From 35a1b70208c3742caf6124ca70bbe38de74002f6 Mon Sep 17 00:00:00 2001 From: Liang Jiaqing Date: Sun, 1 Feb 2026 08:58:26 +0800 Subject: [PATCH] Refactor: restructure directory and clean up redundant scripts --- agentapp.py | 9 ++- assets/global_mem_insight_template.txt | 28 +++++++ .../ljq_web_driver.user.js | 0 make_prompts.py => assets/make_prompts.py | 4 +- sys_prompt.txt => assets/sys_prompt.txt | 0 tools_schema.json => assets/tools_schema.json | 2 +- ga.py | 7 +- simphtml.py | 69 ++++++++++++++++- web_tools.py | 75 ------------------- 9 files changed, 109 insertions(+), 85 deletions(-) create mode 100644 assets/global_mem_insight_template.txt rename ljq_web_driver.user.js => assets/ljq_web_driver.user.js (100%) rename make_prompts.py => assets/make_prompts.py (98%) rename sys_prompt.txt => assets/sys_prompt.txt (100%) rename tools_schema.json => assets/tools_schema.json (96%) delete mode 100644 web_tools.py diff --git a/agentapp.py b/agentapp.py index c2c100a..ffc7bbe 100644 --- a/agentapp.py +++ b/agentapp.py @@ -7,7 +7,7 @@ sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), '..'))) import streamlit as st import time, json, re -with open('tools_schema.json', 'r', encoding='utf-8') as f: +with open('assets/tools_schema.json', 'r', encoding='utf-8') as f: TOOLS_SCHEMA = json.load(f) st.set_page_config(page_title="Cowork", layout="wide") @@ -34,8 +34,11 @@ def get_system_prompt(): if not os.path.exists('memory/global_mem.txt'): with open('memory/global_mem.txt', 'w', encoding='utf-8') as f: f.write('') if not os.path.exists('memory/global_mem_insight.txt'): - with open('memory/global_mem_insight.txt', 'w', encoding='utf-8') as f: f.write('') - with open('sys_prompt.txt', 'r', encoding='utf-8') as f: prompt = f.read() + content = "## Global Memory Index (Logic)\n\n[CONSTITUTION]\n1. 改我自身源码前必须先问用户\n\n[STORES]\n- global_mem: ../memory/global_mem.txt\n\n[ACCESS]\n- global_mem: 按 TOPIC 检索索引\n\n[TOPICS.GLOBAL_MEM]" + if os.path.exists('assets/global_mem_insight_template.txt'): + with open('assets/global_mem_insight_template.txt', 'r', encoding='utf-8') as f: content = f.read() + with open('memory/global_mem_insight.txt', 'w', encoding='utf-8') as f: f.write(content) + with open('assets/sys_prompt.txt', 'r', encoding='utf-8') as f: prompt = f.read() prompt += get_global_memory() return prompt diff --git a/assets/global_mem_insight_template.txt b/assets/global_mem_insight_template.txt new file mode 100644 index 0000000..4d98b24 --- /dev/null +++ b/assets/global_mem_insight_template.txt @@ -0,0 +1,28 @@ +## Global Memory Index (Logic) + +[CONSTITUTION] +1. 改我自身源码前必须先问用户；在 ./ 内自由实验可直接做。 +2. 要下结论/依赖环境/做操作前：先枚举可用存储(store)与索引；再读取相关条目；未验证不硬断言。 +3. 没有用户授权不要直接文本读取或移动密钥/密码文件，能引用即可。或者先请求用户许可。 + +[STORES] +- global_mem: ../memory/global_mem.txt (Facts, Policy, Prefs) +- memory_dir: ../memory/ (SOPs, Tools) + +[ACCESS] +- global_mem: 按 TOPIC 检索索引 → file_read 对应条目。 +- memory_dir: ls ../memory/ + +[TOPICS.GLOBAL_MEM] +# 【引导注释 - 使用后请删除】 +# 请模仿以下概括性 Tag 格式记录 global_mem.txt 中的事实： +# - env/network: 网络代理或环境描述 (Global Facts ## [SECTION]) +# - auth/platform: 平台账号或认证信息 (Global Facts ## [SECTION]) +# - pref/download_dir: 文件存储偏好路径 (Global Facts ## [SECTION]) +# - tool/config: 核心工具运行参数 (Global Facts ## [SECTION]) + +[TOPICS.MEMORY_DIR] +# 【引导注释 - 使用后请删除】 +# 请模仿以下格式登记 memory 目录下的独立文件： +# - sop/workflow: 标准操作流程文档 (.md) +# - tool/automation: 自动化执行脚本 (.py) \ No newline at end of file diff --git a/ljq_web_driver.user.js b/assets/ljq_web_driver.user.js similarity index 100% rename from ljq_web_driver.user.js rename to assets/ljq_web_driver.user.js diff --git a/make_prompts.py b/assets/make_prompts.py similarity index 98% rename from make_prompts.py rename to assets/make_prompts.py index 5905083..72b19e4 100644 --- a/make_prompts.py +++ b/assets/make_prompts.py @@ -15,7 +15,7 @@ def generate_tool_schema(): """ 通过代码内省，将 Handler 的逻辑映射为高语义的工具描述。 """ - with open('ga.py', 'r', encoding='utf-8') as f: + with open('../ga.py', 'r', encoding='utf-8') as f: ga_code = f.read() # 极简且具备高度概括能力的元 Prompt meta_prompt = f""" @@ -79,7 +79,7 @@ def generate_tool_schema(): return None -def make_system_prompt(ga_code_path='ga.py'): +def make_system_prompt(ga_code_path='../ga.py'): with open(ga_code_path, 'r', encoding='utf-8') as f: ga_code = f.read() diff --git a/sys_prompt.txt b/assets/sys_prompt.txt similarity index 100% rename from sys_prompt.txt rename to assets/sys_prompt.txt diff --git a/tools_schema.json b/assets/tools_schema.json similarity index 96% rename from tools_schema.json rename to assets/tools_schema.json index 822b070..3e5fa8e 100644 --- a/tools_schema.json +++ b/assets/tools_schema.json @@ -1,7 +1,7 @@ [ {"type": "function", "function": { "name": "code_run", - "description": "针对 Windows 优化的双模态代码执行器。优先使用python，仅在必要系统操作时使用 powershell。注意：执行的代码必须以 ```python 或 ```powershell 代码块的形式包含在回复正文中。严禁在代码中硬编码大量数据，如有需要应通过文件读取。执行时间限制为 60s。", + "description": "针对 Windows 优化的双模态代码执行器。优先使用python，仅在必要系统操作时使用 powershell。注意：执行的代码必须放在在回复正文中，以 ```python 或 ```powershell 代码块的形式。严禁在代码中硬编码大量数据，如有需要应通过文件读取。", "parameters": {"type": "object", "properties": { "type": {"type": "string", "enum": ["python", "powershell"], "description": "执行环境类型，默认为 python。", "default": "python"}, "timeout": {"type": "integer", "description": "执行超时时间（秒），默认 60。", "default": 60}, diff --git a/ga.py b/ga.py index 605d605..7c8d26c 100644 --- a/ga.py +++ b/ga.py @@ -97,7 +97,7 @@ def ask_user(question: str, candidates: list = None): } } -from web_tools import execute_js_rich, get_html +from simphtml import execute_js_rich, get_html driver = None @@ -109,8 +109,9 @@ def first_init_driver(): time.sleep(1) sess = driver.get_all_sessions() if len(sess) > 0: break - driver.newtab() - time.sleep(5) + if len(sess) == 1: + driver.newtab() + time.sleep(5) def web_scan(focus_item="", switch_tab_id=None): """ diff --git a/simphtml.py b/simphtml.py index 18c277e..4542e86 100644 --- a/simphtml.py +++ b/simphtml.py @@ -862,4 +862,71 @@ def find_changed_elements(before_html, after_html): "changed": len(changed_elements), "similarity": similarity #"changed_elements": changed_elements[:3] - } \ No newline at end of file + } + + + +def get_html(driver, cutlist=False, maxchars=28000, instruction=""): + page = get_main_block(driver) + soup = optimize_html_for_tokens(page) + html = str(soup) + if not cutlist or len(html) <= maxchars: return html + rr = driver.execute_js(js_findMainList + js_findMainContent + """ + return findMainList(findMainContent(document.body));""") + sel = rr.get("selector", None) + if not sel: return html[:maxchars] + s = BeautifulSoup(str(soup), "html.parser"); items = s.select(sel) + hit = [it for it in items if instruction and instruction.strip() and instruction in it.get_text(" ",strip=True)] + keep = hit[:6] if hit else items[:3] + for it in items: + if it not in keep: it.decompose() + s = optimize_html_for_tokens(s) + return str(s)[:maxchars] + +def execute_js_rich(script, driver): + start_temp_monitor(driver) + curr_session = driver.default_session_id + last_html = get_html(driver) + result = None; error_msg = None + new_tab = False; reloaded = False + try: + print(f"⚡ Executing: {script[:250]} ...") + result = driver.execute_js(script, auto_switch_newtab=True) + if type(result) is dict and result.get('closed', 0) == 1: reloaded = True + time.sleep(2) + except Exception as e: + error = e.args[0] if e.args else str(e) + if