@aiyiran/myclaw 1.1.182 → 1.1.184

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@aiyiran/myclaw",
-  "version": "1.1.182",
+  "version": "1.1.184",
   "description": "",
   "main": "index.js",
   "bin": {

package/skills/chat-history-extractor/SKILL.md

@@ -16,15 +16,18 @@ description: Extract and render chat history from OpenClaw session URLs. Use whe
 
 ```bash
 # 单个
-python3 scripts/extract_chat.py "<url>" <output-dir>
+python3 scripts/extract_chat.py "<url>" <my-workspace>
 
 # 批量（逗号分隔）
-python3 scripts/extract_chat.py "url1,url2,url3" <output-dir>
+python3 scripts/extract_chat.py "url1,url2,url3" <my-workspace>
 
 # 批量（JSON 数组）
-python3 scripts/extract_chat.py '["url1","url2","url3"]' <output-dir>
+python3 scripts/extract_chat.py '["url1","url2","url3"]' <my-workspace>
 ```
 
+`<my-workspace>` 是**执行命令的 agent 自己的 workspace 路径**（不是被分析的对话所在的 workspace）。
+输出固定到 `<my-workspace>/chat-history/` 子目录。
+
 ## 输出文件
 
 **只产出两个文件**，不再为每个会话单独落 js（所有会话的完整数据都汇入 `index.js`）：
@@ -74,7 +77,7 @@ URL 示例：`https://claw1.kekouen.cn/chat?session=agent%3Ac108-v1-1811%3Amain`
 ### Step 2: 定位 JSONL 文件
 
 ```
-/root/.openclaw/agents/{agentId}/sessions/sessions.json
+~/.openclaw/agents/{agentId}/sessions/sessions.json
 ```
 
 用 session key 在 `sessions.json` 中查找 `sessionFile` 路径。
@@ -104,22 +107,18 @@ URL 示例：`https://claw1.kekouen.cn/chat?session=agent%3Ac108-v1-1811%3Amain`
 
 如果不使用脚本（如当前 agent 直接操作）：
 
-1. **收集**：解析 URL → 查 `sessions.json` → 复制 JSONL 到输出目录
-2. **运行脚本**：`python3 scripts/extract_chat.py "<urls>" <output-dir>`
-3. **渲染**：复制 HTML 模板到输出目录，用浏览器打开
+1. **收集**：解析 URL → 查 `sessions.json` → 确认 JSONL 存在
+2. **运行脚本**：`python3 scripts/extract_chat.py "<urls>" <my-workspace>`
+3. **渲染**：用浏览器打开 `<my-workspace>/chat-history/index.html`
 
 ## 示例
 
 ### 单个会话
 
 ```bash
-python3 scripts/extract_chat.py "https://claw1.kekouen.cn/chat?session=agent%3Ac108-v1-1811%3Amain" ./output
+python3 scripts/extract_chat.py "https://claw1.kekouen.cn/chat?session=agent%3Ac108-v1-1811%3Amain" /root/.openclaw/workspace-teacher
 ```
 
-输出：
-- `index.js`（含 1 个会话）
-- `index.html`
-
 ### 批量处理
 
 ```bash
@@ -128,9 +127,7 @@ python3 scripts/extract_chat.py '[
   "https://claw6.kekouen.cn/chat?session=agent%3Ausa%3Amain",
   "https://claw6.kekouen.cn/chat?session=agent%3Ac109-v3-1813%3A...",
   "https://claw6.kekouen.cn/chat?session=agent%3Ac108-v1-1810%3A..."
-]' ./output
+]' /root/.openclaw/workspace-teacher
 ```
 
