#!/bin/bash
# ============================================================
# codebuddy-mem 全功能回归测试
# 测试范围: 基础设施 / 数据库 / MCP API / Chroma / 配置
# 使用方式: bash scripts/regression-test.sh
# 或对话中说 "执行 codebuddy-mem 回归测试"
# ============================================================

set -o pipefail
RED='\033[0;31m'; GREEN='\033[0;32m'; YELLOW='\033[0;33m'; CYAN='\033[0;36m'
BOLD='\033[1m'; NC='\033[0m'
PASS=0; FAIL=0; WARN=0; TOTAL=0
DATA_DIR="${HOME}/.codebuddy-mem"

# ---------- helpers ----------
_ok()   { PASS=$((PASS+1)); echo -e "  ${GREEN}[PASS]${NC} $1"; }
_fail() { FAIL=$((FAIL+1)); echo -e "  ${RED}[FAIL]${NC} $1"; }
_warn() { WARN=$((WARN+1)); echo -e "  ${YELLOW}[WARN]${NC} $1"; }
_header() { echo -e "\n${CYAN}${BOLD}━━━ $1 ━━━${NC}"; }

# 给定命令 + 描述, 退出码 0 则 ok, 否则 fail
check() { TOTAL=$((TOTAL+1)); if eval "$1" &>/dev/null; then _ok "$2"; else _fail "$2"; fi; }

# 数值期望断言
assert_eq() {
  TOTAL=$((TOTAL+1))
  local val; val=$(eval "$1" 2>/dev/null)
  if [ "$val" = "$2" ]; then _ok "$3 (expect=$2, got=$val)"; else _fail "$3 (expect=$2, got=$val)"; fi
}
assert_gt() {
  TOTAL=$((TOTAL+1))
  local val; val=$(eval "$1" 2>/dev/null)
  if [ "$val" -gt "$2" ] 2>/dev/null; then _ok "$3 (got=$val > $2)"; else _fail "$3 (got=$val, need > $2)"; fi
}

# ============================================================
echo -e "${BOLD}codebuddy-mem 全功能回归测试${NC}"
echo "时间: $(date '+%Y-%m-%d %H:%M:%S')"
echo "数据目录: $DATA_DIR"
echo ""
echo -e "${YELLOW}注意: 第 10 层 MCP API 测试需在 CodeBuddy Code 对话中由助手手动执行${NC}"
echo ""

# ============================================================
_header "1. 进程状态"

check "pgrep -f 'codebuddy-mem.*mcp-server' > /dev/null" \
  "MCP 服务进程运行中"
check "test -f $DATA_DIR/worker.pid" \
  "worker.pid 存在"
check "[ -n \"\$(cat $DATA_DIR/worker.pid 2>/dev/null | python3 -c 'import sys,json; print(json.load(sys.stdin)[\"pid\"])' 2>/dev/null)\" ] && kill -0 \$(cat $DATA_DIR/worker.pid 2>/dev/null | python3 -c 'import sys,json; print(json.load(sys.stdin)[\"pid\"])' 2>/dev/null) 2>/dev/null" \
  "Worker 进程运行中"
check "pgrep -f 'chroma-mcp.*codebuddy-mem' > /dev/null" \
  "Chroma-MCP 进程运行中"

# ============================================================
_header "2. 文件结构"

check "test -f $DATA_DIR/codebuddy-mem.db" \
  "主数据库 codebuddy-mem.db 存在"
check "test -L $DATA_DIR/claude-mem.db" \
  "符号链接 claude-mem.db 存在"
check "[ \"\$(readlink $DATA_DIR/claude-mem.db)\" = 'codebuddy-mem.db' ]" \
  "符号链接指向 codebuddy-mem.db"
check "test -f $DATA_DIR/codebuddy-mem.db.backup" \
  "数据库备份存在"
check "test -f $DATA_DIR/chroma/chroma.sqlite3" \
  "Chroma 向量库存在"
check "test -d $DATA_DIR/chroma" \
  "Chroma 目录存在"
check "test -f $DATA_DIR/chroma/data_level0.bin" \
  "Chroma HNSW 数据文件存在"
check "test -f $DATA_DIR/chroma/header.bin" \
  "Chroma HNSW 头文件存在"

# ============================================================
_header "3. 数据库架构"

for table in sdk_sessions observations user_prompts session_summaries pending_messages schema_versions; do
  check "sqlite3 $DATA_DIR/claude-mem.db \"SELECT name FROM sqlite_master WHERE type='table' AND name='$table';\" | grep -q $table" \
    "表 $table 存在"
done

