admin/my-assistant
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

docs: add TTS voice reply capability to memory and TOOLS.md

Browse Source
This commit is contained in:
root
2026-03-30 01:36:47 +08:00
parent 84d47fc753
commit 7ce166affb
3 changed files with 233 additions and 85 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
TOOLS.md
View File

@@ -133,6 +133,17 @@ SMTP 服务器: smtp.exmail.qq.com (SSL, 端口465)
**重要提醒人：** `fj@77ircloud.com` 的邮件需要重点关注
## MiniMax TTS 语音回复
**脚本**: `scripts/dingtalk_tts.sh`
**使用方式**: 用户要求语音回复时，执行：
```bash
bash scripts/dingtalk_tts.sh "要说的内容"
```
**注意**: duration 参数使用秒（整数），不是毫秒
## ⚠️ 邮件操作安全规则
**只读不删！** 严禁执行任何删除邮件的操作，包括但不限于：

90
memory/2026-03-29.md Normal file
View File

@@ -0,0 +1,90 @@
# 2026-03-29 Daily Memory
## User Info
- User: 罗国财 (userid: 121922510028034588)
- Role: 产研部门负责人
## MCP Services Configured (4 total)
| Name | URL Key | Purpose |
|------|---------|---------|
| 钉钉文档 | ...?key=70fb6676... | 文档操作 |
| 钉钉MCP | ...?key=1ef173d1... | 日历/日程 |
| 钉钉MCP2 | ...?key=ec174881... | 通讯录/用户 |
| 钉钉待办 | ...?key=7f8ad91c... | 待办管理 |
## DingTalk API Notes
- ⚠️ Time parameters are in **milliseconds**, not seconds
- ⚠️ Field names use lowercase `userid`, not `userId`
- Calendar query: `list_calendar_events` needs millisecond timestamps:
- startTime/endTime in milliseconds
- Example: 2026-03-30 = 1774800000000 to 1774886400000
## DingTalk MCP Mapping Rules
| Feature | Use |
|---------|-----|
| 钉钉文档 | 钉钉文档 MCP |
| 钉钉日程 | 钉钉日程 MCP |
| 钉钉待办 | 钉钉待办 MCP |
| 钉钉通讯录 | 钉钉通讯录 MCP |
| Other DingTalk | Use skills/API |
## Email Config
- Tencent Enterprise Email (IMAP)
- IMAP: imap.exmail.qq.com:993 (SSL)
- SMTP: smtp.exmail.qq.com:465 (SSL)
- Account: lgc@77ircloud.com
- ⚠️ READ ONLY - never delete emails
- Important sender: fj@77ircloud.com
## Email Summary Cron
- Script: scripts/email_summary.py
- Schedule: Daily 9 AM
- Logs: logs/email_summary.log
## Skills Installed Today
- dingtalk-api (from github)
- skill-creator
- pdf
- docx
- canvas-design
- lobster-email
## Work Reports Stored
- File: memory/luoguocai_weekly_reports.md
- 6 weekly reports from 2026-02 to 2026-03 (incomplete due to API 1000-char limit)
## Calendar Tomorrow (2026-03-30)
- 10:00-11:00: 会议 (self-created)
- 11:00-12:00: 产研经理人周会 (陈咏梅)
- 14:00-15:00: 生产健康度review周会 (薛凯杰) - CANCELLED
- 16:45-17:30: 质量管理部周例会 (薛凯杰)
- 17:30-18:00: 前端周会 (刘畅)
- 17:30-18:00: 业务后台周例会 (韦成双)
- 17:30-18:00: 产品及设计周例会 (self)
Note: Multiple meetings overlap at 17:30-18:00
## MiniMax TTS → 钉钉语音
### 脚本
- `scripts/dingtalk_tts.sh` - MiniMax TTS 生成语音并发送到钉钉
### 使用方式
```bash
bash scripts/dingtalk_tts.sh "要说的内容"
```
### 流程
1. MiniMax TTS (hex audio) → Python 解析 → MP3
2. MP3 上传到 oapi.dingtalk.com/media/upload (type=voice)
3. 发送使用 /v1.0/robot/oToMessages/batchSend API
### 关键发现
- DingTalk 语音消息用 MP3 格式即可
- API: POST https://api.dingtalk.com/v1.0/robot/oToMessages/batchSend
- msgKey: sampleAudio
- msgParam: {"mediaId":"xxx","duration":"秒"}
- **重要**: duration 参数使用**秒**（整数），不是毫秒
### ⚠️ 重要规则
**后续回复：如果用户要求语音回复，直接使用 `bash scripts/dingtalk_tts.sh "回复内容"` 发送语音消息**

