novelws 5.2.0 → 5.3.0

package/templates/scripts/db_context.py

@@ -0,0 +1,609 @@
+#!/usr/bin/env python3
+"""db_context.py — 从 DB 生成精确最小上下文
+
+用法:
+    python scripts/db_context.py --chapter 42 --mode write
+    python scripts/db_context.py --chapter 42 --mode expand
+    python scripts/db_context.py --chapter 42 --mode analyze
+    python scripts/db_context.py --dashboard
+
+输出 markdown 格式的上下文摘要到 stdout，可重定向到文件供 AI 读取。
+"""
+
+import argparse
+import json
+import os
+import sys
+import psycopg2
+
+BASE = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+SCHEMA = "novelws"
+
+
+def load_config():
+    config_path = os.path.join(BASE, "resources", "config.json")
+    with open(config_path, "r", encoding="utf-8") as f:
+        return json.load(f)
+
+
+def get_db_config(config):
+    db = config.get("database", {})
+    return {
+        "host": db.get("host", "127.0.0.1"),
+        "port": db.get("port", 5432),
+        "dbname": db.get("dbname", "postgres"),
+        "user": db.get("user", "postgres"),
+        "password": db.get("password", ""),
+    }
+
+
+def get_conn():
+    config = load_config()
+    return psycopg2.connect(**get_db_config(config))
+
+
+def get_chapter_info(cur, global_ch):
+    """获取章节基本信息和所属卷"""
+    cur.execute(f"""
+        SELECT ch.id, ch.chapter_number, ch.time_in_story, ch.status,
+               v.id as vol_id, v.vol_number, v.title as vol_title,
+               p.phase_number, p.title as phase_title
+        FROM {SCHEMA}.chapters ch
+        JOIN {SCHEMA}.volumes v ON ch.volume_id = v.id
+        JOIN {SCHEMA}.phases p ON v.phase_id = p.id
+        WHERE ch.global_chapter_number = %s
+    """, (global_ch,))
+    return cur.fetchone()
+
+
+def context_write(cur, global_ch):
+    """生成 /write 模式的上下文"""
+    info = get_chapter_info(cur, global_ch)
+    if not info:
+        print(f"## 错误: 章节 {global_ch} 不存在于数据库中")
+        return
+
+    ch_id, ch_num, time_in_story, ch_status, vol_id, vol_num, vol_title, phase_num, phase_title = info
+
+    print(f"# /write 上下文 — 第{global_ch}章 (vol-{vol_num:03d} ch-{ch_num:03d})")
+    print(f"纪元{phase_num} [{phase_title}] · {vol_title}")
+    print()
+
+    # 1. 当前卷活跃角色
+    print("## 活跃角色")
+    cur.execute(f"""
+        SELECT c.name, c.role_type, cs.location, cs.state_summary, cs.last_appearance
+        FROM {SCHEMA}.characters c
+        JOIN {SCHEMA}.character_states cs ON c.id = cs.character_id
+        WHERE cs.volume_id = %s
+        ORDER BY cs.last_appearance DESC
+    """, (vol_id,))
+    rows = cur.fetchall()
+    if rows:
+        print("| 角色 | 身份 | 位置 | 状态 | 最后出场 |")
+        print("|------|------|------|------|---------|")
+        for r in rows:
+            print(f"| {r[0]} | {r[1]} | {r[2] or '-'} | {_truncate(r[3], 60)} | ch-{r[4]} |")
+    print()
+
+    # 2. 未解决伏笔
+    print("## 未解决伏笔")
+    cur.execute(f"""
+        SELECT f.code, f.description, f.status, ch.chapter_number as planted_at, f.note
+        FROM {SCHEMA}.foreshadowing f
+        JOIN {SCHEMA}.chapters ch ON f.planted_chapter_id = ch.id
+        WHERE f.status IN ('planted', 'hinted', 'partially_resolved')
+          AND ch.volume_id = %s
+        ORDER BY ch.chapter_number
+    """, (vol_id,))
+    rows = cur.fetchall()
+    if rows:
