weflow-cli 1.15.1 → 1.17.0

package/scripts/imported_chats.py

@@ -0,0 +1,68 @@
+# -*- coding: utf-8 -*-
+"""读 `weflow-cli import qchat` 落下的存储（output/imports/qq/sessions/）。
+
+**单一读者**：chat_notes 与 semantic_search 都从这里取导入会话。两边各解析一遍格式，
+漂了就是"知识卡看得见、搜索看不见"这类静默分叉——`nt_common` 存在的同款理由。
+
+存储形状（写入方是 `src/services/qchatImport.ts`，字段语义对齐 weflow-message/v1）：
+每个会话一个 JSON：schema/session{id,name,type}/importedAt/sourceFiles/messages[]，
+消息字段 serverId(=qchat messageId，全局唯一)/createTime(unix 秒)/isSend(可缺省=
+未知)/senderUsername(QQ 号，敏感)/senderName/messageType(13 稳定名)/parsedContent/
+rawContent。
+
+路径：默认 `output/imports/qq/sessions`（cwd 相对，与 chat_notes 的 OUTPUT_ROOT 同一
+约定）；`WEFLOW_IMPORT_ROOT` 指到别处时用 `<root>/sessions`（测试用，TS 写入方读同
+一个变量）。
+"""
+import json
+import os
+from pathlib import Path
+
+SCHEMA = 'weflow-import/qq-v1'
+
+
+def sessions_dir(root=None):
+    if root is None:
+        root = os.environ.get('WEFLOW_IMPORT_ROOT')
+    if root:
+        return Path(root) / 'sessions'
+    return Path('output/imports/qq/sessions')
+
+
+def sessions(root=None):
+    """全部已导入会话（按导入时间新→旧）。坏文件/非本 schema 的文件跳过不挡路。
+
+    返回的是存储原文（dict），消息已按 createTime 升序。
+    """
+    d = sessions_dir(root)
+    if not d.is_dir():
+        return []
+    out = []
+    for p in sorted(d.glob('*.json')):
+        try:
+            data = json.loads(p.read_text(encoding='utf-8'))
+        except Exception:
+            continue
+        if not isinstance(data, dict):
+            continue
+        if data.get('schema') != SCHEMA or not isinstance(data.get('messages'), list):
+            continue
+        out.append(data)
+    out.sort(key=lambda s: s.get('importedAt') or '', reverse=True)
+    return out
+
+
+def session_messages(sess, days=None, now=None):
+    """一个会话的消息，可选按"最近 days 天"过滤。
+
+    时间语义与 chat_notes/semantic_search 的窗口一致：unix 秒、含当天。
+    """
+    from datetime import datetime, timedelta, timezone
+
+    msgs = [m for m in sess.get('messages', []) if isinstance(m, dict)]
+    if days is None:
+        return msgs
+    tz = timezone(timedelta(hours=8))
+    now = now or datetime.now(tz)
+    cutoff = int((now - timedelta(days=days)).timestamp())
+    return [m for m in msgs if (m.get('createTime') or 0) >= cutoff]

package/scripts/install_tasks.ps1

@@ -1,91 +1,105 @@
-# Register / update the two scheduled tasks for the official-account daily report.
-# Idempotent: running it again overwrites the definitions with the same values.
-#
-# Why this cannot always be done for you: writing to Task Scheduler is refused in a
-# restricted shell / agent sandbox, and refused as a *class* - measured on 2026-10-19 with
-# an elevated Administrator token, all three write paths said "Access is denied" while
-# reads kept working:
-#     schtasks /create ...                      -> Access is denied
-#     (New-Object -ComObject Schedule.Service)   -> RegisterTaskDefinition 0x80070005
-#     New-ScheduledTaskAction / -SettingsSet ... -> 拒绝访问  (even the in-memory builders)
-# So "I can see the tasks but cannot create them" is not an illusion, and it is not fixed
-# by "run as administrator". The one reliable route is a terminal the sandbox does not
-# supervise: double-click scripts\install_tasks.cmd (it asks for elevation itself), or run
-# this file from your own PowerShell window.