217
scripts/email_summary.py
View File

@@ -3,6 +3,7 @@
"""
邮件摘要脚本 - 每天早上9点推送前一天的邮件摘要
重点关注 fj@77ircloud.com 的邮件
需要回应的邮件会标记提醒
⚠️ 安全规则：此脚本仅读取邮件，不执行任何删除操作
"""
@@ -12,8 +13,6 @@ import ssl
import email
from email.header import decode_header
from datetime import datetime, timedelta
import json
import sys
def decode_str(s):
if s is None:
@@ -31,32 +30,46 @@ def decode_str(s):
result.append(part)
return ''.join(result)
def get_email_body(msg):
"""获取邮件正文"""
body = ''
if msg.is_multipart():
for part in msg.walk():
content_type = part.get_content_type()
if content_type == 'text/plain':
try:
charset = part.get_content_charset() or 'utf-8'
body = part.get_payload(decode=True).decode(charset, errors='replace')
break
except:
pass
elif content_type == 'text/html':
try:
charset = part.get_content_charset() or 'utf-8'
body = part.get_payload(decode=True).decode(charset, errors='replace')
except:
pass
else:
try:
charset = msg.get_content_charset() or 'utf-8'
body = msg.get_payload(decode=True).decode(charset, errors='replace')
except:
pass
return body[:500] if body else '(无正文)'
def get_body_preview(msg, max_len=200):
"""获取邮件正文预览"""
try:
if msg.is_multipart():
for part in msg.walk():
if part.get_content_type() == 'text/plain':
payload = part.get_payload(decode=True)
if payload:
charset = part.get_content_charset() or 'utf-8'
body = payload.decode(charset, errors='replace')
return body[:max_len].replace('\n', ' ').strip()
else:
payload = msg.get_payload(decode=True)
if payload:
charset = msg.get_content_charset() or 'utf-8'
body = payload.decode(charset, errors='replace')
return body[:max_len].replace('\n', ' ').strip()
except:
pass
return ''
def needs_response(subject, body, sender):
"""判断邮件是否需要回应"""
text = (subject + ' ' + body).lower()
sender_lower = sender.lower()
# 需要回应的关键词
response_keywords = [
'请回复', '请回复我', '需回复', '需要回复',
'请回复确认', '请回复一下', '请回复告知',
'请确认', '请批准', '请审核', '请审批',
'期待回复', '希望回复', '烦请', '麻烦',
'请知悉', '请查收', '请查阅',
'请协助', '请配合', '请支持', '请安排', '请处理',
]
# 重要发件人
important_senders = ['fj@77ircloud.com']
return any(kw in text for kw in response_keywords) or any(s in sender_lower for s in important_senders)
def fetch_yesterday_emails():
"""获取昨天的邮件"""
@@ -70,88 +83,122 @@ def fetch_yesterday_emails():
email_ids = messages[0].split()
yesterday = (datetime.now() - timedelta(days=1)).date()
yesterday_start = yesterday.strftime('%d-%b-%Y')
all_emails = []
important_emails = []
need_response_emails = []
for uid in reversed(email_ids[-50:] if len(email_ids) >= 50 else email_ids):
status, msg_data = mail.fetch(uid, '(RFC822.HEADER RFC822.TEXT)')
if not msg_data or not msg_data[0]:
continue
raw_email = msg_data[0][1] if isinstance(msg_data[0], tuple) else msg_data[0]
if isinstance(raw_email, tuple):
raw_email = raw_email[1]
# 扫描最近50封
scan_range = email_ids[-50:] if len(email_ids) >= 50 else email_ids
for uid in reversed(scan_range):
try:
msg = email.message_from_bytes(raw_email if isinstance(raw_email, bytes) else raw_email.encode('utf-8'))
except:
# 获取完整邮件（头部+正文）
status, msg_data = mail.fetch(uid, '(RFC822)')
if not msg_data or not msg_data[0]:
continue
raw_email = msg_data[0][1]
if isinstance(raw_email, tuple):
raw_email = raw_email[1]
msg = email.message_from_bytes(raw_email)
subject = decode_str(msg.get('Subject', '(无主题)'))
sender = decode_str(msg.get('From', ''))
date_str = msg.get('Date', '')
# 解析日期
try:
t = email.utils.parsedate_tz(date_str)
