@web-auto/webauto 0.1.1 → 0.1.2

package/dist/modules/workflow/blocks/helpers/commentSectionLocator.js

@@ -0,0 +1,126 @@
+/**
+ * Comment Section Locator Helper
+ *
+ * 共享定位逻辑：WarmupCommentsBlock / ExpandCommentsBlock
+ * - 通过容器锚点 verifyAnchorByContainerId 获取 rect
+ * - 必要时尝试点击 comment_button 激活评论区
+ */
+import { logControllerActionError, logControllerActionResult, logControllerActionStart, } from './operationLogger.js';
+export async function clickCommentButtonByContainerId(options) {
+    const { profile, serviceUrl, controllerUrl, commentButtonContainerId, highlightStyle = '2px solid #ff00ff', highlightMs = 1200, } = options;
+    try {
+        const { verifyAnchorByContainerId } = await import('./containerAnchors.js');
+        const vp = await getViewport(controllerUrl, profile);
+        const btnAnchor = await verifyAnchorByContainerId(commentButtonContainerId, profile, serviceUrl, highlightStyle, highlightMs);
+        if (btnAnchor.found && btnAnchor.rect && vp.innerWidth && vp.innerHeight) {
+            const bx = clamp(btnAnchor.rect.x + btnAnchor.rect.width / 2, 30, vp.innerWidth - 30);
+            const by = clamp(btnAnchor.rect.y + btnAnchor.rect.height / 2, 120, vp.innerHeight - 120);
+            console.log(`[WarmupComments] click comment_button @(${Math.floor(bx)},${Math.floor(by)})`);
+            const { systemClickAt } = await import('./systemInput.js');
+            await systemClickAt(profile, Math.floor(bx), Math.floor(by), undefined, 'click_comment_button');
+            await new Promise((r) => setTimeout(r, 800));
+            return { clicked: true, rect: btnAnchor.rect };
+        }
+        return {
+            clicked: false,
+            rect: btnAnchor.rect,
+            error: btnAnchor.error || 'comment_button anchor not found',
+        };
+    }
+    catch (e) {
+        return { clicked: false, error: e?.message || String(e) };
+    }
+}
+export async function locateCommentSection(options) {
+    const { profile, serviceUrl, controllerUrl, commentSectionContainerId, commentButtonContainerId, canClickCommentButton, highlightStyle = '2px solid #ffaa00', highlightMs = 2000, } = options;
+    const { verifyAnchorByContainerId } = await import('./containerAnchors.js');
+    let clicked = false;
+    const tryClickCommentButton = async (reason) => {
+        if (!canClickCommentButton)
+            return;
+        if (!commentButtonContainerId)
+            return;
+        if (clicked)
+            return;
+        try {
+            const vp = await getViewport(controllerUrl, profile);
+            const btnAnchor = await verifyAnchorByContainerId(commentButtonContainerId, profile, serviceUrl, '2px solid #ff00ff', 1200);
+            if (btnAnchor.found && btnAnchor.rect && vp.innerWidth && vp.innerHeight) {
+                const bx = clamp(btnAnchor.rect.x + btnAnchor.rect.width / 2, 30, vp.innerWidth - 30);
+                const by = clamp(btnAnchor.rect.y + btnAnchor.rect.height / 2, 120, vp.innerHeight - 120);
+                console.log(`[WarmupComments] click comment_button (${reason}) @(${Math.floor(bx)},${Math.floor(by)})`);
