @Sendable () throws -> T) async throws -> T {
let task = Task.detached(operation: operation)
return try await withTaskCancellationHandler { try await task.value } onCancel: { task.cancel() }
}
static let nativeMIMEs: Set<String> = ["image/png", "image/jpeg", "image/gif", "image/webp"]
static func filename(_ value: String) throws -> String {
let value = value.precomposedStringWithCanonicalMapping
guard !value.isEmpty, value.utf8.count <= 255, value != ".", value != "..",
!value.contains("/"), !value.contains("\\"),
!value.unicodeScalars.contains(where: CharacterSet.controlCharacters.contains) else {
throw AgentAttachmentIssue.local("The filename must be a safe basename of 1–255 UTF-8 bytes.")
}
return value
}
static func mime(_ value: String) throws -> String {
let value = value.lowercased().trimmingCharacters(in: .whitespacesAndNewlines)
let parts = value.split(separator: "/", omittingEmptySubsequences: false)
let allowed = CharacterSet(charactersIn: "abcdefghijklmnopqrstuvwxyz0123456789!#$%&'*+-.^_`|~")
guard parts.count == 2, parts.allSatisfy({ !$0.isEmpty && $0.unicodeScalars.allSatisfy(allowed.contains) }), value.utf8.count <= 255 else {
throw AgentAttachmentIssue.local("The selected file has an invalid media type.")
}
return value
}
static func file(_ url: URL, limits: AgentAttachmentCapabilities) throws -> AgentPreparedAttachment {
let accessed = url.startAccessingSecurityScopedResource()
defer { if accessed { url.stopAccessingSecurityScopedResource() } }
let name = try filename(url.lastPathComponent)
let type = (try? url.resourceValues(forKeys: [.contentTypeKey]))?.contentType ?? UTType(filenameExtension: url.pathExtension)
let media = try mime(type?.preferredMIMEType ?? "application/octet-stream")
let handle = try FileHandle(forReadingFrom: url)
defer { try? handle.close() }
var data = Data()
while let chunk = try handle.read(upToCount: 64 * 1024), !chunk.isEmpty {
try Task.checkCancellation()
guard data.count + chunk.count <= (nativeMIMEs.contains(media) ? limits.imageLimit : limits.fileLimit) else {
throw AgentAttachmentIssue.local("\(name) exceeds the file size limit.", code: "attachment_limit_exceeded")
}
data.append(chunk)
}
return try prepare(data, filename: name, mimeType: media, limits: limits)
}
static func prepare(_ data: Data, filename name: String, mimeType: String, limits: AgentAttachmentCapabilities) throws -> AgentPreparedAttachment {
guard limits.supported else { throw AgentAttachmentIssue.unavailable }
let name = try filename(name), media = try mime(mimeType)
guard data.count <= (nativeMIMEs.contains(media) ? limits.imageLimit : limits.fileLimit) else { throw AgentAttachmentIssue.local("\(name) exceeds the file size limit.", code: "attachment_limit_exceeded") }
if nativeMIMEs.contains(media) {
guard limits.imageMimeTypes.contains(media), data.count <= limits.imageLimit else {
throw AgentAttachmentIssue.local("\(name) exceeds the image limit or uses an unavailable image format.", code: "attachment_limit_exceeded")
}
_ = try raster(data, mime: media, pixels: limits.pixelLimit, dimension: limits.dimensionLimit)
}
return AgentPreparedAttachment(filename: name, mimeType: media, data: data)
}
static func raster(_ data: Data, mime: String, pixels: Int = 50_000_000, dimension: Int = 16_384) throws -> CGImage {
guard nativeMIMEs.contains(mime), let source = CGImageSourceCreateWithData(data as CFData, nil),
let type = CGImageSourceGetType(source), UTType(type as String)?.preferredMIMEType == mime,
CGImageSourceGetStatus(source) == .statusComplete,