+        print(f"共 {len(rows)} 条:")
+        for r in rows:
+            note = f" ({r[4]})" if r[4] else ""
+            print(f"- [{r[0]}] ch-{r[3]:03d} [{r[2]}]: {r[1]}{note}")
+    else:
+        print("无")
+    print()
+
+    # 3. 活跃情节线
+    print("## 活跃情节线")
+    cur.execute(f"""
+        SELECT name, status, description
+        FROM {SCHEMA}.plot_threads
+        WHERE status NOT IN ('已完成')
+        ORDER BY id
+    """)
+    rows = cur.fetchall()
+    if rows:
+        for r in rows:
+            print(f"- **{r[0]}** [{r[1]}]: {_truncate(r[2], 80)}")
+    print()
+
+    # 4. 前5章概要摘要
+    print("## 前序章节")
+    start_ch = max(1, ch_num - 5)
+    cur.execute(f"""
+        SELECT chapter_number, time_in_story, synopsis_summary
+        FROM {SCHEMA}.chapters
+        WHERE volume_id = %s AND chapter_number >= %s AND chapter_number < %s
+        ORDER BY chapter_number
+    """, (vol_id, start_ch, ch_num))
+    rows = cur.fetchall()
+    if rows:
+        for r in rows:
+            print(f"- ch-{r[0]:03d} [{r[1] or '?'}]: {r[2] or '(无摘要)'}")
+    else:
+        print("无前序章节")
+    print()
+
+    # 5. 角色关系
+    print("## 关键关系")
+    cur.execute(f"""
+        SELECT ca.name, cb.name, r.relationship_type, r.current_status
+        FROM {SCHEMA}.relationships r
+        JOIN {SCHEMA}.characters ca ON r.character_a_id = ca.id
+        JOIN {SCHEMA}.characters cb ON r.character_b_id = cb.id
+        WHERE r.volume_id = %s
+        ORDER BY r.last_update_chapter DESC
+        LIMIT 10
+    """, (vol_id,))
+    rows = cur.fetchall()
+    if rows:
+        for r in rows:
+            print(f"- {r[0]} → {r[1]} [{r[2]}]: {_truncate(r[3], 60)}")
+    print()
+
+    # 6. 时间线（最近5条）
+    print("## 最近时间线")
+    cur.execute(f"""
+        SELECT ch.chapter_number, te.story_time, te.event_description
+        FROM {SCHEMA}.timeline_events te
+        JOIN {SCHEMA}.chapters ch ON te.chapter_id = ch.id
+        WHERE te.volume_id = %s AND ch.chapter_number < %s
+        ORDER BY ch.chapter_number DESC
+        LIMIT 5
+    """, (vol_id, ch_num))
+    rows = cur.fetchall()
+    if rows:
+        for r in reversed(rows):
+            print(f"- ch-{r[0]:03d} [{r[1]}]: {r[2]}")
+    print()
+
+    # 7. 主角状态
+    print("## 主角状态")
+    cur.execute(f"""
+        SELECT level, progress_pct FROM {SCHEMA}.cultivation_curve
+        ORDER BY chapter_number DESC LIMIT 1
+    """)
+    cult = cur.fetchone()
+    if cult:
+        print(f"- 修炼: {cult[0]} ({cult[1]}%)")
+
+    cur.execute(f"""
+        SELECT skill_name FROM {SCHEMA}.skill_overview
+        WHERE status = 'active' ORDER BY acquired_chapter
+    """)
+    skills = cur.fetchall()
+    if skills:
+        names = '、'.join(r[0] for r in skills[:4])
+        extra = f" 等 (共{len(skills)}项)" if len(skills) > 4 else ""
+        print(f"- 活跃技能: {names}{extra}")
+
+    cur.execute(f"""
+        SELECT item_name, quantity FROM {SCHEMA}.current_inventory
+        ORDER BY item_type LIMIT 5
+    """)
+    items = cur.fetchall()
+    if items:
+        item_strs = [f"{r[0]}×{r[1]}" if r[1] > 1 else r[0] for r in items]
+        extra = " 等" if len(items) >= 5 else ""
+        print(f"- 关键道具: {'、'.join(item_strs)}{extra}")
+    print()
+
+
+def context_expand(cur, global_ch):