+                const { systemClickAt } = await import('./systemInput.js');
+                await systemClickAt(profile, Math.floor(bx), Math.floor(by), undefined, 'click_comment_button');
+                clicked = true;
+                await new Promise((r) => setTimeout(r, 800));
+            }
+        }
+        catch {
+            // ignore
+        }
+    };
+    try {
+        let anchor = await verifyAnchorByContainerId(commentSectionContainerId, profile, serviceUrl, highlightStyle, highlightMs);
+        if (!anchor?.found) {
+            await tryClickCommentButton('comment_section_not_found');
+            anchor = await verifyAnchorByContainerId(commentSectionContainerId, profile, serviceUrl, highlightStyle, highlightMs);
+        }
+        if (anchor.found && anchor.rect) {
+            console.log(`[WarmupComments] comment_section rect: ${JSON.stringify(anchor.rect)}`);
+            return { found: true, rect: anchor.rect, clickedCommentButton: clicked };
+        }
+        return {
+            found: false,
+            error: anchor.error || 'comment_section anchor not found',
+            clickedCommentButton: clicked,
+        };
+    }
+    catch (e) {
+        return {
+            found: false,
+            error: `comment_section anchor verify error: ${e?.message || e}`,
+            clickedCommentButton: clicked,
+        };
+    }
+}
+async function controllerAction(controllerUrl, action, payload = {}) {
+    const opId = logControllerActionStart(action, payload, { source: 'commentSectionLocator' });
+    try {
+        const response = await fetch(controllerUrl, {
+            method: 'POST',
+            headers: { 'Content-Type': 'application/json' },
+            body: JSON.stringify({ action, payload }),
+            signal: AbortSignal.timeout
+                ? AbortSignal.timeout(10000)
+                : undefined,
+        });
+        if (!response.ok) {
+            throw new Error(`HTTP ${response.status}: ${await response.text()}`);
+        }
+        const data = await response.json();
+        const result = data.data || data;
+        logControllerActionResult(opId, action, result, { source: 'commentSectionLocator' });
+        return result;
+    }
+    catch (error) {
+        logControllerActionError(opId, action, error, payload, { source: 'commentSectionLocator' });
+        throw error;
+    }
+}
+async function getViewport(controllerUrl, profile) {
+    try {
+        const result = await controllerAction(controllerUrl, 'browser:execute', {
+            profile,
+            script: '(() => ({ w: window.innerWidth || 0, h: window.innerHeight || 0 }))()',
+        });
+        const payload = result.result || result.data?.result || result;
+        return { innerWidth: Number(payload?.w || 0), innerHeight: Number(payload?.h || 0) };
+    }
+    catch {
+        return { innerWidth: 0, innerHeight: 0 };
+    }
+}
+function clamp(n, min, max) {
+    return Math.min(Math.max(n, min), max);
+}
+//# sourceMappingURL=commentSectionLocator.js.map