-输出：
-- `index.js`（含 4 个会话，全部数据在 `chatAll` 里）
-- `index.html`
+输出在 `<workspace>/chat-history/` 下：`index.js` + `index.html`。

package/skills/chat-history-extractor/scripts/extract_chat.py

@@ -1,47 +1,24 @@
 #!/usr/bin/env python3
 """
 Extract chat history from OpenClaw session URL(s) and generate JS data file(s).
-"""
-import os as _os
-def _get_script_dir():
-    try:
-        return os.path.dirname(os.path.abspath(__file__))
-    except NameError:
-        return os.path.dirname(os.path.abspath(sys.argv[0]))
-
 
-"""
 Usage:
-  # 单个会话（输出到当前 workspace）
-  python3 extract_chat.py <session-url-or-key>
-
-  # 多个会话（逗号分隔 / JSON 数组）
+  python3 extract_chat.py <session-url-or-key> [output-dir]
   python3 extract_chat.py "url1,url2,url3" [output-dir]
-  python3 extract_chat.py '[ "url1","url2","url3"]' [output-dir]
+  python3 extract_chat.py '["url1","url2"]' [output-dir]
+  python3 extract_chat.py --scan [output-dir]
 
-  # 扫描所有 agent 的所有会话，统计表格
-  python3 extract_chat.py --scan [--html]
+输出到 {workspace}/chat-history/ 沙盒目录（或指定的 output-dir）：
+  - index.js：chatIndex + chatAll
+  - index.html：查看器
+  - --scan 模式额外生成 作品扫描_<时间>.html
 
-输出（默认到当前 workspace /root/.openclaw/workspace-teacher）：
-  - index.js：唯一数据产物，含 chatIndex（会话元信息列表，含 work_url 字段）
-    + chatAll（每个会话的完整对话，按 session_id 索引）
-  - index.html：查看器，复制自 assets/chat-history-template.html，只加载 index.js
-  - 不再为每个会话单独落 js 文件
-  - --scan 模式额外生成 作品扫描_<时间>.html（含可点击链接）
-
-Host 检测：
-  - 自动从 /root/.openclaw/myclaw/server/config.json 读取
-  - 可用 --host 覆盖
-
-work_url 字段:
-  - 默认为空字符串
-  - AI 可后续编辑 index.js 中对应会话的 work_url 填入学生作品链接
+Host 自动从 {openclaw_root}/myclaw/server/config.json 读取，可用 --host 覆盖。
 """
 
 import json
 import sys
 import os
-import re
 import glob
 from urllib.parse import urlparse, parse_qs, quote
 from datetime import datetime, timezone, timedelta
@@ -49,10 +26,20 @@ import time
 
 tz_beijing = timezone(timedelta(hours=8))
 
+REPORT_SUBDIR = 'chat-history'
+_base = os.environ.get('OPENCLAW_BASE', os.path.expanduser('~/.openclaw'))
+
+
+def _get_script_dir():
+    try:
+        return os.path.dirname(os.path.abspath(__file__))
+    except NameError:
+        return os.path.dirname(os.path.abspath(sys.argv[0]))
+
 
 def detect_chat_host():
     """Auto-detect chat host from config.json."""
-    config_path = '/root/.openclaw/myclaw/server/config.json'
+    config_path = f'{_base}/myclaw/server/config.json'
     try:
         with open(config_path) as f:
             cfg = json.load(f)
@@ -62,17 +49,6 @@ def detect_chat_host():
         return 'claw5.kekouen.cn'
 
 
-def get_default_output_dir():
-    """Return the current workspace directory."""
-    # OPENCLAW_WORKSPACE is set to the workspace path
-    ws = os.environ.get('OPENCLAW_WORKSPACE', '')
-    if ws and os.path.isdir(ws):
-        return ws
-    # Fallback: derive from cwd (e.g. /root/.openclaw/workspace-teacher)
-    cwd = os.getcwd()
-    if '/workspace-' in cwd:
-        return cwd
-    return cwd
 