+    """生成 /expand 模式的上下文 — 更精确，只加载本章相关数据"""
+    info = get_chapter_info(cur, global_ch)
+    if not info:
+        print(f"## 错误: 章节 {global_ch} 不存在于数据库中")
+        return
+
+    ch_id, ch_num, time_in_story, ch_status, vol_id, vol_num, vol_title, phase_num, phase_title = info
+
+    print(f"# /expand 上下文 — 第{global_ch}章 (vol-{vol_num:03d} ch-{ch_num:03d})")
+    print(f"纪元{phase_num} [{phase_title}] · {vol_title}")
+    print(f"故事时间: {time_in_story or '未知'}")
+    print()
+
+    # 1. 本章关联的伏笔
+    print("## 本章伏笔")
+    cur.execute(f"""
+        SELECT f.code, f.description, f.status, cf.action_type,
+               planted_ch.chapter_number as planted_at
+        FROM {SCHEMA}.chapter_foreshadowing cf
+        JOIN {SCHEMA}.foreshadowing f ON cf.foreshadowing_id = f.id
+        LEFT JOIN {SCHEMA}.chapters planted_ch ON f.planted_chapter_id = planted_ch.id
+        WHERE cf.chapter_id = %s
+        ORDER BY cf.action_type
+    """, (ch_id,))
+    rows = cur.fetchall()
+    if rows:
+        for r in rows:
+            print(f"- [{r[0]}] {r[3]}: {r[1]} (埋设于ch-{r[4]:03d}, 状态:{r[2]})")
+    else:
+        # 回退：显示当前卷所有未解决伏笔
+        print("(本章无直接关联伏笔，显示当前卷未解决伏笔)")
+        cur.execute(f"""
+            SELECT f.code, f.description, f.status, ch.chapter_number
+            FROM {SCHEMA}.foreshadowing f
+            JOIN {SCHEMA}.chapters ch ON f.planted_chapter_id = ch.id
+            WHERE f.status IN ('planted', 'hinted', 'partially_resolved')
+              AND ch.volume_id = %s
+            ORDER BY ch.chapter_number
+        """, (vol_id,))
+        rows = cur.fetchall()
+        for r in rows:
+            print(f"- [{r[0]}] ch-{r[3]:03d} [{r[2]}]: {r[1]}")
+    print()
+
+    # 2. 当前卷活跃角色（精简版）
+    print("## 活跃角色状态")
+    cur.execute(f"""
+        SELECT c.name, c.role_type, cs.location, cs.state_summary, cs.last_appearance
+        FROM {SCHEMA}.characters c
+        JOIN {SCHEMA}.character_states cs ON c.id = cs.character_id
+        WHERE cs.volume_id = %s
+        ORDER BY cs.last_appearance DESC
+    """, (vol_id,))
+    rows = cur.fetchall()
+    if rows:
+        for r in rows:
+            print(f"- **{r[0]}** ({r[1]}): {r[2] or '?'} — {_truncate(r[3], 80)} [ch-{r[4]}]")
+    print()
+
+    # 3. 相关角色关系
+    print("## 角色关系")
+    cur.execute(f"""
+        SELECT ca.name, cb.name, r.relationship_type, r.current_status
+        FROM {SCHEMA}.relationships r
+        JOIN {SCHEMA}.characters ca ON r.character_a_id = ca.id
+        JOIN {SCHEMA}.characters cb ON r.character_b_id = cb.id
+        WHERE r.volume_id = %s
+        ORDER BY r.last_update_chapter DESC
+    """, (vol_id,))
+    rows = cur.fetchall()
+    if rows:
+        for r in rows:
+            print(f"- {r[0]} → {r[1]} [{r[2]}]: {_truncate(r[3], 60)}")
+    print()
+
+    # 4. 前2章时间线（衔接用）
+    print("## 前序衔接")
+    cur.execute(f"""
+        SELECT ch.chapter_number, te.story_time, te.event_description
+        FROM {SCHEMA}.timeline_events te
+        JOIN {SCHEMA}.chapters ch ON te.chapter_id = ch.id
+        WHERE te.volume_id = %s AND ch.chapter_number < %s
+        ORDER BY ch.chapter_number DESC
+        LIMIT 3
+    """, (vol_id, ch_num))
+    rows = cur.fetchall()
+    if rows:
+        for r in reversed(rows):