if t:
local_dt = datetime.fromtimestamp(email.utils.mktime_tz(t))
email_date = local_dt.date()
else:
continue
except:
continue
# 只取昨天的邮件
if email_date != yesterday:
continue
body = get_body_preview(msg, 300)
email_info = {
'subject': subject,
'sender': sender,
'date': date_str,
'body': body,
}
all_emails.append(email_info)
# 重点关注
if 'fj@77ircloud.com' in sender:
important_emails.append(email_info)
# 需要回应
if needs_response(subject, body, sender):
need_response_emails.append(email_info)
except Exception as e:
continue
subject = decode_str(msg.get('Subject', '(无主题)'))
sender = decode_str(msg.get('From', ''))
date_str = msg.get('Date', '')
# 解析日期
try:
email_date = email.utils.parsedate_to_datetime(date_str).date()
except:
continue
# 只取昨天的邮件
if email_date != yesterday:
continue
email_info = {
'subject': subject,
'sender': sender,
'date': date_str,
'body': get_email_body(msg)
}
all_emails.append(email_info)
# 重点关注 fj@77ircloud.com
if 'fj@77ircloud.com' in sender:
important_emails.append(email_info)
mail.logout()
return all_emails, important_emails, yesterday.strftime('%Y年%m月%d日')
return all_emails, important_emails, need_response_emails, yesterday.strftime('%Y年%m月%d日')
except Exception as e:
return [], [], str(e)
return [], [], [], str(e)
def format_summary():
"""格式化邮件摘要"""
emails, important, date_str = fetch_yesterday_emails()
emails, important, need_response, date_str = fetch_yesterday_emails()
if isinstance(important, str):
return f"⚠️ 获取邮件失败: {important}"
if not emails:
return f"📬 {date_str} 无新邮件"
return f"📬 {date_str} 无新邮件\n\n祝你一天好心情！😊"
lines = [f"📬 {date_str} 邮件摘要（共 {len(emails)} 封）\n"]
lines = [f"📬 {date_str} 邮件摘要\n"]
lines.append(f"共收到 **{len(emails)}** 封邮件\n")
if important:
lines.append(f"⭐ 重点关注（fj@77ircloud.com）：{len(important)} 封\n")
for i, mail in enumerate(important, 1):
# 需要回应的邮件（最优先）
if need_response:
lines.append(f"🔔 **需要您回应（{len(need_response)} 封）**\n")
lines.append("─" * 40)
for i, mail in enumerate(need_response, 1):
lines.append(f" {i}. 📧 {mail['subject']}")
lines.append(f" 发件人: {mail['sender']}")
lines.append(f" 时间: {mail['date']}")
if mail['body'] and mail['body'] != '(无正文)':
lines.append(f" 摘要: {mail['body'][:200]}...")
if mail.get('body'):
lines.append(f" 摘要: {mail['body'][:150]}...")
lines.append("")
lines.append(f"\n📋 其他邮件（{len(emails) - len(important)} 封）：")
other = [e for e in emails if e not in important]
for i, mail in enumerate(other[:5], 1):
lines.append(f" {i}. 📧 {mail['subject']} - {mail['sender']} ({mail['date']})")
# 重点关注（排除已标记需要回应的）
important_only = [e for e in important if e not in need_response]
if important_only:
lines.append(f"⭐ **重点关注（fj@77ircloud.com）（{len(important_only)} 封）**\n")
lines.append("─" * 40)
for i, mail in enumerate(important_only, 1):
lines.append(f" {i}. 📧 {mail['subject']}")
lines.append(f" 发件人: {mail['sender']}")
lines.append("")
if len(other) > 5:
lines.append(f" ... 还有 {len(other) - 5} 封邮件")
# 其他邮件
other = [e for e in emails if e not in important and e not in need_response]
if other:
lines.append(f"\n📋 **其他邮件（{len(other)} 封）**\n")
lines.append("─" * 40)
for i, mail in enumerate(other[:8], 1):
lines.append(f" {i}. 📧 {mail['subject']}")
lines.append(f" 发件人: {mail['sender']}")
lines.append("")
if len(other) > 8:
lines.append(f" ... 还有 {len(other) - 8} 封邮件")
lines.append("\n" + "=" * 40)
lines.append("💡 提示：以上标记「需要回应」的邮件，请及时处理！")
return "\n".join(lines)