 
 def parse_time(ts_str):
@@ -119,7 +95,7 @@ def find_session_file(session_key):
     if len(parts) < 2:
         return None
     agent_id = parts[1]
-    sessions_json_path = f'/root/.openclaw/agents/{agent_id}/sessions/sessions.json'
+    sessions_json_path = f'{_base}/agents/{agent_id}/sessions/sessions.json'
 
     if not os.path.exists(sessions_json_path):
         return None
@@ -130,9 +106,7 @@ def find_session_file(session_key):
     if session_key in sessions:
         session_file = sessions[session_key].get('sessionFile')
         if session_file and not os.path.isabs(session_file):
-            session_file = os.path.join(
-                f'/root/.openclaw/agents/{agent_id}/sessions', session_file
-            )
+            session_file = os.path.join(f'{_base}/agents/{agent_id}/sessions', session_file)
         return session_file
     return None
 
@@ -185,13 +159,13 @@ def count_conversations(jsonl_path):
                     ai_messages.append(messages[j])
                 j += 1
             ai_text = '\n\n'.join([m['text'] for m in ai_messages if m['has_text']])
-            user_time = msg.get('timestamp', '')
-            ai_time = ai_messages[0].get('timestamp', '') if ai_messages else ''
+            user_time = parse_time(msg.get('timestamp', '')) or ''
+            ai_time = parse_time(ai_messages[0].get('timestamp', '')) if ai_messages else ''
             conversations.append({
                 'user': user_text,
                 'ai': ai_text,
-                'user_time': user_time,
-                'ai_time': ai_time,
+                'user_time': user_time or '',
+                'ai_time': ai_time or '',
             })
             i += 1
         else:
@@ -200,30 +174,6 @@ def count_conversations(jsonl_path):
     return conversations, first_timestamp, last_timestamp
 
 
-def parse_jsonl_to_conversations(jsonl_path):
-    """Full parse returning conversation pairs with timestamps."""
-    conversations, first_ts, last_ts = count_conversations(jsonl_path)
-    return conversations, first_ts, last_ts
-
-
-def sanitize_filename(name):
-    safe = re.sub(r'[^\w\u4e00-\u9fff\u3400-\u4dbf-]', '-', name)
-    safe = re.sub(r'-+', '-', safe).strip('-')
-    return safe or 'session'
-
-
-def build_filename(session_key):
-    parts = session_key.split(':')
-    agent_id = parts[1] if len(parts) > 1 else 'unknown'
-    session_name = parts[2] if len(parts) > 2 else 'main'
-    # Use last part (UUID or 'main') as unique suffix to avoid collisions
-    unique_id = parts[-1] if len(parts) > 3 else session_name
-    agent_safe = sanitize_filename(agent_id)
-    name_safe = sanitize_filename(session_name)
-    unique_safe = sanitize_filename(unique_id)
-    if unique_safe == name_safe:
-        return f"{agent_safe}_{name_safe}.js"
-    return f"{agent_safe}_{name_safe}_{unique_safe}.js"
 
 
 def build_chat_data(conversations, session_key, first_ts, last_ts):
@@ -256,18 +206,15 @@ def session_key_to_url(session_key, host=None):
 
 
 def scan_all_agents():
-    """
-    Scan /root/.openclaw/agents/*/sessions/sessions.json
-    Return list of (session_key, agent_id, session_name, jsonl_path) for all sessions.
-    """
+    """Scan {_base}/agents/*/sessions/sessions.json."""
     results = []
-    base = '/root/.openclaw/agents'
-    for agent_dir in sorted(glob.glob(f'{base}/*/sessions/sessions.json')):
+    agents_dir = f'{_base}/agents'
+    for agent_dir in sorted(glob.glob(f'{agents_dir}/*/sessions/sessions.json')):
         agent_id = agent_dir.split('/agents/')[1].split('/sessions/')[0]
         try:
             with open(agent_dir) as f:
                 sessions = json.load(f)
-        except:
+        except Exception:
             continue
         for session_key, meta in sessions.items():
             if not isinstance(meta, dict):
@@ -276,7 +223,7 @@ def scan_all_agents():
             if not session_file:
                 continue
             if not os.path.isabs(session_file):
-                session_file = os.path.join(f'{base}/{agent_id}/sessions', session_file)
+                session_file = os.path.join(agents_dir, agent_id, 'sessions', session_file)
             if not os.path.exists(session_file):
                 continue
             parts = session_key.split(':')
@@ -286,12 +233,9 @@ def scan_all_agents():
 
