nolo-cli 0.1.31 → 0.1.33

package/chat/web/fileProcessor.ts

@@ -1,442 +0,0 @@
-// src/features/chat/web/fileProcessor.ts
-
-import { nanoid } from "nanoid";
-import { Descendant } from "slate";
-import toast from "react-hot-toast";
-import { TFunction } from "i18next";
-
-import { createPageAndAddReference } from "../dialog/dialogSlice";
-import { AppDispatch } from "../../app/store";
-import { upload } from "../../database/dbSlice";
-
-// ========= 常量 & 类型 =========
-
-/**
- * 认为是“表格类文件”的扩展名集合
- */
-const EXCEL_LIKE_EXTENSIONS = [
-  ".xlsx",
-  ".xls",
-  ".csv",
-  ".ods",
-  ".xlsm",
-  ".xlsb",
-] as const;
-
-/**
- * 认为是“纯文本类文件”的扩展名集合
- * - 这些文件不需要复杂结构解析，只需要按纯文本展示即可
- * - JSON 单独处理（见下方 isJsonFile），这里不要包含 .json
- */
-const PLAIN_TEXT_EXTENSIONS = [
-  ".txt",
-  ".md",
-  ".mdx",
-  ".log",
-  ".ini",
-  ".cfg",
-  ".yaml",
-  ".yml",
-  ".env",
-  ".toml",
-  ".xml",
-  ".html",
-  ".htm",
-  ".js",
-  ".jsx",
-  ".ts",
-  ".tsx",
-  ".css",
-  ".scss",
-  ".less",
-] as const;
-
-interface ProcessDocumentFileArgs {
-  file: File;
-  /**
-   * 由调用方生成，用于跟踪整个文件操作；
-   * Excel 多 sheet 的情况会为每个 sheet 单独生成 sheetFileId。
-   */
-  fileId: string;
-  dispatch: AppDispatch;
-  t: TFunction<"chat", undefined>;
-  /** PDF OCR 模型，"none" 或未传 时使用 pdf.js 提取文字 */
-  ocrModel?: string;
-  /** OCR 接口上下文，避免零散透传 server/token */
-  ocrRequest?: {
-    serverOrigin?: string;
-    accessToken?: string;
-  };
-  dialogKey?: string;
-}
-
-// ========= 工具函数 =========
-
-/**
- * 是否为表格类文件（Excel / CSV / ODS 等）。
- */
-const isExcelLikeFile = (fileName: string): boolean => {
-  const lower = fileName.toLowerCase();
-  return EXCEL_LIKE_EXTENSIONS.some((ext) => lower.endsWith(ext));
-};
-
-/**
- * 是否为 JSON 文件
- * - 通过扩展名 .json 或 常见 JSON MIME 判断
- */
-const isJsonFile = (file: File, fileNameLower: string): boolean => {
-  return (
-    fileNameLower.endsWith(".json") ||
-    file.type === "application/json" ||
-    file.type === "application/ld+json" ||
-    file.type === "application/jsonl"
-  );
-};
-
-/**
- * 是否为“纯文本类文件”
- *
- * 注意：
- * - 这里不包含 JSON，JSON 有单独的处理逻辑
- * - 只要是 text/* 都认为可以按纯文本展示
- */
-const isPlainTextFile = (file: File, fileNameLower: string): boolean => {
-  // JSON 单独处理
-  if (fileNameLower.endsWith(".json")) return false;
-
-  if (file.type && file.type.startsWith("text/")) return true;
-
-  return PLAIN_TEXT_EXTENSIONS.some((ext) => fileNameLower.endsWith(ext));
-};
-
-/**
- * Excel -> JSON 的通用配置：
- * - raw: true     保留原始值（日期用 Date 或数字表示），不做字符串格式化
- * - cellDates: true  尽量把日期单元格解析成 JS Date 对象
- *
- * 具体的「显示格式」在 convertExcelToSlate 里处理（其中 displayData 会带有 Excel 的显示文本）。
- */
-const getSheetToJsonOptions = () => ({
-  raw: true,
-  cellDates: true,
-});
-
-/**
- * 生成一个 code-block 类型的 Slate 节点内容。
- *
- * 如果你的 schema 不同（比如 type: "code" 或 data.language），
- * 在这里统一改结构即可。
- */
-const createCodeBlockSlate = (
-  text: string,
-  language?: string
-): Descendant[] => {
-  return [
-    {
-      type: "code-block",
-      language,
-      children: [{ text }],
-    } as any,
-  ];
-};
-
-/**
- * JSON 文本 -> Slate 数据（code-block + pretty print）
- * - 若解析成功，则以缩进 2 的格式美化 JSON 字符串
- * - 若解析失败，则原样作为纯文本展示，不抛异常
- */
-const createSlateFromJsonText = (text: string): Descendant[] => {
-  let prettyJson = text;
-
-  try {
-    const parsed = JSON.parse(text);
-    prettyJson = JSON.stringify(parsed, null, 2);
-  } catch {
-    // 忽略解析错误，按原始文本展示
-  }
-
-  return createCodeBlockSlate(prettyJson, "json");
-};
-
-/**
- * 根据扩展名猜测适合的代码高亮语言。
- * 这里只处理“配置 / 数据 / 日志 / 代码”这些适合用 code-block 展示的类型。
- *
- * 说明：
- * - txt / md / mdx 仍然当自然语言文本处理，不在这里返回 language。
- * - JSON 已在 isJsonFile 分支中单独处理，不在这里判断 .json。
- */
-const guessCodeLanguageFromExtension = (
-  fileNameLower: string
-): string | undefined => {
-  if (fileNameLower.endsWith(".yaml") || fileNameLower.endsWith(".yml")) {
-    return "yaml";
-  }
-  if (fileNameLower.endsWith(".ini") || fileNameLower.endsWith(".cfg")) {
-    return "ini";
-  }
-  if (fileNameLower.endsWith(".log")) {
-    return "log";
-  }
-  if (fileNameLower.endsWith(".env")) {
