import_glossary_csv

Instructions

Implementation Reference

• src/mcp/tools/import_glossary_csv.ts:27-49 (handler)

  The main handler function `importGlossaryCsv` that validates input via Zod schema, writes CSV content to a temp file, and calls `lara.glossaries.importCsv()`. Includes a 5MB size limit and cleans up the temp directory in a finally block.

  export async function importGlossaryCsv(args: any, lara: Translator) {
    const validatedArgs = importGlossaryCsvSchema.parse(args);
    const { id, csv_content, content_type, gzip } = validatedArgs;
  
    // File size limit: 5MB
    const MAX_CSV_SIZE = 5 * 1024 * 1024;
    if (Buffer.byteLength(csv_content, 'utf8') > MAX_CSV_SIZE) {
      throw new InvalidInputError("CSV file too large. Maximum allowed size is 5MB.");
    }
  
    const tempDir = fs.mkdtempSync(path.join(os.tmpdir(), 'lara-csv-'));
    const tempFilePath = path.join(tempDir, 'import.csv');
  
    try {
      fs.writeFileSync(tempFilePath, csv_content, { mode: 0o600 });
  
      return await lara.glossaries.importCsv(id, tempFilePath, content_type, gzip);
    } finally {
      try {
        fs.rmSync(tempDir, { recursive: true, force: true });
      } catch (_) { /* best-effort cleanup */ }
    }
  }

• src/mcp/tools/import_glossary_csv.ts:8-25 (schema)

  Zod schema `importGlossaryCsvSchema` defining four input fields: id (glossary ID matching gls_* pattern), csv_content (the CSV file content string), content_type (enum csv/table-uni or csv/table-multi, defaults to csv/table-uni), and gzip (optional boolean).

  export const importGlossaryCsvSchema = z.object({
    id: z.string()
      .min(1)
      .max(255)
      .regex(/^gls_[a-zA-Z0-9_-]+$/, "Invalid glossary ID format")
      .describe("The glossary ID (format: gls_*, e.g., 'gls_xyz123')"),
    csv_content: z
      .string()
      .describe("The content of the CSV file to upload"),
    content_type: z
      .enum(["csv/table-uni", "csv/table-multi"])
      .default("csv/table-uni")
      .describe("The format of the CSV file. 'csv/table-uni' for unidirectional, 'csv/table-multi' for multidirectional"),
    gzip: z
      .boolean()
      .optional()
      .describe("Whether the CSV content is gzip compressed"),
  });

• src/mcp/tools.ts:30-30 (registration)

  Import of `importGlossaryCsv` and `importGlossaryCsvSchema` from the tool module.

  import { importGlossaryCsv, importGlossaryCsvSchema } from "./tools/import_glossary_csv.js";

• src/mcp/tools.ts:62-62 (registration)

  Registration of `import_glossary_csv` handler in the handlers record mapping tool names to handler functions.

  import_glossary_csv: importGlossaryCsv,

• src/mcp/tools.ts:401-413 (registration)

  MCP tool definition including name, description, inputSchema (JSON schema from Zod), annotations (title, readOnlyHint false, destructiveHint false), and _meta invocation messages.

  {
    name: "import_glossary_csv",
    description:
      "Imports a CSV file into a glossary. Supports unidirectional and multidirectional formats. This is an async operation that returns an import job object containing an import_id. Poll with check_glossary_import_status using the returned import_id until the import is complete.",
    inputSchema: z.toJSONSchema(importGlossaryCsvSchema),
    annotations: {
      title: "Import glossary CSV",
      readOnlyHint: false,
      destructiveHint: false,
      openWorldHint: false,
    },
    _meta: invocationMeta("Queuing glossary import…", "Glossary import queued"),
  },