 
 def print_summary_table(entries, host=None):
+    """Print the summary table to stdout."""
     if host is None:
         host = detect_chat_host()
-    """
-    Print the summary table.
-    entries: list of dicts with keys: index, session_name, agent_id, total_pairs, work_url, session_key, first_time, last_time
-    """
     print(f"\n{'#':>3}  {'workspace':<16} {'会话名':<28} {'对话数':>6}  {'起始时间':<20}  {'最后更新':<20}  {'作品链接':<14}  {'对话链接'}")
     print(f"{'-'*3}  {'-'*16} {'-'*28} {'-'*6}  {'-'*20}  {'-'*20}  {'-'*14}  {'-'*50}")
     for e in entries:
@@ -390,7 +334,7 @@ def process_one(url_or_key, output_dir, index):
         print(f"[{index}] ERROR: JSONL not found: {jsonl_path}")
         return None
 
-    conversations, first_ts, last_ts = parse_jsonl_to_conversations(jsonl_path)
+    conversations, first_ts, last_ts = count_conversations(jsonl_path)
     print(f"[{index}] {len(conversations)} pairs")
 
     data = build_chat_data(conversations, session_key, first_ts, last_ts)
@@ -513,25 +457,34 @@ def main():
 
     if not positional and not scan_mode:
         print("Usage:")
-        print("  python3 extract_chat.py <url-or-key> [output-dir] [--host HOST] [--html]")
-        print("  python3 extract_chat.py 'url1,url2,url3' [output-dir] [--host HOST] [--html]")
-        print('  python3 extract_chat.py \'["url1","url2"]\' [output-dir] [--host HOST] [--html]')
-        print("  python3 extract_chat.py --scan [output-dir] [--host HOST] [--html]")
-        print("Flags:")
+        print("  python3 extract_chat.py <url-or-key> <workspace> [--host HOST]")
+        print("  python3 extract_chat.py 'url1,url2,url3' <workspace> [--host HOST]")
+        print('  python3 extract_chat.py \'["url1","url2"]\' <workspace> [--host HOST]')
+        print("  python3 extract_chat.py --scan <workspace> [--host HOST]")
+        print("\n  <workspace>  执行命令的 agent 自己的 workspace 路径")
+        print("  输出固定到: <workspace>/%s/" % REPORT_SUBDIR)
+        print("\nFlags:")
         print("  --host HOST     Override chat host (default: auto-detect from config.json)")
-        print("  --html          Also generate summary.html with clickable links")
         print("  --scan          Scan all agents")
         sys.exit(1)
 
     # --scan mode
     if scan_mode:
-        output_dir = positional[0] if positional else get_default_output_dir()
+        if not positional:
+            print("ERROR: --scan requires <workspace> argument")
+            sys.exit(1)
+        workspace = positional[0]
+        output_dir = os.path.join(workspace, REPORT_SUBDIR)
         run_scan(output_dir, host=host)
         return
 
-    # Normal mode
+    # Normal mode: first arg = urls, second arg = workspace
+    if len(positional) < 2:
+        print("ERROR: missing <workspace> argument")
+        sys.exit(1)
     arg1 = positional[0]
-    output_dir = positional[1] if len(positional) > 1 else get_default_output_dir()
+    workspace = positional[1]
+    output_dir = os.path.join(workspace, REPORT_SUBDIR)
     items = parse_input(arg1)
     print("Input: %d session(s)" % len(items))
     print("Host: %s" % detect_chat_host())