# FTS5 虚拟表在 sqlite_master 中 type='table'，用 table 名存在性即可
for fts_table in observations_fts session_summaries_fts user_prompts_fts; do
  check "sqlite3 $DATA_DIR/claude-mem.db \"SELECT name FROM sqlite_master WHERE name='$fts_table';\" | grep -q $fts_table" \
    "FTS 虚拟表 $fts_table 存在"
done

# ============================================================
_header "4. 数据统计"

assert_gt "sqlite3 $DATA_DIR/claude-mem.db 'SELECT count(*) FROM observations;'" 0 \
  "observations 有数据"
assert_gt "sqlite3 $DATA_DIR/claude-mem.db 'SELECT count(*) FROM sdk_sessions;'" 0 \
  "sdk_sessions 有数据"
assert_gt "sqlite3 $DATA_DIR/claude-mem.db 'SELECT count(*) FROM user_prompts;'" 0 \
  "user_prompts 有数据"
assert_gt "sqlite3 $DATA_DIR/claude-mem.db 'SELECT count(*) FROM session_summaries;'" 0 \
  "session_summaries 有数据"

# FTS 与主表行数一致
check "test \"\$(sqlite3 $DATA_DIR/claude-mem.db 'SELECT count(*) FROM observations;')\" -eq \"\$(sqlite3 $DATA_DIR/claude-mem.db 'SELECT count(*) FROM observations_fts;')\"" \
  "observations 与 observations_fts 行数一致"

# ============================================================
_header "5. 数据完整性"

# 必须有至少一个项目的数据
check "sqlite3 $DATA_DIR/claude-mem.db \"SELECT count(DISTINCT project) FROM observations;\" | grep -q '[1-9]'" \
  "observations 包含多个项目"

# 必须有常见类型
for otype in discovery change decision; do
  check "sqlite3 $DATA_DIR/claude-mem.db \"SELECT count(*) FROM observations WHERE type='$otype';\" | grep -q '[1-9]'" \
    "observations 包含类型 $otype"
done

# 外键完整性: 孤儿 observation 比例 < 20% (迁移残留正常，严重才告警)
ORPHAN=$(sqlite3 $DATA_DIR/claude-mem.db "SELECT count(*) FROM observations o LEFT JOIN sdk_sessions s ON o.memory_session_id = s.memory_session_id WHERE s.memory_session_id IS NULL;" 2>/dev/null)
TOTAL_OBS=$(sqlite3 $DATA_DIR/claude-mem.db "SELECT count(*) FROM observations;" 2>/dev/null)
if [ -n "$ORPHAN" ] && [ -n "$TOTAL_OBS" ] && [ "$TOTAL_OBS" -gt 0 ]; then
  if [ "$ORPHAN" -eq 0 ]; then
    _ok "observations 外键完整性 (0 条孤儿)"
  elif [ "$ORPHAN" -lt "$((TOTAL_OBS / 5))" ]; then
    _ok "observations 外键完整性 ($ORPHAN/$TOTAL_OBS 孤儿 < 20%, 迁移残留正常)"
  else
    _fail "observations 外键完整性 ($ORPHAN/$TOTAL_OBS 孤儿 >= 20%)"
  fi
fi

# ============================================================
_header "6. Chroma 同步状态"

check "test -f $DATA_DIR/chroma-sync-state.json" \
  "chroma-sync-state.json 存在"

# sync state 中的项目与 observations 中的项目一致
check "python3 -c \"
import json
with open('$DATA_DIR/chroma-sync-state.json') as f:
    state = json.load(f)
# 检查主要项目都有 >0 的 observations watermark
for proj in ['观星阁','筑基阁','mac']:
    if proj not in state:
        raise Exception(f'{proj} not in sync state')
    if state[proj].get('observations',0) == 0:
        raise Exception(f'{proj} observations watermark is 0')
print('OK')
\"" \
  "主要项目同步水位 > 0"

# Chroma 向量索引有数据
check "test -f $DATA_DIR/chroma/chroma.sqlite3 && sqlite3 $DATA_DIR/chroma/chroma.sqlite3 'SELECT count(*) FROM embeddings;' | grep -q '[1-9]'" \
  "Chroma embeddings 有数据"

check "sqlite3 $DATA_DIR/chroma/chroma.sqlite3 'SELECT count(*) FROM collections;' | grep -q '[1-9]'" \
  "Chroma collections 有数据"