-#
-# Usage:
-#   powershell -NoProfile -ExecutionPolicy Bypass -File <repo>\scripts\install_tasks.ps1
-#   powershell -NoProfile -ExecutionPolicy Bypass -File <repo>\scripts\install_tasks.ps1 -Uninstall
-#
-# The two tasks:
-#   WeFlow Daily       00:05 + at logon   -> YESTERDAY (the last complete day)
-#   WeFlow Daily Noon  13:00              -> TODAY so far (half a day of pushes)
-#
-# Why 00:05 the next day instead of 24:00 tonight: it is the same instant, but "yesterday"
-# then lines up with the date directory the run writes into, and it does not collide with
-# other jobs on the hour. Official accounts keep pushing until 24:00, so this is the only
-# moment that has the complete day.
-#
-# This file is intentionally ASCII-only: Windows PowerShell 5.1 reads a BOM-less file as
-# the ANSI codepage, so non-ASCII text here turns into mojibake *and* a parse error.
-param(
-    [switch]$Uninstall
-)
-
-$ErrorActionPreference = 'Stop'
-
-$repo = Split-Path -Parent $PSScriptRoot
-$script = Join-Path $repo 'scripts\daily_task.ps1'
-if (-not (Test-Path $script)) { throw "not found: $script (this script belongs in the repo's scripts\ folder)" }
-
-$isAdmin = (New-Object Security.Principal.WindowsPrincipal(
-    [Security.Principal.WindowsIdentity]::GetCurrent())
-).IsInRole([Security.Principal.WindowsBuiltInRole]::Administrator)
-if (-not $isAdmin) {
-    Write-Host 'Need elevation: reopen PowerShell as administrator, then run this again.' -ForegroundColor Yellow
-    exit 1
-}
-
-if ($Uninstall) {
-    foreach ($name in @('WeFlow Daily Noon', 'WeFlow Daily')) {
-        if (Get-ScheduledTask -TaskName $name -ErrorAction SilentlyContinue) {
-            Unregister-ScheduledTask -TaskName $name -Confirm:$false
-            Write-Host "removed: $name"
-        }
-    }
-    exit 0
-}
-
-$settings = New-ScheduledTaskSettingsSet -StartWhenAvailable -MultipleInstances IgnoreNew
-$psExe = 'powershell.exe'
-$common = '-NoProfile -ExecutionPolicy Bypass -File ' + $script
-
-# End of day: yesterday. The at-logon trigger is the catch-up: this machine slept for
-# 37 hours and StartWhenAvailable did NOT actually start the missed run (NumberOfMissedRuns=2,
-# no run record), so "it will catch up by itself" is not something to rely on.
-Register-ScheduledTask -TaskName 'WeFlow Daily' `
-    -Action (New-ScheduledTaskAction -Execute $psExe -Argument $common) `
-    -Trigger @((New-ScheduledTaskTrigger -Daily -At '00:05'),
-               (New-ScheduledTaskTrigger -AtLogOn)) `
-    -Settings $settings `
-    -Description 'WeChat daily report for yesterday (00:05 + at logon, idempotent)' -Force | Out-Null
-
-# Noon: today so far. Deliberately NO at-logon trigger: a morning logon would produce a
-# very thin "noon" report and write the .noon_done marker, which would then make the real
-# 13:00 run skip itself (the guard keys on that marker).