isinstance(error, dict): error.pop('stack', None) + error_msg = str(error) + print(f"❌ Error: {error_msg}") + + transients = get_temp_texts(driver) + + if driver.default_session_id != curr_session: + curr_session = driver.latest_session_id + print('Session changed') + new_tab = True + + current_html = get_html(driver) + diff_summary = "无需对比 (报错)" + is_significant_change = False + if not error_msg: + diff_data = find_changed_elements(last_html, current_html) + change_count = diff_data.get('changed', 0) + diff_summary = f"DOM变化量: {change_count}" + if change_count < 5 and not transients and not new_tab: + diff_summary += " (页面几乎无静默变化)" + else: + is_significant_change = True + return { + "status": "failed" if error_msg else "success", + "js_return": result, + "error": error_msg, + "transients": transients, + "environment": { + "new_tab": new_tab, + "reloaded": reloaded + }, + "diff": diff_summary, + "suggestion": "" if is_significant_change else "页面无明显变化" + } diff --git a/web_tools.py b/web_tools.py deleted file mode 100644 index ec591cc..0000000 --- a/web_tools.py +++ /dev/null @@ -1,75 +0,0 @@ -import sys, os, re -import pyperclip -import json, time -import subprocess -import tempfile -sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), '..'))) - -from simphtml import get_main_block, start_temp_monitor, get_temp_texts, find_changed_elements, optimize_html_for_tokens -from simphtml import js_findMainContent, js_findMainList -from bs4 import BeautifulSoup - -def get_html(driver, cutlist=False, maxchars=28000, instruction=""): - page = get_main_block(driver) - soup = optimize_html_for_tokens(page) - html = str(soup) - if not cutlist or len(html) <= maxchars: return html - rr = driver.execute_js(js_findMainList + js_findMainContent + """ - return findMainList(findMainContent(document.body));""") - sel = rr.get("selector", None) - if not sel: return html[:maxchars] - s = BeautifulSoup(str(soup), "html.parser"); items = s.select(sel) - hit = [it for it in items if instruction and instruction.strip() and instruction in it.get_text(" ",strip=True)] - keep = hit[:6] if hit else items[:3] - for it in items: - if it not in keep: it.decompose() - s = optimize_html_for_tokens(s) - return str(s)[:maxchars] - -def execute_js_rich(script, driver): - start_temp_monitor(driver) - curr_session = driver.default_session_id - last_html = get_html(driver) - result = None; error_msg = None - new_tab = False; reloaded = False - try: - print(f"⚡ Executing: {script[:250]} ...") - result = driver.execute_js(script, auto_switch_newtab=True) - if type(result) is dict and result.get('closed', 0) == 1: reloaded = True - time.sleep(2) - except Exception as e: - error = e.args[0] if e.args else str(e) - if isinstance(error, dict): error.pop('stack', None) - error_msg = str(error) - print(f"❌ Error: {error_msg}") - - transients = get_temp_texts(driver) - - if driver.default_session_id != curr_session: - curr_session = driver.latest_session_id - print('Session changed') - new_tab = True - - current_html = get_html(driver) - diff_summary = "无需对比 (报错)" - is_significant_change = False - if not error_msg: - diff_data = find_changed_elements(last_html, current_html) - change_count = diff_data.get('changed', 0) - diff_summary = f"DOM变化量: {change_count}" - if change_count < 5 and not transients and not new_tab: - diff_summary += " (页面几乎无静默变化)" - else: - is_significant_change = True - return { - "status": "failed" if error_msg else "success", - "js_return": result, - "error": error_msg, - "transients": transients, - "environment": { - "new_tab": new_tab, - "reloaded": reloaded - }, - "diff": diff_summary, - "suggestion": "" if is_significant_change else "页面无明显变化" - }