# ============================================================
_header "7. 配置文件"

check "test -f $DATA_DIR/settings.json && python3 -c 'import json; json.load(open(\"$DATA_DIR/settings.json\"))' 2>/dev/null" \
  "settings.json 格式正确"
check "test -f $DATA_DIR/supervisor.json && python3 -c 'import json; json.load(open(\"$DATA_DIR/supervisor.json\"))' 2>/dev/null" \
  "supervisor.json 格式正确"

# 必要配置项
check "python3 -c \"
import json
with open('$DATA_DIR/settings.json') as f:
    s = json.load(f)
assert 'CODEBUDDY_MEM_DATA_DIR' in s
assert 'CODEBUDDY_MEM_WORKER_PORT' in s
assert 'CODEBUDDY_MEM_PROVIDER' in s
print('OK')
\"" \
  "settings.json 包含必要配置项"

# ============================================================
_header "8. 日志健康"

check "test -d $DATA_DIR/logs" \
  "日志目录存在"
check "ls $DATA_DIR/logs/*.log 2>/dev/null | head -1 | grep -q log" \
  "存在日志文件"

# 最近日志中无严重错误 (ERROR 但排除已知无害的)
LATEST_LOG=$(ls -t $DATA_DIR/logs/*.log 2>/dev/null | head -1)
if [ -n "$LATEST_LOG" ]; then
  ERRORS=$(grep -c '\[ERROR\]' "$LATEST_LOG" 2>/dev/null || echo 0)
  if [ "$ERRORS" -eq 0 ]; then
    _ok "最新日志无 ERROR ($LATEST_LOG)"
  else
    # 检查是否是"无害"错误 (insufficient disk / pollution cleanup)
    HARMFUL=$(grep '\[ERROR\]' "$LATEST_LOG" | grep -vc 'Insufficient disk\|pollution cleanup' || echo 0)
    if [ "$HARMFUL" -eq 0 ]; then
      _ok "最新日志 $ERRORS 条 ERROR 均为已知无害警告"
    else
      _warn "最新日志有 $HARMFUL 条需要关注的 ERROR"
      echo "       $(grep '\[ERROR\]' "$LATEST_LOG" | grep -v 'Insufficient disk\|pollution cleanup' | tail -3)"
    fi
  fi
fi

# ============================================================
_header "9. HTTP 端口可用性"

PORT=$(python3 -c "import json; print(json.load(open('$DATA_DIR/settings.json'))['CODEBUDDY_MEM_WORKER_PORT'])" 2>/dev/null)
if [ -n "$PORT" ]; then
  check "curl -s -o /dev/null -w '%{http_code}' http://127.0.0.1:$PORT/health 2>/dev/null | grep -q '200'" \
    "Worker HTTP 健康检查 (port $PORT) 返回 200"
fi

# ============================================================
_header "10. Hook 配置与连通性"

CB_SETTINGS="${HOME}/.codebuddy/settings.json"
SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]:-$0}")" && pwd)"
REPO_ROOT="$(cd "$SCRIPT_DIR/.." && pwd)"

# 10.1 settings.json hook 事件完整性
check "python3 -c \"
import json
with open('${CB_SETTINGS}') as f: s = json.load(f)
hooks = s.get('hooks', {})
required = ['PreToolUse', 'PostToolUse', 'UserPromptSubmit', 'Stop', 'SessionStart']
for h in required:
    assert h in hooks, f'{h} missing'
assert hooks['SessionStart'][0].get('matcher') == 'startup|clear|compact'
print('OK')
\"" \
  "settings.json 包含全部 5 类 hook 事件"

# 10.2 hook 命令包含数据目录环境变量
check "python3 -c \"
import json
with open('${CB_SETTINGS}') as f: s = json.load(f)
cmds = []
for event in ['PreToolUse','PostToolUse','UserPromptSubmit','Stop','SessionStart']:
    for g in s['hooks'].get(event,[]):
        for h in g.get('hooks',[]):
            cmds.append(h['command'])
bun_cmds = [c for c in cmds if 'bun-runner.js' in c]
assert len(bun_cmds) >= 6, f'expected >=6 bun-runner hooks, got {len(bun_cmds)}'
for c in bun_cmds:
    assert 'CLAUDE_MEM_DATA_DIR' in c, f'missing CLAUDE_MEM_DATA_DIR'
    assert 'CODEBUDDY_MEM_DATA_DIR' in c, f'missing CODEBUDDY_MEM_DATA_DIR'
print(f'{len(bun_cmds)} hook commands OK')
\"" \
  "settings.json hook 命令全部包含 DATA_DIR 环境变量"