package/dist/modules/workflow/blocks/helpers/containerAnchors.js

@@ -0,0 +1,301 @@
+/**
+ * 容器锚点验证辅助函数（不依赖 containers:match）
+ * 直接从容器定义 JSON 读取 selector，然后用 browser:execute 高亮 + Rect 回环
+ *
+ * 注意：这里不使用 dist 里的 ContainerDefinitionLoader（V2 结构只保留了 classes，丢失 css），
+ * 而是直接读取 container-library 下的原始 JSON，保证 selector.css 可用。
+ */
+import fs from 'node:fs';
+import path from 'node:path';
+import os from 'node:os';
+import { logError, logOperation } from './operationLogger.js';
+const repoRoot = path.resolve(process.cwd());
+const userContainerRoot = process.env.WEBAUTO_USER_CONTAINER_ROOT || path.join(os.homedir(), '.webauto', 'container-lib');
+const containerIndexPath = process.env.WEBAUTO_CONTAINER_INDEX || path.join(repoRoot, 'apps/webauto/resources/container-library.index.json');
+const definitionCache = new Map();
+function loadContainerIndex() {
+    try {
+        const content = fs.readFileSync(containerIndexPath, 'utf-8');
+        return JSON.parse(content);
+    }
+    catch {
+        return null;
+    }
+}
+function findSiteKeyForContainer(containerId, index) {
+    if (!index)
+        return null;
+    const keys = Object.keys(index);
+    for (const key of keys) {
+        if (containerId.startsWith(`${key}_`))
+            return key;
+    }
+    return null;
+}
+function findContainerJsonPath(containerId) {
+    const index = loadContainerIndex();
+    const siteKey = findSiteKeyForContainer(containerId, index) || 'xiaohongshu';
+    const rootDir = path.join(repoRoot, 'apps/webauto/resources/container-library', siteKey);
+    if (!fs.existsSync(rootDir)) {
+        return null;
+    }
+    const stack = [rootDir];
+    while (stack.length > 0) {
+        const current = stack.pop();
+        const stat = fs.statSync(current);
+        if (stat.isDirectory()) {
+            const entries = fs.readdirSync(current);
+            for (const entry of entries) {
+                stack.push(path.join(current, entry));
+            }
+        }
+        else if (current.endsWith('.json')) {
+            try {
+                const content = fs.readFileSync(current, 'utf-8');
+                const json = JSON.parse(content);
+                if (json && json.id === containerId) {
+                    return current;
+                }
+            }
+            catch {
+                // ignore malformed JSON
+            }
+        }
+    }
+    return null;
+}
+async function findContainerDefinition(containerId) {
+    if (definitionCache.has(containerId)) {
+        const cached = definitionCache.get(containerId);
+        if (cached) {
+            return cached;
+        }
+        // 如果之前没有找到（缓存为 null），重新扫描一次，避免负缓存长期生效
+    }
+    // 优先从用户自定义容器根查找
+    if (userContainerRoot && fs.existsSync(userContainerRoot)) {
+        const userRoot = path.join(userContainerRoot, 'xiaohongshu');
+        if (fs.existsSync(userRoot)) {
+            const stack = [userRoot];
+            while (stack.length > 0) {
+                const current = stack.pop();
+                const stat = fs.statSync(current);
+                if (stat.isDirectory()) {
+                    const entries = fs.readdirSync(current);
+                    for (const entry of entries) {
+                        stack.push(path.join(current, entry));
+                    }
+                }
+                else if (current.endsWith('.json')) {
+                    try {
+                        const content = fs.readFileSync(current, 'utf-8');
+                        const json = JSON.parse(content);
+                        if (json && json.id === containerId) {
+                            definitionCache.set(containerId, json);
+                            return json;
+                        }
+                    }
+                    catch {
+                        // ignore
+                    }
+                }
+            }
+        }
+    }
+    const jsonPath = findContainerJsonPath(containerId);
+    if (!jsonPath) {
+        return null;
+    }
+    try {
+        const content = fs.readFileSync(jsonPath, 'utf-8');
+        const json = JSON.parse(content);
+        definitionCache.set(containerId, json);
+        return json;
+    }
+    catch (error) {
+        return null;
+    }
+}
+function pickPrimarySelector(container) {
+    const selectors = container.selectors || [];
+    if (!selectors.length)
+        return null;
+    const primary = selectors.find((item) => item.variant === 'primary') || selectors[0];
+    return primary?.css || null;
+}
+export async function getPrimarySelectorByContainerId(containerId) {
+    const container = await findContainerDefinition(containerId);
+    if (!container) {
+        return null;
+    }
+    return pickPrimarySelector(container);
+}
+export async function getContainerExtractorsById(containerId) {
+    const container = await findContainerDefinition(containerId);
+    if (!container || !container.extractors || typeof container.extractors !== 'object') {
+        return null;
+    }
+    return container.extractors;
+}
+export async function verifyAnchorByContainerId(containerId, sessionId, serviceUrl = 'http://127.0.0.1:7701', highlightStyle = '3px solid #ff4444', highlightDuration = 2000) {
+    console.log(`[verifyAnchorByContainerId] Looking for container: ${containerId}`);
+    console.log(`[verifyAnchorByContainerId] containerIndexPath: ${containerIndexPath}`);
+    console.log(`[verifyAnchorByContainerId] repoRoot: ${repoRoot}`);
+    const container = await findContainerDefinition(containerId);