+            print(f"- ch-{r[0]:03d} [{r[1]}]: {r[2]}")
+    print()
+
+    # 5. 主角能力
+    print("## 主角能力")
+    cur.execute(f"""
+        SELECT level, progress_pct FROM {SCHEMA}.cultivation_curve
+        ORDER BY chapter_number DESC LIMIT 1
+    """)
+    cult = cur.fetchone()
+    if cult:
+        print(f"- 当前修为: {cult[0]} ({cult[1]}%)")
+
+    cur.execute(f"""
+        SELECT skill_name, skill_category, description FROM {SCHEMA}.skill_overview
+        WHERE status = 'active' ORDER BY acquired_chapter
+    """)
+    skills = cur.fetchall()
+    if skills:
+        print("- 可用技能:")
+        for s in skills:
+            print(f"  - {s[0]} [{s[1]}]: {_truncate(s[2], 40)}")
+
+    cur.execute(f"""
+        SELECT item_name, quantity FROM {SCHEMA}.current_inventory
+        ORDER BY item_type
+    """)
+    items = cur.fetchall()
+    if items:
+        print("- 可用道具:")
+        for it in items:
+            qty = f" ×{it[1]}" if it[1] > 1 else ""
+            print(f"  - {it[0]}{qty}")
+    print()
+
+
+def context_analyze(cur, global_ch):
+    """生成 /analyze 模式的上下文 — 一致性校验数据"""
+    info = get_chapter_info(cur, global_ch)
+    if not info:
+        print(f"## 错误: 章节 {global_ch} 不存在于数据库中")
+        return
+
+    ch_id, ch_num, time_in_story, ch_status, vol_id, vol_num, vol_title, phase_num, phase_title = info
+
+    print(f"# /analyze 上下文 — 第{global_ch}章 (vol-{vol_num:03d} ch-{ch_num:03d})")
+    print()
+
+    # 1. 本章应有的伏笔
+    print("## 本章伏笔检查清单")
+    cur.execute(f"""
+        SELECT f.code, f.description, cf.action_type
+        FROM {SCHEMA}.chapter_foreshadowing cf
+        JOIN {SCHEMA}.foreshadowing f ON cf.foreshadowing_id = f.id
+        WHERE cf.chapter_id = %s
+    """, (ch_id,))
+    rows = cur.fetchall()
+    if rows:
+        for r in rows:
+            print(f"- [{r[0]}] 应{r[2]}: {r[1]}")
+    else:
+        print("无直接关联伏笔")
+    print()
+
+    # 2. 角色状态（用于一致性对比）
+    print("## 角色状态基准")
+    cur.execute(f"""
+        SELECT c.name, cs.location, cs.state_summary, cs.last_appearance
+        FROM {SCHEMA}.characters c
+        JOIN {SCHEMA}.character_states cs ON c.id = cs.character_id
+        WHERE cs.volume_id = %s
+        ORDER BY c.name
+    """, (vol_id,))
+    rows = cur.fetchall()
+    for r in rows:
+        print(f"- {r[0]}: {r[1]} — {_truncate(r[2], 60)} [最后ch-{r[3]}]")
+    print()
+
+    # 3. 时间线连续性
+    print("## 时间线连续性")
+    cur.execute(f"""
+        SELECT ch.chapter_number, te.story_time, te.event_description
+        FROM {SCHEMA}.timeline_events te
+        JOIN {SCHEMA}.chapters ch ON te.chapter_id = ch.id
+        WHERE te.volume_id = %s AND ch.chapter_number BETWEEN %s AND %s
+        ORDER BY ch.chapter_number
+    """, (vol_id, max(1, ch_num - 2), ch_num + 1))
+    rows = cur.fetchall()
+    for r in rows:
+        marker = " <<<" if r[0] == ch_num else ""
+        print(f"- ch-{r[0]:03d} [{r[1]}]: {r[2]}{marker}")
+    print()
+
+    # 4. 全局一致性警告
+    print("## 一致性警告")
+    # 检测遗忘伏笔（埋设超过50章未解决）
+    cur.execute(f"""
+        SELECT f.code, f.description, ch.chapter_number as planted_at
+        FROM {SCHEMA}.foreshadowing f
+        JOIN {SCHEMA}.chapters ch ON f.planted_chapter_id = ch.id
+        WHERE f.status = 'planted' AND ch.chapter_number < %s - 50
