Jarvis-Brain 0.1.9.15__py3-none-any.whl → 0.1.10.0__py3-none-any.whl

{jarvis_brain-0.1.9.15.dist-info → jarvis_brain-0.1.10.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: Jarvis_Brain
-Version: 0.1.9.15
+Version: 0.1.10.0
 Summary: Jarvis brain mcp
 Requires-Python: >=3.10
 Requires-Dist: beautifulsoup4

jarvis_brain-0.1.10.0.dist-info/RECORD

@@ -0,0 +1,12 @@
+mcp_tools/__init__.py,sha256=_ZzedxbxmVkdRH00ONOjQR31pZW-B8WPn5P2qE6T-Iw,3012
+mcp_tools/chrome_devtools_tools.py,sha256=LEcDb2gamjJdSzqA9_Wy-lCtx8B-9p1M97Ik9GScXrY,284
+mcp_tools/dp_tools.py,sha256=_2ys1eedp2AxfRxeQ5U_o6qE_eII6J7mxAWLXaKy9Sg,17208
+mcp_tools/main.py,sha256=k5YOuf2USW7vnlFEBxvLSfxPPHP7nrM1J8lLmVEcSuY,1116
+tools/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+tools/browser_manager.py,sha256=zeYcWuzxoohMdnYUoZbRH7axFC_VtV8MsncfN8y0yw0,2023
+tools/browser_proxy.py,sha256=YDBp1bDEX08ev-Bvbuv0TkWz_nkShe95GS7tfd8GBLc,7606
+tools/tools.py,sha256=E2NyOwJtFh5hT18nvCwIYs23OipKyjJEHIfyZJ1Bsjg,8960
+jarvis_brain-0.1.10.0.dist-info/METADATA,sha256=o6l0jL2sUiifn_rx2zEY4Xtma6Ajqi2NP7quCTtB0NQ,242
+jarvis_brain-0.1.10.0.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+jarvis_brain-0.1.10.0.dist-info/entry_points.txt,sha256=YFQT4xpkUqt5dM5wlKPQQOqcjMuFrT9iuRAzIpAyH7U,51
+jarvis_brain-0.1.10.0.dist-info/RECORD,,

mcp_tools/chrome_devtools_tools.py

@@ -0,0 +1,11 @@
+"""
+这个文件中提供的工具作为谷歌官方mcp：chrome-devtools-mcp的辅助工具，仅提供功能补充与浏览器实例管理
+"""
+import hashlib
+import json
+import os
+from typing import Any
+
+from fastmcp import FastMCP
+
+from tools.browser_manager import BrowserManager

mcp_tools/dp_tools.py

@@ -1,3 +1,6 @@
+"""
+这个文件中提供的工具作为独立的Drissionpage mcp工具
+"""
 import hashlib
 import json
 import os
@@ -6,7 +9,8 @@ from typing import Any
 from fastmcp import FastMCP
 
 from tools.browser_manager import BrowserManager
-from tools.tools import compress_html, requests_html, dp_headless_html, assert_waf_cookie, dp_mcp_message_pack
+from tools.tools import compress_html, requests_html, dp_headless_html, assert_waf_cookie, dp_mcp_message_pack, \
+    compress_html_js
 from tools.browser_proxy import DPProxyClient, DPProxyClientManager
 
 html_source_code_local_save_path = os.path.join(os.getcwd(), "html-source-code")
@@ -94,10 +98,11 @@ def register_get_html(mcp: FastMCP, browser_manager):
         file_name_prefix = hashlib.md5(str(tab.title).encode('utf-8')).hexdigest()
         if not os.path.exists(html_source_code_local_save_path):
             os.makedirs(html_source_code_local_save_path)
-        min_html, compress_rate = compress_html(tab.html)
-        html_str_list = [min_html[i:i + one_turn_max_token] for i in range(0, len(min_html), one_turn_max_token)]
+        # min_html, compress_rate = compress_html(tab.html)
+        min_html = tab.run_js(compress_html_js)
+        # html_str_list = [min_html[i:i + one_turn_max_token] for i in range(0, len(min_html), one_turn_max_token)]
         html_file_list = []
-        for index, html_str in enumerate(html_str_list):
+        for index, html_str in enumerate([min_html]):
             file_name = file_name_prefix + f"_{tab_id}_segment{index}.html"
             abs_path = os.path.join(html_source_code_local_save_path, file_name)
             with open(abs_path, "w", encoding="utf-8") as f:

tools/tools.py

@@ -7,6 +7,87 @@ from bs4 import BeautifulSoup
 from curl_cffi import requests
 from lxml import html, etree
 