+    if (!container) {
+        console.error(`[verifyAnchorByContainerId] Container not found: ${containerId}`);
+        return { found: false, highlighted: false, error: `Container not found: ${containerId}` };
+    }
+    const selectors = (container.selectors || [])
+        .map((s) => s?.css)
+        .filter((css) => typeof css === 'string' && css.trim().length > 0);
+    if (!selectors.length) {
+        return { found: false, highlighted: false, error: `No selector defined for: ${containerId}` };
+    }
+    const matchOpId = logOperation({
+        kind: 'anchor_match_start',
+        action: 'anchor:verify',
+        sessionId,
+        payload: {
+            containerId,
+            selectorsCount: selectors.length,
+            serviceUrl,
+        },
+    });
+    try {
+        const selectorsJson = JSON.stringify(selectors);
+        const script = `(() => {
+      const selectors = ${selectorsJson};
+
+      const isVisible = (el) => {
+        if (!el) return false;
+        const style = window.getComputedStyle(el);
+        if (style.display === 'none' || style.visibility === 'hidden' || style.opacity === '0') return false;
+        const r = el.getBoundingClientRect();
+        if (!r.width || !r.height) return false;
+        if (r.bottom <= 0 || r.top >= window.innerHeight) return false;
+        return true;
+      };
+
+      let el = null;
+      let usedSelector = null;
+      for (const sel of selectors) {
+        const candidate = document.querySelector(sel);
+        if (candidate && isVisible(candidate)) {
+          el = candidate;
+          usedSelector = sel;
+          break;
+        }
+      }
+
+      if (!el) {
+        return { found: false, error: 'Element not found' };
+      }
+
+      // 创建或复用一个 overlay 高亮框，避免被页面样式覆盖
+      let overlay = document.getElementById('webauto-anchor-highlight');
+      if (!overlay) {
+        overlay = document.createElement('div');
+        overlay.id = 'webauto-anchor-highlight';
+        overlay.style.position = 'fixed';
+        overlay.style.pointerEvents = 'none';
+        overlay.style.zIndex = '2147483647';
+        document.body.appendChild(overlay);
+      }
+      const r = el.getBoundingClientRect();
+      overlay.style.left = r.x + 'px';
+      overlay.style.top = r.y + 'px';
+      overlay.style.width = r.width + 'px';
+      overlay.style.height = r.height + 'px';
+      overlay.style.border = '${highlightStyle.replace(/'/g, "\\'")}';
+      overlay.style.boxSizing = 'border-box';
+      overlay.style.background = 'transparent';
+
+      setTimeout(() => {
+        if (overlay && overlay.parentElement) {
+          overlay.parentElement.removeChild(overlay);
+        }
+      }, ${highlightDuration});
+      const rect = el.getBoundingClientRect();
+      return {
+        found: true,
+        selector: usedSelector,
+        rect: { x: rect.x, y: rect.y, width: rect.width, height: rect.height }
+      };
+    })()`;
+        const response = await fetch(`${serviceUrl}/v1/controller/action`, {
+            method: 'POST',
+            headers: { 'Content-Type': 'application/json' },
+            body: JSON.stringify({
+                action: 'browser:execute',
+                payload: {
+                    profile: sessionId,
+                    script,
+                },
+            }),
+            // 本地 Unified API 调用在部分页面初始化时可能稍慢，适当放宽超时时间
+            signal: AbortSignal.timeout ? AbortSignal.timeout(30000) : undefined,
+        });
+        if (!response.ok) {
+            return {
+                found: false,
+                highlighted: false,
+                error: `HTTP ${response.status}: ${await response.text()}`,
+            };
+        }
+        const data = await response.json();
+        const result = data.data?.result || data.result;
+        if (!result?.found) {
+            logOperation({
+                kind: 'anchor_match_result',
+                action: 'anchor:verify',
+                sessionId,
+                result: {
+                    found: false,
+                    selector: result?.selector || null,
+                    error: result?.error || 'Element not found',
+                },
+                meta: { opId: matchOpId, containerId },
+            });
+            return {
+                found: false,
+                highlighted: false,
+                selector: result?.selector,
+                error: result?.error || 'Element not found',
+            };
+        }
+        logOperation({
+            kind: 'anchor_match_result',
+            action: 'anchor:verify',
+            sessionId,
+            result: {
+                found: true,
+                selector: result?.selector || null,
+                rect: result?.rect || null,
+            },
+            meta: { opId: matchOpId, containerId },
+        });
+        return {
+            found: true,
+            highlighted: true,
+            rect: result.rect,
+            selector: result.selector,
+        };
+    }
+    catch (error) {
+        logError({
+            kind: 'anchor_match_error',
+            action: 'anchor:verify',
+            sessionId,
+            error: error?.message || String(error),
+            payload: { containerId, serviceUrl },
+            meta: { opId: matchOpId },
+        });
+        return {
+            found: false,
+            highlighted: false,
+            error: `verifyAnchor failed: ${error.message}`,
+        };
+    }
+}
+//# sourceMappingURL=containerAnchors.js.map

