Add Revit CSV transport

2026-08-02 15:56:02 +02:00
parent 43cee890e7
commit bfe8b790b2
4 changed files with 521 additions and 5 deletions
@@ -0,0 +1,124 @@
+import assert from "node:assert/strict";
+import { describe, it } from "node:test";
+import {
+  ExternalCsvParseError,
+  parseExternalCsv,
+  serializeExternalCsv,
+} from "../src/external-model/csv/external-csv-transport.js";
+import {
+  externalCsvTestConfiguration,
+  firstIfcGuid,
+  quotedRevitCsv,
+  utf8Bytes,
+} from "./fixtures/revit-csv-fixtures.js";
+
+describe("external CSV transport", () => {
+  it("detects the observed Revit dialect and classifies every row", () => {
+    const document = parseExternalCsv(
+      utf8Bytes(quotedRevitCsv, true),
+      externalCsvTestConfiguration
+    );
+
+    assert.deepEqual(document.dialect, {
+      encoding: "utf-8",
+      hasBom: true,
+      delimiter: ";",
+      lineEnding: "\r\n",
+      quoteCharacter: '"',
+      quoteAllFields: true,
+      hasTrailingLineEnding: true,
+    });
+    assert.equal(document.headerRowIndex, 1);
+    assert.deepEqual(
+      document.rows.map((row) => row.classification),
+      [
+        "metadata",
+        "header",
+        "passthrough",
+        "object",
+        "passthrough",
+        "suspect-object",
+        "object",
+      ]
+    );
+    assert.equal(document.rows[6].cells[2].value, 'Büro "Nord"');
+  });
+
+  it("round-trips BOM, CRLF, quotation, order and cell values byte for byte", () => {
+    const source = utf8Bytes(quotedRevitCsv, true);
+    const document = parseExternalCsv(source, externalCsvTestConfiguration);
+
+    assert.deepEqual(serializeExternalCsv(document), source);
+  });
+
+  it("preserves embedded delimiters and line endings inside quoted cells", () => {
+    const csv = quotedRevitCsv.replace(
+      '"Steckdose: Doppelsteckdose"',
+      '"Steckdose; Doppel\r\nsteckdose"'
+    );
+    const source = utf8Bytes(csv, true);
+    const document = parseExternalCsv(source, externalCsvTestConfiguration);
+
+    assert.equal(
+      document.rows[3].cells[3].value,
+      "Steckdose; Doppel\r\nsteckdose"
+    );
+    assert.deepEqual(serializeExternalCsv(document), source);
+  });
+
+  it("blocks missing and ambiguous configured headers", () => {
+    assertParseError(
+      utf8Bytes(quotedRevitCsv.replace("IfcGUID", "Andere ID")),
+      "header-not-found"
+    );
+    const duplicatedHeader = quotedRevitCsv.replace(
+      '"";"";"";"";"";"";"";""',
+      quotedRevitCsv.split("\r\n")[1]
+    );
+    assertParseError(utf8Bytes(duplicatedHeader), "ambiguous-header");
+
+    const duplicateConfiguredColumn = quotedRevitCsv.replace(
+      '"Stromkreis";"Raumnummer"',
+      '"Stromkreis";"Raumnummer";"Raumnummer"'
+    );
+    assertParseError(
+      utf8Bytes(duplicateConfiguredColumn),
+      "ambiguous-header"
+    );
+  });
+
+  it("blocks invalid and duplicate non-empty IfcGUIDs", () => {
+    assertParseError(
+      utf8Bytes(quotedRevitCsv.replace(firstIfcGuid, "not-an-ifc-guid")),
+      "invalid-ifc-guid"
+    );
+    assertParseError(
+      utf8Bytes(quotedRevitCsv.replace(firstIfcGuid, ` ${firstIfcGuid}`)),
+      "invalid-ifc-guid"
+    );
+    const duplicate = quotedRevitCsv.replace(
+      "1AbCdEfGhIjKlMnOpQrStu",
+      firstIfcGuid
+    );
+    assertParseError(utf8Bytes(duplicate), "duplicate-ifc-guid");
+  });
+
+  it("blocks invalid UTF-8, unterminated quotes and mixed row endings", () => {
+    assertParseError(new Uint8Array([0xff, 0xfe, 0xfd]), "invalid-encoding");
+    assertParseError(
+      utf8Bytes(quotedRevitCsv.replace(/"\r\n$/, "\r\n")),
+      "invalid-csv"
+    );
+    assertParseError(
+      utf8Bytes(quotedRevitCsv.replace("\r\n", "\n")),
+      "invalid-csv"
+    );
+  });
+});
+
+function assertParseError(input: Uint8Array, code: ExternalCsvParseError["code"]) {
+  assert.throws(
+    () => parseExternalCsv(input, externalCsvTestConfiguration),
+    (error: unknown) => error instanceof ExternalCsvParseError && error.code === code
+  );
+}
@@ -0,0 +1,42 @@
+import {
+  createDefaultExternalCsvConfiguration,
+  type ExternalCsvColumnMapping,
+} from "../../src/external-model/csv/external-csv-contracts.js";
+
+export const externalCsvTestColumns: ExternalCsvColumnMapping = {
+  ifcGuid: "IfcGUID",
+  roomNumber: "Raumnummer",
+  roomName: "Raumname",
+  familyAndType: "Familie und Typ",
+  selectionMarker: "CAx_Auswahlkenner",
+  circuitIdentifier: "Stromkreis",
+  power: "Elektrische Leistung",
+  quantity: "Anzahl",
+};
+
+export const externalCsvTestConfiguration =
+  createDefaultExternalCsvConfiguration(externalCsvTestColumns);
+
+export const firstIfcGuid = "0AbCdEfGhIjKlMnOpQrStu";
+export const secondIfcGuid = "1AbCdEfGhIjKlMnOpQrStu";
+
+export const quotedRevitCsv = [
+  '"Projekt-Export";"";"";"";"";"";"";""',
+  '"Stromkreis";"Raumnummer";"Raumname";"Familie und Typ";"CAx_Auswahlkenner";"Elektrische Leistung";"Anzahl";"IfcGUID"',
+  '"";"";"";"";"";"";"";""',
+  `"UV_AV_01-2F1";"01/101";"Technik";"Steckdose: Doppelsteckdose";"Arbeitsplatz";"120,5";"2";"${firstIfcGuid}"`,
+  '"UV_AV_01-2F1: 2";"";"";"";"";"";"";""',
+  '"";"01/102";"Büro";"Steckdose: Standard";"Arbeitsplatz";"100";"1";""',
+  `"";"01/103";"Büro \"\"Nord\"\"";"Steckdose: Standard";"Arbeitsplatz";"100";"1";"${secondIfcGuid}"`,
+].join("\r\n") + "\r\n";
+
+export function utf8Bytes(value: string, withBom = false) {
+  const encoded = new TextEncoder().encode(value);
+  if (!withBom) {
+    return encoded;
+  }
+  const result = new Uint8Array(encoded.length + 3);
+  result.set([0xef, 0xbb, 0xbf]);
+  result.set(encoded, 3);
+  return result;
+}