-    // 大部分高亮库没有专门的 dotenv 语言，可以按 shell/ini 处理
-    return "bash";
-  }
-  if (fileNameLower.endsWith(".toml")) {
-    return "toml";
-  }
-  if (fileNameLower.endsWith(".xml")) {
-    return "xml";
-  }
-  if (fileNameLower.endsWith(".html") || fileNameLower.endsWith(".htm")) {
-    return "html";
-  }
-  if (fileNameLower.endsWith(".js") || fileNameLower.endsWith(".jsx")) {
-    return "javascript";
-  }
-  if (fileNameLower.endsWith(".ts") || fileNameLower.endsWith(".tsx")) {
-    return "typescript";
-  }
-  if (fileNameLower.endsWith(".css")) {
-    return "css";
-  }
-  if (fileNameLower.endsWith(".scss") || fileNameLower.endsWith(".sass")) {
-    return "scss";
-  }
-  if (fileNameLower.endsWith(".less")) {
-    return "less";
-  }
-
-  return undefined;
-};
-
-// ========= 主函数 =========
-
-/**
- * 处理单个文档文件：
- *
- * - Excel / CSV / ODS 等多工作表文件：
- *   为每个工作表创建一个页面引用，并为它们附加一个共享的 groupId。
- * - 其他单文件类型（DOCX / PDF / TXT / JSON / 其他纯文本）：
- *   解析后直接生成一个页面引用。
- *
- * 内部使用动态 import，对不同类型文件按需加载各自解析库。
- * 若文件类型不受支持或解析失败会抛出错误。
- */
-export const processDocumentFile = async ({
-  file,
-  fileId,
-  dispatch,
-  t,
-  ocrModel,
-  ocrRequest,
-  dialogKey,
-}: ProcessDocumentFileArgs): Promise<void> => {
-  const toastId = toast.loading(
-    t("processingFile", "正在处理 {{fileName}}...", { fileName: file.name })
-  );
-
-  try {
-    const fileNameLower = file.name.toLowerCase();
-    let fileProcessed = false;
-
-    // 当前语言，用于日期本地化（优先 i18next，其次浏览器语言，最后 en-US）
-    const currentLocale =
-      (t as any)?.i18n?.language ||
-      (typeof navigator !== "undefined" && navigator.language) ||
-      "en-US";
-
-    // ========= 1. Excel / CSV / ODS 等表格类文件 =========
-    if (isExcelLikeFile(fileNameLower)) {
-      fileProcessed = true;
-
-      // 按需加载 xlsx & table actions
-      const XLSX = await import("xlsx");
-      const { createTable, addRow } = await import("../../render/table/tableSlice");
-      const { addPendingFile } = await import("../dialog/dialogSlice");
-
-      const buffer = await file.arrayBuffer();
-      const workbook = XLSX.read(buffer, { type: "array", cellDates: true });
-
-      if (!workbook.SheetNames || workbook.SheetNames.length === 0) {
-        throw new Error(t("excelNoSheets", "Excel 文件中没有找到工作表。"));
-      }
-
-      // 一个工作簿共享一个 groupId
-      const workbookGroupId = nanoid();
-
-      for (const sheetName of workbook.SheetNames) {
-        const worksheet = workbook.Sheets[sheetName];
-
-        const jsonData = XLSX.utils.sheet_to_json<Record<string, any>>(
-          worksheet,
-          getSheetToJsonOptions()
-        );
-
-        if (!jsonData.length) continue;
-
-        const tableTitle = `${file.name} (${sheetName})`;
-
-        // 1. 动态生成列定义
-        const firstRow = jsonData[0];
-        const columns = Object.keys(firstRow).map((key) => ({
-          name: key,
-          label: key,
-          type: "text" as const,
-        }));
-
-        // 2. 创建真正的内置表格
-        const dbKey = await dispatch(
-          createTable({
-            title: tableTitle,
-            columns,
-            withDefaultRows: false,
-          }) as any
-        ).unwrap();
-
-        const parts = dbKey.split("-");
-        const tableId = parts[parts.length - 1];
-        const tenantId = parts[1]; // meta-tenantId-tableId
-
-        // 3. 注入数据行
-        // 注意：这里由于行数可能较多，我们并行处理以提高效率，但生产环境建议分批或后台处理
-        await Promise.all(
-          jsonData.map((row) =>
-            dispatch(
-              addRow({
-                tenantId,
-                tableId,
-                values: row,
-              }) as any
-            ).unwrap()
-          )
-        );
-
-        // 4. 将该表格引用加入对话（直接作为 table 类型）
-        dispatch(
-          addPendingFile({
-            id: nanoid(),
-            name: tableTitle,
-            pageKey: dbKey,
-            type: "table",
-            groupId: workbookGroupId,
-          })
-        );
-      }
-    }
-
-    // ========= 2. 其他单文件类型：DOCX / PDF / TXT / JSON / 其他纯文本 =========
-    else {
-      let slateData: Descendant[] | undefined;