package/dist/modules/workflow/blocks/helpers/debugArtifacts.js

@@ -0,0 +1,6 @@
+export function isDebugArtifactsEnabled() {
+    return (process.env.WEBAUTO_DEBUG === '1' ||
+        process.env.WEBAUTO_DEBUG_ARTIFACTS === '1' ||
+        process.env.WEBAUTO_DEBUG_SCREENSHOT === '1');
+}
+//# sourceMappingURL=debugArtifacts.js.map

package/dist/modules/workflow/blocks/helpers/downloadPaths.js

@@ -0,0 +1,29 @@
+import os from 'node:os';
+import path from 'node:path';
+export function sanitizeForPath(name) {
+    if (!name)
+        return '';
+    return name.replace(/[\\/:"*?<>|]+/g, '_').trim();
+}
+export function resolveDownloadRoot(custom, homeDir) {
+    if (custom && custom.trim())
+        return custom;
+    if (homeDir && homeDir.trim())
+        return path.join(homeDir, '.webauto', 'download');
+    const envHome = process.env.HOME || process.env.USERPROFILE;
+    if (envHome && envHome.trim())
+        return path.join(envHome, '.webauto', 'download');
+    try {
+        return path.join(os.homedir(), '.webauto', 'download');
+    }
+    catch {
+        return path.join(process.cwd(), '.webauto', 'download');
+    }
+}
+export function resolveKeywordDir(input) {
+    const { platform, env, keyword, homeDir, downloadRoot } = input;
+    const safeKeyword = sanitizeForPath(keyword) || 'unknown';
+    const root = resolveDownloadRoot(downloadRoot, homeDir);
+    return path.join(root, platform, env, safeKeyword);
+}
+//# sourceMappingURL=downloadPaths.js.map

package/dist/modules/workflow/blocks/helpers/expandCommentsController.js

@@ -0,0 +1,53 @@
+/**
+ * expandCommentsController.ts
+ *
+ * ExpandComments 专用的 controller 调用封装
+ */
+import { logControllerActionError, logControllerActionResult, logControllerActionStart, } from './operationLogger.js';
+/**
+ * 创建 ExpandComments Controller 客户端
+ */
+export function createExpandCommentsControllerClient(config) {
+    const { profile, controllerUrl } = config;
+    async function controllerAction(action, payload = {}) {
+        const opId = logControllerActionStart(action, payload, { source: 'expandCommentsController' });
+        try {
+            const response = await fetch(controllerUrl, {
+                method: 'POST',
+                headers: { 'Content-Type': 'application/json' },
+                body: JSON.stringify({ action, payload }),
+                signal: AbortSignal.timeout
+                    ? AbortSignal.timeout(10000)
+                    : undefined,
+            });
+            if (!response.ok) {
+                throw new Error(`HTTP ${response.status}: ${await response.text()}`);
+            }
+            const data = await response.json();
+            const result = data.data || data;
+            logControllerActionResult(opId, action, result, { source: 'expandCommentsController' });
+            return result;
+        }
+        catch (error) {
+            logControllerActionError(opId, action, error, payload, { source: 'expandCommentsController' });
+            throw error;
+        }
+    }
+    async function getCurrentUrl() {
+        const response = await fetch(controllerUrl, {
+            method: 'POST',
+            headers: { 'Content-Type': 'application/json' },
+            body: JSON.stringify({
+                action: 'browser:execute',
+                payload: {
+                    profile,
+                    script: 'location.href',
+                },
+            }),
+        });
+        const data = await response.json();
+        return data.data?.result || data.result || '';
+    }
+    return { controllerAction, getCurrentUrl };
+}
+//# sourceMappingURL=expandCommentsController.js.map