let properties = CGImageSourceCopyPropertiesAtIndex(source, 0, nil) as? [CFString: Any],
let width = properties[kCGImagePropertyPixelWidth] as? Int, let height = properties[kCGImagePropertyPixelHeight] as? Int,
width > 0, height > 0, width <= dimension, height <= dimension, width <= pixels / height,
let image = CGImageSourceCreateThumbnailAtIndex(source, 0, [kCGImageSourceCreateThumbnailFromImageAlways: true,
kCGImageSourceThumbnailMaxPixelSize: 600, kCGImageSourceCreateThumbnailWithTransform: true] as CFDictionary) else {
throw AgentAttachmentIssue.local("The image is invalid, has a mismatched media type, or exceeds the dimension/pixel limit.", code: "invalid_image")
}
return image
}
static func photo(_ data: Data, filename originalName: String?, mimeType: String?, limits: AgentAttachmentCapabilities) throws -> AgentPreparedAttachment {
let media = mimeType ?? "application/octet-stream"
if nativeMIMEs.contains(media) {
return try prepare(data, filename: originalName ?? "photo.\(UTType(mimeType: media)?.preferredFilenameExtension ?? "jpg")", mimeType: media, limits: limits)
}
// Photo Library is an explicit visual-input selection. Convert HEIC/HEIF with bounded decoding.
guard let source = CGImageSourceCreateWithData(data as CFData, nil), let sourceType = CGImageSourceGetType(source),
["public.heic", "public.heif"].contains(sourceType as String), limits.imageMimeTypes.contains("image/jpeg"),
let properties = CGImageSourceCopyPropertiesAtIndex(source, 0, nil) as? [CFString: Any],
let width = properties[kCGImagePropertyPixelWidth] as? Int, let height = properties[kCGImagePropertyPixelHeight] as? Int,
width > 0, height > 0, width <= limits.dimensionLimit, height <= limits.dimensionLimit, width <= limits.pixelLimit / height,
let image = CGImageSourceCreateThumbnailAtIndex(source, 0, [kCGImageSourceCreateThumbnailFromImageAlways: true,
kCGImageSourceThumbnailMaxPixelSize: min(2000, limits.dimensionLimit), kCGImageSourceCreateThumbnailWithTransform: true] as CFDictionary),
let jpeg = UIImage(cgImage: image).jpegData(compressionQuality: 0.88) else {
throw AgentAttachmentIssue.local("This photo could not be prepared as a supported image. Select its original in Files to upload it as a file.", code: "invalid_image")
}
let safe = try filename(originalName ?? "photo.heic")
let stem = (safe as NSString).deletingPathExtension
return try prepare(jpeg, filename: stem + ".jpg", mimeType: "image/jpeg", limits: limits)
}
}
struct AgentMultipart: Sendable {
let boundary: String
let body: Data
init(file: AgentPreparedAttachment, uploadID: String) throws {
guard UUID(uuidString: uploadID) != nil else { throw AgentAttachmentIssue.local("Invalid upload identity.") }
let name = try AgentAttachmentPreparation.filename(file.filename)
let media = try AgentAttachmentPreparation.mime(file.mimeType)
var boundary: String
repeat { boundary = "Sybil-" + UUID().uuidString } while file.data.range(of: Data(boundary.utf8)) != nil
self.boundary = boundary
let quoted = name.replacingOccurrences(of: "\"", with: "\\\"")
var bytes = Data("--\(boundary)\r\nContent-Disposition: form-data; name=\"uploadId\"\r\n\r\n\(uploadID)\r\n--\(boundary)\r\nContent-Disposition: form-data; name=\"file\"; filename=\"\(quoted)\"\r\nContent-Type: \(media)\r\n\r\n".utf8)
bytes.append(file.data)
bytes.append(Data("\r\n--\(boundary)--\r\n".utf8))
guard bytes.count <= 8 * 1024 * 1024 + 64 * 1024 else { throw AgentAttachmentIssue.local("The upload exceeds the request size limit.", code: "attachment_limit_exceeded") }
body = bytes
}
}