-Register-ScheduledTask -TaskName 'WeFlow Daily Noon' `
-    -Action (New-ScheduledTaskAction -Execute $psExe -Argument ($common + ' noon')) `
-    -Trigger (New-ScheduledTaskTrigger -Daily -At '13:00') `
-    -Settings $settings `
-    -Description 'WeChat daily report for today so far (13:00)' -Force | Out-Null
-
-foreach ($name in @('WeFlow Daily', 'WeFlow Daily Noon')) {
-    $t = Get-ScheduledTask -TaskName $name
-    $i = Get-ScheduledTaskInfo -TaskName $name
-    Write-Host ("{0,-20} {1,-8} next {2}" -f $t.TaskName, $t.State, $i.NextRunTime) -ForegroundColor Green
-}
-Write-Host ''
-Write-Host 'Run one now instead of waiting:'
-Write-Host '  Start-ScheduledTask -TaskName "WeFlow Daily Noon"   # today so far'
-Write-Host '  Start-ScheduledTask -TaskName "WeFlow Daily"        # yesterday, complete'
-Write-Host 'Logs: <repo>\logs\daily-<date>[-noon].log'
+# Register / update the two scheduled tasks for the official-account daily report.
+# Idempotent: running it again overwrites the definitions with the same values.
+#
+# Why this cannot always be done for you: writing to Task Scheduler is refused in a
+# restricted shell / agent sandbox, and refused as a *class* - measured on 2026-10-19 with
+# an elevated Administrator token, all three write paths said "Access is denied" while
+# reads kept working:
+#     schtasks /create ...                      -> Access is denied
+#     (New-Object -ComObject Schedule.Service)   -> RegisterTaskDefinition 0x80070005
+#     New-ScheduledTaskAction / -SettingsSet ... -> 拒绝访问  (even the in-memory builders)
+# So "I can see the tasks but cannot create them" is not an illusion, and it is not fixed
+# by "run as administrator". The one reliable route is a terminal the sandbox does not
+# supervise: double-click scripts\install_tasks.cmd (it asks for elevation itself), or run
+# this file from your own PowerShell window.
+#
+# Usage:
+#   powershell -NoProfile -ExecutionPolicy Bypass -File <repo>\scripts\install_tasks.ps1
+#   powershell -NoProfile -ExecutionPolicy Bypass -File <repo>\scripts\install_tasks.ps1 -Uninstall
+#
+# The two tasks:
+#   WeFlow Daily       00:05 + at logon   -> YESTERDAY (the last complete day)
+#   WeFlow Daily Noon  13:00              -> TODAY so far (half a day of pushes)
+#
+# Why 00:05 the next day instead of 24:00 tonight: it is the same instant, but "yesterday"
+# then lines up with the date directory the run writes into, and it does not collide with
+# other jobs on the hour. Official accounts keep pushing until 24:00, so this is the only
+# moment that has the complete day.
+#
+# This file is intentionally ASCII-only: Windows PowerShell 5.1 reads a BOM-less file as
+# the ANSI codepage, so non-ASCII text here turns into mojibake *and* a parse error.
+param(
+    [switch]$Uninstall
+)
+
+$ErrorActionPreference = 'Stop'
+
+$repo = Split-Path -Parent $PSScriptRoot
+$script = Join-Path $repo 'scripts\daily_task.ps1'
+if (-not (Test-Path $script)) { throw "not found: $script (this script belongs in the repo's scripts\ folder)" }
+
+$isAdmin = (New-Object Security.Principal.WindowsPrincipal(
+    [Security.Principal.WindowsIdentity]::GetCurrent())
+).IsInRole([Security.Principal.WindowsBuiltInRole]::Administrator)
+if (-not $isAdmin) {
+    Write-Host 'Need elevation: reopen PowerShell as administrator, then run this again.' -ForegroundColor Yellow
+    exit 1
+}
+
+if ($Uninstall) {
+    foreach ($name in @('WeFlow Daily Noon', 'WeFlow Daily')) {
+        if (Get-ScheduledTask -TaskName $name -ErrorAction SilentlyContinue) {
+            Unregister-ScheduledTask -TaskName $name -Confirm:$false
+            Write-Host "removed: $name"
+        }
+    }
+    exit 0
+}
+
+$settings = New-ScheduledTaskSettingsSet -StartWhenAvailable -MultipleInstances IgnoreNew
+$psExe = 'powershell.exe'
+$common = '-NoProfile -ExecutionPolicy Bypass -File ' + $script
+
+# Launch through wscript.exe, not powershell.exe directly.