package/dist/modules/workflow/blocks/helpers/expandCommentsExtractor.js

@@ -0,0 +1,129 @@
+/**
+ * expandCommentsExtractor.ts
+ *
+ * ExpandComments 的 DOM 提取逻辑封装（只读 query，不做 click/scroll）
+ */
+export function buildExtractCommentsScript(cfg) {
+    const domConfig = {
+        rootSelectors: cfg.rootSelectors,
+        itemSelector: cfg.itemSelector,
+        fields: cfg.extractors,
+    };
+    return `(() => {
+    const cfg = ${JSON.stringify(domConfig)};
+    const pickRoot = () => {
+      const roots = cfg.rootSelectors || [];
+      for (const sel of roots) {
+        try {
+          const el = document.querySelector(sel);
+          if (el) return el;
+        } catch (_) {}
+      }
+      return null;
+    };
+
+    const root = pickRoot();
+    const scope = root || document;
+
+    const DEFAULT_ITEM_SELECTOR = [
+      '.comment-item',
+      "[class*='comment-item']",
+      // 一些页面的“回复/子评论”可能不叫 comment-item，但仍属于评论计数
+      // 注意：这里必须避免把“展开更多/显示更多”按钮误当成评论项
+      "[class*='reply-item']",
+      "[class*='replyItem']",
+      "[class*='sub-comment']",
+      "[class*='subComment']",
+    ].join(', ');
+
+    const rawItems = Array.from(scope.querySelectorAll(cfg.itemSelector || DEFAULT_ITEM_SELECTOR));
+    const items = rawItems.filter((el) => {
+      if (!(el instanceof HTMLElement)) return false;
+      // 排除明显的“展开更多/显示更多”按钮（这些会被 replyExpander 处理）
+      if (el.matches('.show-more, .reply-expand, [class*=\"show-more\"], [class*=\"reply-expand\"], [class*=\"expand\"][class*=\"more\"]')) {
+        return false;
+      }
+      const tag = String(el.tagName || '').toUpperCase();
+      if (tag === 'BUTTON') return false;
+      const t = (el.textContent || '').trim();
+      if (!t) return false;
+      // 常见展开按钮文案很短；避免把它当成评论
+      if (t.length <= 16 && t.includes('展开')) return false;
+      return true;
+    });
+
+    const comments = items.map((el, idx) => {
+      const item = {};
+      const fields = cfg.fields || {};
+
+      const getAttrValue = (node, attr) => {
+        if (!node) return '';
+        if (!attr || attr === 'textContent') {
+          return (node.textContent || '').trim();
+        }
+        if (attr === 'href') {
+          return (node.href || node.getAttribute('href') || '').trim();
+        }
+        const v = node.getAttribute(attr);
+        return v ? v.trim() : '';
+      };
+
+      for (const fieldName of Object.keys(fields)) {
+        const fieldCfg = fields[fieldName] || {};
+        const sels = Array.isArray(fieldCfg.selectors) ? fieldCfg.selectors : [];
+        let value = '';
+        for (const sel of sels) {
+          try {
+            const node = el.querySelector(sel);
+            if (!node) continue;
+            value = getAttrValue(node, fieldCfg.attr);
+            if (value) break;
+          } catch (_) {}
+        }
+        item[fieldName] = value;
+      }
+
+      item._idx = idx;
+      item.comment_id =
+        el.getAttribute('data-id') ||
+        el.getAttribute('data-comment-id') ||
+        el.getAttribute('data-commentid') ||
+        el.getAttribute('id') ||
+        '';
+
+      if (!item.text) {
+        item.text = (el.textContent || '').trim();
+      }
+      item.is_reply = !!el.closest('.reply-container');
+      return item;
+    });
+
+    return { found: true, comments };
+  })()`;
+}
+export function mergeExtractedComments(params) {
+    const { rawList, seenKeys, out, maxOut } = params;
+    const limit = typeof maxOut === 'number' && Number.isFinite(maxOut) && maxOut > 0 ? Math.floor(maxOut) : null;
+    for (const c of rawList) {
+        if (limit && out.length >= limit)
+            break;
+        if (!c || typeof c !== 'object')
+            continue;
+        const hasContent = Boolean(c.text && String(c.text).trim()) ||
+            Boolean(c.user_name && String(c.user_name).trim());
+        if (!hasContent)
+            continue;
+        const cid = c.comment_id || c.commentId || c.id || '';
+        const idx = typeof c._idx === 'number' ? String(c._idx) : '';
+        const key = cid
+            ? `id:${cid}`
+            : `idx:${idx}||${c.user_id || ''}||${c.user_name || ''}||${c.text || ''}||${c.timestamp || ''}`;
+        if (seenKeys.has(key))
+            continue;
+        seenKeys.add(key);
+        // 供上层做跨 tab / 跨批次去重
+        c._key = key;
+        out.push(c);
+    }
+}
+//# sourceMappingURL=expandCommentsExtractor.js.map