@truto/truto-jsonata 1.0.12 → 1.0.13

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/dist/main.cjs CHANGED
@@ -12,7 +12,6 @@ var $dxT2C$mime = require("mime");
12
12
  var $dxT2C$pmap = require("p-map");
13
13
  var $dxT2C$pretry = require("p-retry");
14
14
  var $dxT2C$officeparser = require("officeparser");
15
- var $dxT2C$filetype = require("file-type");
16
15
  var $dxT2C$pdfjsserverless = require("pdfjs-serverless");
17
16
  var $dxT2C$langchaintextsplitters = require("@langchain/textsplitters");
18
17
 
@@ -1409,6 +1408,17 @@ var $caaee789061bb8bb$export$2e2bcd8739ae039 = $caaee789061bb8bb$var$generateEmb
1409
1408
 
1410
1409
 
1411
1410
 
1411
+ async function $0f748d6318103cdf$var$bufferToString(value, encoding) {
1412
+ if (value instanceof ReadableStream) {
1413
+ const chunks = [];
1414
+ for await (const chunk of value)chunks.push(Buffer.from(chunk));
1415
+ const buffer = Buffer.concat(chunks);
1416
+ return buffer.toString(encoding);
1417
+ }
1418
+ return value.toString(encoding);
1419
+ }
1420
+ var $0f748d6318103cdf$export$2e2bcd8739ae039 = $0f748d6318103cdf$var$bufferToString;
1421
+
1412
1422
 
1413
1423
  async function $05e3378f7c17d263$var$parsePdf(buffer) {
1414
1424
  const data = buffer.buffer.slice(buffer?.byteOffset, buffer?.byteOffset + buffer?.byteLength);
@@ -1433,16 +1443,23 @@ async function $05e3378f7c17d263$var$parsePdf(buffer) {
1433
1443
  // Return the results as JSON
1434
1444
  return (0, $dxT2C$lodashes.join)(output, "\n");
1435
1445
  }
1436
- async function $05e3378f7c17d263$var$parseDocument(file) {
1446
+ async function $05e3378f7c17d263$var$parseDocument(file, fileType) {
1447
+ let buffer;
1437
1448
  if (file instanceof ReadableStream) {
1438
1449
  const chunks = [];
1439
1450
  for await (const chunk of file)chunks.push(Buffer.from(chunk));
1440
- const buffer = Buffer.concat(chunks);
1441
- const fileExt = (await (0, $dxT2C$filetype.fileTypeFromBuffer)(buffer))?.ext;
1442
- if (fileExt === "pdf") return await $05e3378f7c17d263$var$parsePdf(buffer);
1443
- return await (0, $dxT2C$officeparser.parseOfficeAsync)(buffer);
1451
+ buffer = Buffer.concat(chunks);
1444
1452
  }
1445
- return await (0, $dxT2C$officeparser.parseOfficeAsync)(file);
1453
+ if ((0, $dxT2C$lodashes.includes)([
1454
+ "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
1455
+ "application/vnd.openxmlformats-officedocument.presentationml.presentation",
1456
+ "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
1457
+ "application/vnd.oasis.opendocument.text",
1458
+ "application/vnd.oasis.opendocument.presentation",
1459
+ "application/vnd.oasis.opendocument.spreadsheet"
1460
+ ], fileType)) return await (0, $dxT2C$officeparser.parseOfficeAsync)(buffer);
1461
+ if (fileType === "application/pdf") return await $05e3378f7c17d263$var$parsePdf(buffer);
1462
+ return (0, $0f748d6318103cdf$export$2e2bcd8739ae039)(buffer, "utf-8");
1446
1463
  }
1447
1464
  var $05e3378f7c17d263$export$2e2bcd8739ae039 = $05e3378f7c17d263$var$parseDocument;
1448
1465
 
@@ -1457,21 +1474,19 @@ function $fe4dcef142601b8c$export$e600492876ee595b(text, options = {
1457
1474
  }
1458
1475
 
1459
1476
 
1460
-
1461
- async function $15c432f5f036a88a$var$getDataUri(file) {
1477
+ async function $15c432f5f036a88a$var$getDataUri(file, mimeType) {
1478
+ if (!mimeType) throw new Error("Mime type is required");
1462
1479
  if (file instanceof ReadableStream) {
1463
1480
  const chunks = [];
1464
1481
  for await (const chunk of file)chunks.push(Buffer.from(chunk));
1465
1482
  const buffer = Buffer.concat(chunks);
1466
1483
  const base64Image = buffer.toString("base64");
1467
- const mimeType = (await (0, $dxT2C$filetype.fileTypeFromBuffer)(buffer))?.mime;
1468
1484
  // Construct the data URI for a PNG image
1469
1485
  return `data:${mimeType};base64,${base64Image}`;
1470
1486
  }
1471
1487
  const arrayBuffer = file instanceof Blob ? await file.arrayBuffer() : file;
1472
1488
  const buffer = Buffer.from(arrayBuffer);
1473
1489
  const base64Image = buffer.toString("base64");
1474
- const mimeType = (await (0, $dxT2C$filetype.fileTypeFromBuffer)(buffer))?.mime;
1475
1490
  return `data:${mimeType};base64,${base64Image}`;
1476
1491
  }
1477
1492
  var $15c432f5f036a88a$export$2e2bcd8739ae039 = $15c432f5f036a88a$var$getDataUri;
@@ -1483,6 +1498,7 @@ async function $9a2529096849a04f$var$teeStream(stream) {
1483
1498
  var $9a2529096849a04f$export$2e2bcd8739ae039 = $9a2529096849a04f$var$teeStream;
1484
1499
 
1485
1500
 
1501
+
1486
1502
  function $af351c41b7fd6f79$export$2e2bcd8739ae039(expression) {
1487
1503
  expression.registerFunction("dtFromIso", (0, $bab42b5e4be720d3$export$2e2bcd8739ae039));
1488
1504
  expression.registerFunction("base64decode", (0, $77081a2d6d46cd50$export$2e2bcd8739ae039));
@@ -1564,6 +1580,7 @@ function $af351c41b7fd6f79$export$2e2bcd8739ae039(expression) {
1564
1580
  expression.registerFunction("recursiveCharacterTextSplitter", (0, $fe4dcef142601b8c$export$e600492876ee595b));
1565
1581
  expression.registerFunction("getDataUri", (0, $15c432f5f036a88a$export$2e2bcd8739ae039));
1566
1582
  expression.registerFunction("teeStream", (0, $9a2529096849a04f$export$2e2bcd8739ae039));
1583
+ expression.registerFunction("bufferToString", (0, $0f748d6318103cdf$export$2e2bcd8739ae039));
1567
1584
  return expression;
1568
1585
  }
1569
1586