# 10.3 仓库 hooks.json 存在且格式正确
check "test -f ${REPO_ROOT}/hooks/hooks.json && python3 -c 'import json; json.load(open(\"${REPO_ROOT}/hooks/hooks.json\"))' 2>/dev/null" \
  "仓库 hooks.json 存在且格式正确"

# 10.4 仓库 hooks.json 也包含 DATA_DIR
check "python3 -c \"
import json
with open('${REPO_ROOT}/hooks/hooks.json') as f: s = json.load(f)
cmds = []
for event in ['PreToolUse','PostToolUse','UserPromptSubmit','Stop','SessionStart']:
    for g in s['hooks'].get(event,[]):
        for h in g.get('hooks',[]):
            cmds.append(h['command'])
bun_cmds = [c for c in cmds if 'bun-runner.js' in c]
assert len(bun_cmds) >= 6
for c in bun_cmds:
    assert 'CLAUDE_MEM_DATA_DIR' in c
print(f'{len(bun_cmds)} hook commands OK')
\"" \
  "仓库 hooks.json 所有 hook 命令包含 DATA_DIR"

# 10.5 无残留 CAPTURE_BROKEN
CAPTURE_BROKEN="${DATA_DIR}/CAPTURE_BROKEN"
if [ -f "$CAPTURE_BROKEN" ]; then
  _warn "CAPTURE_BROKEN 文件存在，表明近期 hook 执行异常"
  echo "       $(head -3 $CAPTURE_BROKEN)"
else
  _ok "无 CAPTURE_BROKEN 残留"
fi

# 10.6 Hook 连通性: 执行一次 session-init
check "export CLAUDE_MEM_DATA_DIR='${DATA_DIR}' CODEBUDDY_MEM_DATA_DIR='${DATA_DIR}' PATH=\"\$($SHELL -lc 'echo \$PATH' 2>/dev/null):\$PATH\" _R='${REPO_ROOT}' && node \"\$_R/scripts/bun-runner.js\" \"\$_R/scripts/worker-service.cjs\" hook codebuddy session-init 2>&1 | grep -q 'bun-runner'; rm -f '${DATA_DIR}/CAPTURE_BROKEN'" \
  "Hook 链路连通 (bun-runner → worker-service)"

# ============================================================
# 汇总
# ============================================================
echo -e "\n${CYAN}${BOLD}━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━${NC}"
echo -e "${BOLD}测试结果汇总${NC}"
echo -e "  通过: ${GREEN}$PASS${NC}"
echo -e "  失败: ${RED}$FAIL${NC}"
echo -e "  警告: ${YELLOW}$WARN${NC}"
echo -e "  总计: $TOTAL"
echo ""

if [ $FAIL -gt 0 ]; then
  echo -e "${RED}${BOLD}[不通过] 存在 $FAIL 个失败项，请排查后重试。${NC}"
  exit 1
elif [ $WARN -gt 0 ]; then
  echo -e "${YELLOW}${BOLD}[通过(有告警)] $WARN 个告警项，建议关注。${NC}"
  exit 0
else
  echo -e "${GREEN}${BOLD}[全通过] 所有 $PASS 项检查通过。${NC}"
  exit 0
fi

# ============================================================
# 第 11 层: MCP API 测试 (由 CodeBuddy Code 助手执行)
# 以下测试无法通过 bash 脚本完成，需对话中执行:
#
#   mcp__codebuddy-mem__search      { query:"观星阁", limit:5 }
#   mcp__codebuddy-mem__get_observations { ids:[1,200,438] }
#   mcp__codebuddy-mem__timeline    { query:"观星阁 Git", limit:3 }
#   mcp__codebuddy-mem__list_corpora {}
#   mcp__codebuddy-mem__smart_search { query:"function check" }
#   mcp__codebuddy-mem__smart_outline { file:"scripts/regression-test.sh" }
#
# 验证点:
#   1. search 返回结果数 > 0，结果包含 title/type/created_at
#   2. get_observations 返回完整字段: title,facts,narrative,project,type
#   3. timeline 返回带 Anchor 的上下文时间线
#   4. list_corpora 能正常调用 (corpora 可为空)
#   5. smart_search 能在代码库中搜索
#   6. smart_outline 能解析文件结构
# ============================================================