-      // 目前明确区分 docx / pdf / txt / json，其他纯文本统一归为 txt
-      let fileType: "docx" | "pdf" | "txt" | "json" | undefined;
-
-      // DOCX
-      if (fileNameLower.endsWith(".docx")) {
-        fileType = "docx";
-
-        const { convertDocxToSlate } = await import("./docxToSlate");
-
-        // 将 docx 内嵌图片上传到服务器，避免 base64 膨胀消息体
-        const uploadImage = async (blob: Blob, filename: string): Promise<string> => {
-          const imageFile = new File([blob], filename, { type: blob.type });
-          const result = await dispatch(upload({ file: imageFile }) as any).unwrap();
-          return result.id as string;
-        };
-
-        slateData = await convertDocxToSlate(file, uploadImage);
-      }
-      // PDF
-      else if (fileNameLower.endsWith(".pdf")) {
-        fileType = "pdf";
-
-        if (
-          ocrModel &&
-          ocrModel !== "none" &&
-          ocrRequest?.serverOrigin &&
-          ocrRequest?.accessToken
-        ) {
-          // 使用 OCR 模型识别 PDF 页面（适合扫描件等无文字层文档）
-          const { processPdfWithOcr } = await import("./pdfOcrProcessor");
-          slateData = await processPdfWithOcr(file, ocrModel, ocrRequest);
-        } else {
-          // 内部再按需加载 pdfjs-dist & worker
-          const { convertPdfToSlate } = await import(
-            "../../create/editor/utils/pdfToSlate"
-          );
-          slateData = await convertPdfToSlate(file);
-        }
-      }
-      // JSON（通过扩展名或 MIME 类型识别，按“格式化 code-block”展示）
-      else if (isJsonFile(file, fileNameLower)) {
-        fileType = "json";
-
-        const textContent = await file.text();
-        slateData = createSlateFromJsonText(textContent);
-      }
-      // 纯文本类文件（txt / md / log / yaml / ini / cfg / env / toml / js / ts / css 等）
-      else if (isPlainTextFile(file, fileNameLower)) {
-        const textContent = await file.text();
-        const codeLanguage = guessCodeLanguageFromExtension(fileNameLower);
-
-        // 更偏“配置 / 日志 / 代码 / 数据”的扩展名：用 code-block 高亮展示
-        if (codeLanguage) {
-          fileType = "txt"; // 业务上仍然归为 txt，展示层通过 code-block + language 区分
-          slateData = createCodeBlockSlate(textContent, codeLanguage);
-        }
-        // 更偏“自然语言”的 txt / md / mdx：按普通文本处理
-        else {
-          fileType = "txt";
-
-          const { convertTxtToSlate } = await import(
-            "../../create/editor/utils/txtToSlate"
-          );
-          slateData = await convertTxtToSlate(textContent);
-        }
-      }
-
-      if (fileType && slateData) {
-        fileProcessed = true;
-
-        await dispatch(
-          createPageAndAddReference({
-            slateData,
-            title: file.name,
-            type: fileType,
-            fileId,
-            size: file.size,
-            dialogKey,
-          })
-        ).unwrap();
-      }
-    }
-
-    // ========= 3. 文件类型不支持 =========
-    if (!fileProcessed) {
-      throw new Error(t("unsupportedFileType", "不支持的文件类型"));
-    }
-
-    toast.success(
-      t("fileProcessedSuccess", "{{fileName}} 处理成功!", {
-        fileName: file.name,
-      }),
-      { id: toastId }
-    );
-  } catch (err) {
-    const msg =
-      err instanceof Error
-        ? err.message
-        : t("fileProcessDefaultError", "处理文件时发生未知错误");
-
-    toast.error(
-      t("fileProcessedError", "处理 {{fileName}} 时出错: {{error}}", {
-        fileName: file.name,
-        error: msg,
-      }),
-      { id: toastId }
-    );
-
-    // 重新抛出错误，以便调用方可以更新其状态（例如 fileErrors）
-    throw new Error(msg);
-  }
-};

package/chat/web/hooks/useProgressiveDisplayCount.ts

@@ -1,102 +0,0 @@
-import { useState, useLayoutEffect, useEffect } from "react";
-
-interface ProgressiveDisplayOptions {
-    initialCount?: number; // Initial number of items to render
-    batchSize?: number; // Number of items to add per batch
-    itemHeight?: number; // Estimated height of each item in pixels
-}
-
-/**
- * useProgressiveDisplayCount
- * A custom hook to progressively increase the number of displayed items
- * to avoid blocking the main thread during initial render.
