feat: polish scenarios and stabilize keyboard layout height

Add preset-driven polish scenarios (Settings, onboarding, ScenarioChip) with ScenarioPrompt and style directives; drive keyboard height from content (240pt) and use viewIsAppearing encapsulated-height offset for smoother keyboard switches; remove redundant StatusBadge and hide system dictation via hasDictationKey.
2026-06-29 00:08:00 +08:00
parent 4fec0da7f0
commit 1bdb8824ac
32 changed files with 1247 additions and 340 deletions
@@ -0,0 +1,59 @@
+// ScenarioPrompt.swift
+// OSGKeyboard · Shared
+//
+// Builds the system prompt for a selected polish scenario. Output
+// format rules come from `ScenarioStyleDirective` (shared with
+// `TranslationPrompt`).
+
+import Foundation
+
+public enum ScenarioPrompt {
+
+    public static func make(
+        scenarioId: String,
+        providerId: String,
+        uiLanguage: AppUILanguage? = nil
+    ) -> String {
+        let isChineseNative = Self.isChineseNativeProvider(providerId)
+        let directive = ScenarioStyleDirective.make(
+            scenarioId: scenarioId,
+            providerId: providerId,
+            uiLanguage: uiLanguage
+        )
+        return isChineseNative
+            ? """
+            \(chineseBase)
+            \(directive)
+            """
+            : """
+            \(englishBase)
+            \(directive)
+            """
+    }
+
+    private static func isChineseNativeProvider(_ providerId: String) -> Bool {
+        ["zhipu", "moonshot", "qwen", "deepseek"].contains(providerId)
+    }
+
+    private static let chineseBase = """
+    你是一位语音输入润色助手。用户用 ASR 转写了一段可能含噪声的口述。
+    硬性要求:
+    1) 保留原意,不编造事实;保持输入语言。
+    2) 修复 ASR 噪声(同音错字、漏字、断句错乱)。
+    3) 去掉无意义的口头禅(嗯、啊、那个)。
+    4) 简洁;若下方场景未要求列表/分段,不超出原长 1.5 倍。
+    5) 若场景格式要求列表或分段,允许按格式组织;总长度不超过原长 2 倍。
+    6) 只输出润色后的正文,不要解释、不要加引号。
+    """
+
+    private static let englishBase = """
+    You are a voice-input polishing assistant. The user spoke informally and the transcript may contain ASR noise.
+    Hard rules:
+    1) Preserve meaning; do not invent facts; keep the input language.
+    2) Fix ASR noise (homophone errors, missing characters, broken segmentation).
+    3) Drop filler words (um, uh, like).
+    4) Stay concise; if the scenario below does not require lists/sections, do not exceed 1.5x the spoken length.
+    5) If the scenario requires bullets or paragraph breaks, use that layout; total length may be up to 2x when listing.
+    6) Output ONLY the polished text. No quotes, no explanation, no preamble.
+    """
+}