+        ORDER BY ch.chapter_number
+    """, (ch_num,))
+    rows = cur.fetchall()
+    if rows:
+        print("### 可能遗忘的伏笔（埋设超过50章）:")
+        for r in rows:
+            print(f"- [{r[0]}] 埋设于ch-{r[2]:03d}: {r[1]}")
+    else:
+        print("无一致性警告")
+    print()
+
+    # 5. 主角能力基准
+    print("## 主角能力基准")
+    cur.execute(f"""
+        SELECT level FROM {SCHEMA}.cultivation_curve
+        ORDER BY chapter_number DESC LIMIT 1
+    """)
+    cult = cur.fetchone()
+    if cult:
+        print(f"- 修炼等级: {cult[0]}")
+
+    cur.execute(f"""
+        SELECT skill_name, acquired_chapter FROM {SCHEMA}.skill_overview
+        WHERE status = 'active' ORDER BY acquired_chapter
+    """)
+    skills = cur.fetchall()
+    if skills:
+        print("- 已习得技能:")
+        for s in skills:
+            print(f"  - {s[0]} (ch-{s[1]:03d})")
+
+    cur.execute(f"""
+        SELECT item_name, quantity FROM {SCHEMA}.current_inventory
+        ORDER BY item_type
+    """)
+    items = cur.fetchall()
+    if items:
+        print("- 持有道具:")
+        for it in items:
+            qty = f" ×{it[1]}" if it[1] > 1 else ""
+            print(f"  - {it[0]}{qty}")
+    print()
+
+
+def dashboard(cur):
+    """输出创作进度仪表盘"""
+    print("# 创作进度仪表盘")
+    print()
+
+    cur.execute(f"SELECT * FROM {SCHEMA}.writing_dashboard")
+    rows = cur.fetchall()
+
+    print("## 纪元进度")
+    print("| 纪 | 标题 | 总卷 | 完成卷 | 总章 | 终稿章 | 概要章 | 总字数 |")
+    print("|----|------|------|--------|------|--------|--------|--------|")
+    for r in rows:
+        words = f"{r[7]:,}" if r[7] else "0"
+        print(f"| {r[0]} | {r[1]} | {r[2]} | {r[3]} | {r[4]} | {r[5]} | {r[6]} | {words} |")
+    print()
+
+    # 伏笔统计
+    print("## 伏笔统计")
+    cur.execute(f"""
+        SELECT status, COUNT(*) FROM {SCHEMA}.foreshadowing GROUP BY status ORDER BY status
+    """)
+    rows = cur.fetchall()
+    for r in rows:
+        print(f"- {r[0]}: {r[1]} 条")
+    print()
+
+    # 角色统计
+    print("## 角色统计")
+    cur.execute(f"""
+        SELECT status, COUNT(*) FROM {SCHEMA}.characters GROUP BY status ORDER BY status
+    """)
+    rows = cur.fetchall()
+    for r in rows:
+        print(f"- {r[0]}: {r[1]} 个")
+    print()
+
+    # 情节线统计
+    print("## 情节线统计")
+    cur.execute(f"""
+        SELECT status, COUNT(*) FROM {SCHEMA}.plot_threads GROUP BY status ORDER BY status
+    """)
+    rows = cur.fetchall()
+    for r in rows:
+        print(f"- {r[0]}: {r[1]} 条")
+    print()
+
+    # 遗忘伏笔检测
+    print("## 遗忘伏笔检测")
+    cur.execute(f"""
+        SELECT f.code, f.description, ch.chapter_number, v.vol_number
+        FROM {SCHEMA}.foreshadowing f
+        JOIN {SCHEMA}.chapters ch ON f.planted_chapter_id = ch.id
+        JOIN {SCHEMA}.volumes v ON ch.volume_id = v.id
+        WHERE f.status = 'planted'
+        ORDER BY ch.chapter_number
+    """)
+    rows = cur.fetchall()
+    # 获取最大章节号作为当前进度
+    cur.execute(f"SELECT MAX(chapter_number) FROM {SCHEMA}.chapters WHERE synopsis_summary IS NOT NULL")
+    max_ch = cur.fetchone()[0] or 0
+
+    warn_count = 0
+    crit_count = 0
+    if rows:
+        for r in rows:
+            age = max_ch - r[2]
+            if age > 50:
+                crit_count += 1
+                print(f"- ❌ [{r[0]}] vol-{r[3]:03d} ch-{r[2]:03d} (已过{age}章): {_truncate(r[1], 60)}")