+#
+# A scheduled action that runs powershell.exe is a *console* program: Windows creates the console
+# window before PowerShell can hide it, so every run flashes a black box - and while a game is in
+# the foreground that is a real interruption. wscript.exe is a GUI process with no console at all,
+# and run_hidden.vbs starts PowerShell with window style 0, so nothing appears.
+$wscript = Join-Path $env:SystemRoot 'System32\wscript.exe'
+$vbs = Join-Path $PSScriptRoot 'run_hidden.vbs'
+if (-not (Test-Path $vbs)) {
+    Write-Host "Missing launcher: $vbs" -ForegroundColor Red
+    exit 1
+}
+$action = '//B //Nologo "' + $vbs + '"'
+
+# End of day: yesterday. The at-logon trigger is the catch-up: this machine slept for
+# 37 hours and StartWhenAvailable did NOT actually start the missed run (NumberOfMissedRuns=2,
+# no run record), so "it will catch up by itself" is not something to rely on.
+Register-ScheduledTask -TaskName 'WeFlow Daily' `
+    -Action (New-ScheduledTaskAction -Execute $wscript -Argument $action) `
+    -Trigger @((New-ScheduledTaskTrigger -Daily -At '00:05'),
+               (New-ScheduledTaskTrigger -AtLogOn)) `
+    -Settings $settings `
+    -Description 'WeChat daily report for yesterday (00:05 + at logon, idempotent, no window)' -Force | Out-Null
+
+# Noon: today so far. Deliberately NO at-logon trigger: a morning logon would produce a
+# very thin "noon" report and write the .noon_done marker, which would then make the real
+# 13:00 run skip itself (the guard keys on that marker).
+Register-ScheduledTask -TaskName 'WeFlow Daily Noon' `
+    -Action (New-ScheduledTaskAction -Execute $wscript -Argument ($action + ' noon')) `
+    -Trigger (New-ScheduledTaskTrigger -Daily -At '13:00') `
+    -Settings $settings `
+    -Description 'WeChat daily report for today so far (13:00, no window)' -Force | Out-Null
+
+foreach ($name in @('WeFlow Daily', 'WeFlow Daily Noon')) {
+    $t = Get-ScheduledTask -TaskName $name
+    $i = Get-ScheduledTaskInfo -TaskName $name
+    Write-Host ("{0,-20} {1,-8} next {2}" -f $t.TaskName, $t.State, $i.NextRunTime) -ForegroundColor Green
+}
+Write-Host ''
+Write-Host 'Run one now instead of waiting:'
+Write-Host '  Start-ScheduledTask -TaskName "WeFlow Daily Noon"   # today so far'
+Write-Host '  Start-ScheduledTask -TaskName "WeFlow Daily"        # yesterday, complete'
+Write-Host 'Logs: <repo>\logs\daily-<date>[-noon].log'

package/scripts/leads.py

@@ -0,0 +1,179 @@
+"""线索雷达：把聊天里"可能是生意"的对话挑出来，并指回原话。
+
+为什么它值得存在：微信里最容易丢的不是重要消息，是**混在闲聊里的机会** —— 有人问价、
+有人问能不能做、有人留了联系方式，然后被下一条消息压过去。它和 `promises` 用同一个骨架：
+规则先捞候选、再到同会话的后续对话里找"有没有跟上"，**每条结论都带原话**，不做黑箱判断。
+
+默认不调模型。信号分五类（联系方式 / 报价 / 采购 / 合作 / 需求），按命中类型加权排序，
+高分的往往是"问了价 + 留了联系方式"这种最该马上回的组合。
+
+用法：
+    python scripts/leads.py --days 30
+    python scripts/leads.py --days 90 --min-score 3
+    python scripts/leads.py --days 30 --json
+"""
+from __future__ import annotations
+
+import argparse
+import json
+import pathlib
+import re
+import sys
+from collections import defaultdict
+from datetime import datetime, timezone, timedelta