+compress_html_js = """
+    function getSimplifiedDOM(node) {
+        // 1. 处理文本节点
+        if (node.nodeType === Node.TEXT_NODE) {
+            const text = node.textContent.trim();
+            // 限制文本长度，避免大段文章消耗 token，保留前100个字符通常足够定位
+            return text ? text.slice(0, 100) + (text.length > 100 ? '...' : '') : null;
+        }
+
+        // 2. 过滤无用标签
+        const ignoreTags = ['SCRIPT', 'STYLE', 'NOSCRIPT', 'IFRAME', 'SVG', 'LINK', 'META'];
+        if (ignoreTags.includes(node.tagName)) return null;
+        if (node.nodeType !== Node.ELEMENT_NODE) return null;
+
+        // 3. 过滤不可见元素
+        const style = window.getComputedStyle(node);
+        if (style.display === 'none' || style.visibility === 'hidden' || style.opacity === '0') return null;
+        // 过滤宽高太小的元素（往往是埋点空像素）
+        const rect = node.getBoundingClientRect();
+        if (rect.width === 0 || rect.height === 0) return null;
+
+        // --- 开始构建标签字符串 ---
+        const tagName = node.tagName.toLowerCase();
+        let tagStr = tagName;
+
+        // A. 基础标识符 (ID 和 Class)
+        if (node.id) tagStr += `#${node.id}`;
+        if (node.className && typeof node.className === 'string') {
+            // 过滤掉 Tailwind 等太长且无语义的 class，保留有意义的业务 class
+            // 这里简单处理，全部保留，让 LLM 自己判断
+            const classes = node.className.trim().split(/\s+/);
+            if (classes.length > 0) tagStr += `.${classes.join('.')}`;
+        }
+
+        // B. 关键属性白名单 (这是你指出问题的核心修复)
+        const props = [];
+
+        // 通用重要属性
+        if (node.getAttribute('role')) props.push(`role="${node.getAttribute('role')}"`);
+        if (node.getAttribute('aria-label')) props.push(`aria-label="${node.getAttribute('aria-label')}"`);
+        if (node.getAttribute('title')) props.push(`title="${node.getAttribute('title')}"`);
+
+        // 特定标签的特定属性
+        if (tagName === 'a') {
+            const href = node.getAttribute('href');
+            // 只保留有意义的链接，忽略 javascript:;
+            if (href && !href.startsWith('javascript')) props.push(`href="${href}"`);
+        } else if (tagName === 'input' || tagName === 'textarea' || tagName === 'select') {
+            if (node.getAttribute('type')) props.push(`type="${node.getAttribute('type')}"`);
+            if (node.getAttribute('name')) props.push(`name="${node.getAttribute('name')}"`);
+            if (node.getAttribute('placeholder')) props.push(`placeholder="${node.getAttribute('placeholder')}"`);
+            if (node.disabled) props.push('disabled');
+            if (node.checked) props.push('checked');
+        } else if (tagName === 'button') {
+            if (node.getAttribute('type')) props.push(`type="${node.getAttribute('type')}"`);
+        } else if (tagName === 'img') {
+            if (node.getAttribute('alt')) props.push(`alt="${node.getAttribute('alt')}"`);
+        }
+
+        if (props.length > 0) {
+            tagStr += ` ${props.join(' ')}`;
+        }
+
+        // 4. 递归子节点
+        const children = Array.from(node.childNodes)
+            .map(getSimplifiedDOM)
+            .filter(n => n !== null);
+
+        // 5. 组装输出
+        // 如果没有子节点，也没有ID/Class，也不是输入框/图片/链接，那这个标签可能只是布局用的 div，可以考虑跳过它直接返回子节点内容
+        // 但为了保持结构完整，我们暂时保留它
+        if (children.length === 0) {
+            // 自闭合标签或空标签
+            return `<${tagStr} />`;
+        }
+        return `<${tagStr}>${children.join('')}</${tagName}>`; // 结束标签只保留 tagName 节省 token
+    }
+
+    return getSimplifiedDOM(document.body);
+"""
+
 
 # 使用requests获取html，用于测试是否使用了瑞数和jsl
 def requests_html(url):

jarvis_brain-0.1.9.15.dist-info/RECORD

@@ -1,11 +0,0 @@
-mcp_tools/__init__.py,sha256=_ZzedxbxmVkdRH00ONOjQR31pZW-B8WPn5P2qE6T-Iw,3012
-mcp_tools/dp_tools.py,sha256=ugzHnOf43fJ2LutWlDrYplc4TPn8ZLP9Byb85yENvfM,17059
-mcp_tools/main.py,sha256=k5YOuf2USW7vnlFEBxvLSfxPPHP7nrM1J8lLmVEcSuY,1116
-tools/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-tools/browser_manager.py,sha256=zeYcWuzxoohMdnYUoZbRH7axFC_VtV8MsncfN8y0yw0,2023
-tools/browser_proxy.py,sha256=YDBp1bDEX08ev-Bvbuv0TkWz_nkShe95GS7tfd8GBLc,7606
-tools/tools.py,sha256=zYJCvwy7OUmYWKOhHbItSfttAmJLbC-ixXv9oW-TNTA,5033
-jarvis_brain-0.1.9.15.dist-info/METADATA,sha256=g9Bp_gpYdKAXYMaiWgZGesmaLYpF5KxSJ9ijuwdUn8E,242
-jarvis_brain-0.1.9.15.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
-jarvis_brain-0.1.9.15.dist-info/entry_points.txt,sha256=YFQT4xpkUqt5dM5wlKPQQOqcjMuFrT9iuRAzIpAyH7U,51
-jarvis_brain-0.1.9.15.dist-info/RECORD,,