+            elif age > 30:
+                warn_count += 1
+                print(f"- ⚠️ [{r[0]}] vol-{r[3]:03d} ch-{r[2]:03d} (已过{age}章): {_truncate(r[1], 60)}")
+        if warn_count == 0 and crit_count == 0:
+            print("✅ 无超期伏笔")
+        else:
+            print(f"\n合计: ⚠️ 待推进 {warn_count} 条, ❌ 严重遗忘 {crit_count} 条")
+    else:
+        print("无 planted 伏笔")
+    print()
+
+    # 角色活跃度
+    print("## 角色活跃度")
+    cur.execute(f"""
+        SELECT DISTINCT ON (c.name)
+            c.name, c.role_type, cs.last_appearance, v.vol_number, c.status
+        FROM {SCHEMA}.characters c
+        JOIN {SCHEMA}.character_states cs ON c.id = cs.character_id
+        JOIN {SCHEMA}.volumes v ON cs.volume_id = v.id
+        ORDER BY c.name, v.vol_number DESC, cs.last_appearance DESC
+    """)
+    rows = cur.fetchall()
+    if rows:
+        print("| 角色 | 类型 | 最后出场 | 状态 |")
+        print("|------|------|---------|------|")
+        for r in sorted(rows, key=lambda x: x[2] or 0, reverse=True):
+            print(f"| {r[0]} | {r[1]} | vol-{r[3]:03d} ch-{r[2] or '?'} | {r[4]} |")
+    else:
+        print("(无角色数据)")
+    print()
+
+    # 主角成长
+    print("## 主角成长")
+    cur.execute(f"""
+        SELECT COUNT(*) FILTER (WHERE status = 'active'),
+               COUNT(*) FILTER (WHERE status = 'sealed'),
+               COUNT(*)
+        FROM {SCHEMA}.protagonist_skills
+    """)
+    sk = cur.fetchone()
+    if sk:
+        print(f"- 技能: {sk[2]} 项 (活跃 {sk[0]} / 封印 {sk[1]})")
+
+    cur.execute(f"""
+        SELECT level, progress_pct FROM {SCHEMA}.cultivation_curve
+        ORDER BY chapter_number DESC LIMIT 1
+    """)
+    cult = cur.fetchone()
+    if cult:
+        print(f"- 修炼: {cult[0]} ({cult[1]}%)")
+
+    cur.execute(f"""
+        SELECT COUNT(*) FROM {SCHEMA}.current_inventory
+    """)
+    inv = cur.fetchone()
+    if inv:
+        print(f"- 道具: {inv[0]} 件持有")
+    print()
+
+
+def _truncate(text, max_len=80):
+    """截断文本"""
+    if not text:
+        return "-"
+    text = text.replace("\n", " ")
+    if len(text) > max_len:
+        return text[:max_len] + "..."
+    return text
+
+
+def main():
+    parser = argparse.ArgumentParser(description="DB -> 精确上下文生成器")
+    parser.add_argument("--chapter", type=int, help="目标章节号（全局）")
+    parser.add_argument("--mode", choices=["write", "expand", "analyze"], default="write",
+                        help="上下文模式: write/expand/analyze")
+    parser.add_argument("--dashboard", action="store_true", help="输出创作进度仪表盘")
+    args = parser.parse_args()
+
+    if not args.chapter and not args.dashboard:
+        print("用法: python db_context.py --chapter 42 --mode write")
+        print("      python db_context.py --dashboard")
+        sys.exit(1)
+
+    conn = get_conn()
+    cur = conn.cursor()
+
+    try:
+        if args.dashboard:
+            dashboard(cur)
+        elif args.mode == "write":
+            context_write(cur, args.chapter)
+        elif args.mode == "expand":
+            context_expand(cur, args.chapter)
+        elif args.mode == "analyze":
+            context_analyze(cur, args.chapter)
+    finally:
+        cur.close()
+        conn.close()
+
+
+if __name__ == "__main__":
+    main()