+
+SCRIPT_DIR = pathlib.Path(__file__).resolve().parent
+sys.path.insert(0, str(SCRIPT_DIR))
+
+from _utils import load_config, decrypt_lock  # noqa: E402
+from promises import load_messages, same_account  # noqa: E402
+import nt_decrypt as nd  # noqa: E402
+
+TZ = timezone(timedelta(hours=8))
+
+#: 信号类型 → (正则, 权重)。权重表示"这条信号有多像生意"，不是"多重要"。
+SIGNALS = {
+    '联系方式': (re.compile(r'(微信|vx|wx|威信|手机号|电话|联系我|加我)|1[3-9]\d{9}'
+                          r'|[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}', re.I), 3),
+    '报价': (re.compile(r'(多少钱|报价|预算|费用|价格|单价|收费|怎么算钱|贵不贵)'), 3),
+    '采购': (re.compile(r'(采购|下单|合同|发票|订购|样品|打款|付款)'), 2),
+    '合作': (re.compile(r'(合作|对接|渠道|代理|分销|外包|接单|项目合作)'), 2),
+    '需求': (re.compile(r'(需要.{0,8}(做|开发|方案|系统|服务|设计)|能不能做|可以做吗'
+                      r'|有没有.{0,4}(做|卖|提供)|求推荐|想找.{0,6}(做|人))'), 2),
+}
+#: 我跟上了的迹象（在同一条线索之后的同会话消息里找）
+FOLLOWED = re.compile(r'(已加|加了|私聊|发你了|发您了|发过去|安排|稍后|好的.{0,4}(我|马上)|收到)')
+#: 已经明确黄掉的迹象
+DEAD = re.compile(r'(太贵|不考虑|算了|暂时不需要|下次吧|再看看)')
+
+MERGE_WINDOW = 3600      # 同一会话里这一秒数内的信号合成一条线索
+
+
+#: 广告 / 招聘 / 促销：这些恰好塞满"合作、采购、报价、联系方式"这些词，但**不是线索**。
+#: 2026-10-23 第一版没滤它们，30 天里 8 条"线索"全是校招宣讲和电商促销，没有一条真生意。
+NOISE = re.compile(
+    r'(招聘|校招|秋招|春招|宣讲会|双选会|简历|offer|求职|面试|直播|预告'
+    r'|优惠|券|折扣|秒杀|包邮|抽奖|福利|返现|下单链接|u\.jd\.com|拼团|预售|凑单)', re.I)
+#: 真线索几乎都带**请求或提问**。纯陈述句（广告、通知）不进来。
+ASK = re.compile(r'(吗|呢|？|\?|能否|可以吗|可不可以|麻烦|请问|想找|有没有|多少钱|怎么收费|求推荐)')
+
+
+def _score(text):
+    """(信号类型, 分数)。**先过噪点与提问两道闸，再算信号分。**
+
+    匹配前先剥掉 `@昵称`：群昵称里常带手机号（`@某人13800000000`），不剥的话每一句
+    亲戚群闲聊都会被判成"留了联系方式"（2026-10-23 实测，误报全来自这里）。
+    """
+    bare = re.sub(r'@[^\s]{1,40}\s*', '', text)
+    # URL 里的长数字串会命中手机号规则（`?id=4713032079053994316` 这类），先剥链接。
+    bare = re.sub(r'https?://\S+', '', bare)
+    if len(bare) > 300 or NOISE.search(bare) or not ASK.search(bare):
+        return [], 0
+    hit = []
+    score = 0
+    for name, (rx, weight) in SIGNALS.items():
+        if rx.search(bare):
+            hit.append(name)
+            score += weight
+    return hit, score
+
+
+def main():
+    ap = argparse.ArgumentParser(description='线索雷达：聊天里的报价/合作/需求信号')
+    ap.add_argument('--days', type=int, default=30)
+    ap.add_argument('--min-score', type=int, default=2, help='最低分（默认 2）')
+    ap.add_argument('--limit', type=int, default=20)
+    ap.add_argument('--json', action='store_true')
+    args = ap.parse_args()
+
+    cfg = load_config()
+    db = cfg.get('ntDbPath') or ''
+    my_wxid = cfg.get('wxid') or ''
+    if not db:
+        raise SystemExit('配置里没有 ntDbPath')
+    key = decrypt_lock(cfg.get('ntKey', '') or '')
+    salt = decrypt_lock(cfg.get('ntSalt', '') or '')
+    pw = decrypt_lock(cfg.get('favPassphrase', '') or cfg.get('decryptKey', '') or '')
+    conns = nd.connect_message_shards(db, key, salt, pw)
+    if not conns:
+        raise SystemExit('打不开消息分片')
+
+    try:
+        by_talker = load_messages(conns, args.days, my_wxid)
+        name_map = {}
+        try:
+            from network import load_graph
+            people, _rooms, _members = load_graph(cfg)
