Configurable recap templates (categories per meeting type, in Settings)

Takeaways categories are no longer hardcoded — they're editable templates. A
template = the always-on TLDR + an ordered list of sections, each with a title, a
type (attributed items / bulleted list / paragraph), and an instruction (the prompt
text for that category). The analyzer assembles the LLM prompt FROM the template
and parses generically, so adding/removing/renaming a category needs zero code and
the output always renders.

- RecapTemplate / TemplateSection / SectionKind + TopicGranularity; built-in
  defaults (Internal Meeting, 1:1, Company/Sales Call), all editable.
- Generic extras: RecapExtras{tldr, primarySpeakers, sections:[RenderedSection]} +
  RecapItem{text,who,when,note} replaces the fixed MeetingExtras. Analyzer builds
  per-section sec_N fields + parses by kind; renderer + remap are generic.
- Topic granularity (coarse/auto/fine) answers 'should chunking be configurable' —
  it scales the target topic count; raw window sizes stay as tuned defaults.
- AppSettings persists templates + defaultTemplateId (seeded once). Settings gets a
  default-template picker + 'Manage…' → TemplatesView (CRUD, edit sections/
  instructions, set default, **Preview prompt** for full transparency).
- Recap editor gains a template picker; Regenerate uses the chosen template. Auto
  recap uses the default template.

54/54 XCTest (template prompt build, generic parse/remap/render updated).

Ten31Transcripts/Recap/RecapEditModel.swift

@@ -18,6 +18,8 @@ final class RecapEditModel: ObservableObject {
     private var originalSpeakers: [String]
     private var renameOps: [(from: String, to: String)] = []
 
+    let templates: [RecapTemplate]
+    @Published var selectedTemplateId: String
     @Published private(set) var segments: [SpeakersFile.Segment]
     @Published private(set) var speakers: [String]
     @Published private(set) var dirty = false
@@ -25,7 +27,8 @@ final class RecapEditModel: ObservableObject {
     @Published private(set) var hasRecap: Bool
     @Published private(set) var status: String?
 
-    init?(folder: URL, voiceprints: VoiceprintStore, baseURL: String, skipTLS: Bool) {
+    init?(folder: URL, voiceprints: VoiceprintStore, baseURL: String, skipTLS: Bool,
+          templates: [RecapTemplate], defaultTemplateId: String) {
         let speakersURL = folder.appendingPathComponent("speakers.json")
         guard let data = try? Data(contentsOf: speakersURL),
               let file = try? JSONDecoder().decode(SpeakersFile.self, from: data),
@@ -34,6 +37,9 @@ final class RecapEditModel: ObservableObject {
         self.voiceprints = voiceprints
         self.baseURL = baseURL
         self.skipTLS = skipTLS
+        self.templates = templates.isEmpty ? RecapTemplate.builtIns : templates
+        self.selectedTemplateId = (templates.contains { $0.id == defaultTemplateId } ? defaultTemplateId : templates.first?.id)
+            ?? RecapTemplate.builtIns.first!.id
         self.base = file
         self.segments = file.segments
         self.speakers = SpeakerEditing.orderedSpeakers(file.segments)
@@ -102,13 +108,14 @@ final class RecapEditModel: ObservableObject {
         defer { regenerating = false }
 
         let file = commitCorrections()
+        let template = templates.first { $0.id == selectedTemplateId } ?? templates.first ?? .internalMeeting
         let llm = GatewayLLMClient(baseURL: baseURL, skipTLS: skipTLS)
         guard let model = await llm.chatModelId() else {
             status = "No language model on the gateway — saved corrections only."
             rebaseline(); return
         }
         let analyzer = RecapAnalyzer(llm: llm, model: model)
-        guard let result = try? await analyzer.recap(file: file) else {
+        guard let result = try? await analyzer.recap(file: file, template: template) else {
             status = "Recap regeneration failed — corrections were saved."
             rebaseline(); return
         }