- *
- * @param totalItems The total number of items to eventually display.
- * @param resetDeps Dependencies that trigger a reset (e.g., activeTab, spaceId).
- * @param options Configuration options.
- * @returns The current number of items to display and loading status.
- */
-export const useProgressiveDisplayCount = (
-    totalItems: number,
-    resetDeps: any[] = [],
-    options: ProgressiveDisplayOptions = {}
-) => {
-    const {
-        initialCount = 20,
-        batchSize = 50,
-        itemHeight = 37, // Default based on sidebar item height
-    } = options;
-
-    const [displayCount, setDisplayCount] = useState(initialCount);
-    const [isFullyLoaded, setIsFullyLoaded] = useState(false);
-
-    // 1. Reset logic when dependencies change
-    // using useLayoutEffect to prevent initial flash if changing tabs
-    useLayoutEffect(() => {
-        // Calculate how many items fit in the viewport + buffer (1.5x)
-        const viewportCount = Math.ceil((window.innerHeight / itemHeight) * 1.5);
-        const startCount = Math.max(initialCount, viewportCount);
-
-        setDisplayCount(startCount);
-        setIsFullyLoaded(false);
-    }, resetDeps); // eslint-disable-line react-hooks/exhaustive-deps
-
-    // 2. Schedule loading of remaining items using requestIdleCallback
-    useEffect(() => {
-        if (isFullyLoaded || displayCount >= totalItems) {
-            if (!isFullyLoaded) setIsFullyLoaded(true);
-            return;
-        }
-
-        let requestIdleId: any;
-
-        const loadMore = (deadline: IdleDeadline) => {
-            // If there is time remaining or deadline did expire
-            if (deadline.timeRemaining() > 0 || deadline.didTimeout) {
-                setDisplayCount((prev) => {
-                    const nextCount = prev + batchSize;
-                    if (nextCount >= totalItems) {
-                        setIsFullyLoaded(true);
-                        return totalItems;
-                    }
-                    return nextCount;
-                });
-            } else {
-                // No time left, try again in next idle callback
-                // This prevents blocking if the browser is extremely busy
-                // But we re-schedule here just in case? Or let the next effect cycle handle it?
-                // Actually, if we don't update state, effect doesn't re-run.
-                // So we MUST re-schedule.
-                if (typeof window.requestIdleCallback !== 'undefined') {
-                    // @ts-ignore
-                    requestIdleId = window.requestIdleCallback(loadMore);
-                }
-                return;
-            }
-        };
-
-        // Check for browser support
-        if (typeof window.requestIdleCallback !== 'undefined') {
-            // @ts-ignore
-            requestIdleId = window.requestIdleCallback(loadMore);
-        } else {
-            // Fallback for Safari / environments without requestIdleCallback
-            const timer = setTimeout(() => {
-                setDisplayCount(totalItems);
-                setIsFullyLoaded(true);
-            }, 200);
-            return () => clearTimeout(timer);
-        }
-
-        return () => {
-            if (typeof window.cancelIdleCallback !== 'undefined') {
-                // @ts-ignore
-                window.cancelIdleCallback(requestIdleId);
-            }
-        };
-    }, [isFullyLoaded, displayCount, totalItems, batchSize]);
-
-    // If displayCount exceeds totalItems due to batching, cap it (handled in setter logic but safe to keep consistent)
-
-    return { displayCount, isFullyLoaded };
-};