+            name_map = {p['username']: (p['remark'] or p['nick'] or p['username'])
+                        for p in people.values()}
+        except Exception as exc:
+            print(f'[WARN] 读联系人失败（{type(exc).__name__}），显示 id', file=sys.stderr)
+    finally:
+        for conn in conns:
+            try:
+                conn.close()
+            except Exception:
+                pass
+
+    leads = []
+    for talker, msgs in by_talker.items():
+        hits = []
+        for i, (ts, is_me, who, text) in enumerate(msgs):
+            if is_me:
+                continue                      # 线索来自别人说的，不是我自己发的
+            kinds, score = _score(text)
+            if score < args.min_score:
+                continue
+            hits.append({'i': i, 'ts': ts, 'text': text, 'kinds': kinds, 'score': score})
+
+        # 同一会话里时间挨近的信号合成一条线索，避免"问价"和"留手机号"被拆成两条。
+        merged = []
+        for h in hits:
+            if merged and h['ts'] - merged[-1]['last'] <= MERGE_WINDOW:
+                cur = merged[-1]
+                cur['last'] = h['ts']
+                cur['score'] = max(cur['score'], h['score']) + 1
+                cur['kinds'] = sorted(set(cur['kinds']) | set(h['kinds']))
+                if h['text'] not in cur['parts']:
+                    cur['parts'].append(h['text'])
+                cur['idx'].append(h['i'])
+            else:
+                merged.append({'first': h['ts'], 'last': h['ts'], 'score': h['score'],
+                               'kinds': list(h['kinds']), 'parts': [h['text']],
+                               'idx': [h['i']]})
+
+        for m in merged:
+            after = msgs[m['idx'][-1] + 1: m['idx'][-1] + 25]
+            replied = any(is_me for _, is_me, _w, _t in after)
+            followed = any(is_me and FOLLOWED.search(t) for _, is_me, _w, t in after)
+            dead = any(DEAD.search(t) for _, _me, _w, t in after)
+            state = ('已结束' if dead else '已跟进' if followed else '我回过' if replied else '没回')
+            leads.append({
+                'time': datetime.fromtimestamp(m['first'], tz=TZ).strftime('%Y-%m-%d %H:%M'),
+                'talker': name_map.get(talker, talker),
+                'kinds': m['kinds'],
+                'score': m['score'],
+                'said': ' ／ '.join(p[:80] for p in m['parts'][:3]),
+                'state': state,
+            })
+
+    leads.sort(key=lambda x: (-x['score'], x['time']), reverse=False)
+    leads = leads[:args.limit]
+
+    if args.json:
+        print(json.dumps({'success': True, 'days': args.days, 'leads': leads},
+                         ensure_ascii=False, indent=2))
+        return
+
+    if not leads:
+        print(f'最近 {args.days} 天里没有达到分数 {args.min_score} 的线索信号')
+        return
+    print(f'最近 {args.days} 天的线索信号 {len(leads)} 条（分数越高越像生意）：\n')
+    for it in leads:
+        print(f"[{it['score']}分 · {'/'.join(it['kinds'])} · {it['state']}] {it['time']}  {it['talker']}")
+        print(f"    {it['said']}")
+        print()
+
+
+if __name__ == '__main__':
+    main()