@truto/truto-jsonata 1.0.11 → 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
@@ -12,10 +12,8 @@ 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
- var $dxT2C$nodestreamweb = require("node:stream/web");
19
17
 
20
18
  function $parcel$interopDefault(a) {
21
19
  return a && a.__esModule ? a.default : a;
@@ -1410,6 +1408,17 @@ var $caaee789061bb8bb$export$2e2bcd8739ae039 = $caaee789061bb8bb$var$generateEmb
1410
1408
 
1411
1409
 
1412
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
+
1413
1422
 
1414
1423
  async function $05e3378f7c17d263$var$parsePdf(buffer) {
1415
1424
  const data = buffer.buffer.slice(buffer?.byteOffset, buffer?.byteOffset + buffer?.byteLength);
@@ -1434,16 +1443,23 @@ async function $05e3378f7c17d263$var$parsePdf(buffer) {
1434
1443
  // Return the results as JSON
1435
1444
  return (0, $dxT2C$lodashes.join)(output, "\n");
1436
1445
  }
1437
- async function $05e3378f7c17d263$var$parseDocument(file) {
1438
- if (file) {
1446
+ async function $05e3378f7c17d263$var$parseDocument(file, fileType) {
1447
+ let buffer;
1448
+ if (file instanceof ReadableStream) {
1439
1449
  const chunks = [];
1440
1450
  for await (const chunk of file)chunks.push(Buffer.from(chunk));
1441
- const buffer = Buffer.concat(chunks);
1442
- const fileExt = (await (0, $dxT2C$filetype.fileTypeFromBuffer)(buffer))?.ext;
1443
- if (fileExt === "pdf") return await $05e3378f7c17d263$var$parsePdf(buffer);
1444
- return await (0, $dxT2C$officeparser.parseOfficeAsync)(buffer);
1451
+ buffer = Buffer.concat(chunks);
1445
1452
  }
1446
- 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");
1447
1463
  }
1448
1464
  var $05e3378f7c17d263$export$2e2bcd8739ae039 = $05e3378f7c17d263$var$parseDocument;
1449
1465
 
@@ -1458,27 +1474,31 @@ function $fe4dcef142601b8c$export$e600492876ee595b(text, options = {
1458
1474
  }
1459
1475
 
1460
1476
 
1461
-
1462
-
1463
- async function $15c432f5f036a88a$var$getDataUri(file) {
1464
- if (file instanceof (0, $dxT2C$nodestreamweb.ReadableStream)) {
1477
+ async function $15c432f5f036a88a$var$getDataUri(file, mimeType) {
1478
+ if (!mimeType) throw new Error("Mime type is required");
1479
+ if (file instanceof ReadableStream) {
1465
1480
  const chunks = [];
1466
1481
  for await (const chunk of file)chunks.push(Buffer.from(chunk));
1467
1482
  const buffer = Buffer.concat(chunks);
1468
1483
  const base64Image = buffer.toString("base64");
1469
- const mimeType = (await (0, $dxT2C$filetype.fileTypeFromBuffer)(buffer))?.mime;
1470
1484
  // Construct the data URI for a PNG image
1471
1485
  return `data:${mimeType};base64,${base64Image}`;
1472
1486
  }
1473
1487
  const arrayBuffer = file instanceof Blob ? await file.arrayBuffer() : file;
1474
1488
  const buffer = Buffer.from(arrayBuffer);
1475
1489
  const base64Image = buffer.toString("base64");
1476
- const mimeType = (await (0, $dxT2C$filetype.fileTypeFromBuffer)(buffer))?.mime;
1477
1490
  return `data:${mimeType};base64,${base64Image}`;
1478
1491
  }
1479
1492
  var $15c432f5f036a88a$export$2e2bcd8739ae039 = $15c432f5f036a88a$var$getDataUri;
1480
1493
 
1481
1494
 
1495
+ async function $9a2529096849a04f$var$teeStream(stream) {
1496
+ return stream.tee();
1497
+ }
1498
+ var $9a2529096849a04f$export$2e2bcd8739ae039 = $9a2529096849a04f$var$teeStream;
1499
+
1500
+
1501
+
1482
1502
  function $af351c41b7fd6f79$export$2e2bcd8739ae039(expression) {
1483
1503
  expression.registerFunction("dtFromIso", (0, $bab42b5e4be720d3$export$2e2bcd8739ae039));
1484
1504
  expression.registerFunction("base64decode", (0, $77081a2d6d46cd50$export$2e2bcd8739ae039));
@@ -1559,6 +1579,8 @@ function $af351c41b7fd6f79$export$2e2bcd8739ae039(expression) {
1559
1579
  expression.registerFunction("parseDocument", (0, $05e3378f7c17d263$export$2e2bcd8739ae039));
1560
1580
  expression.registerFunction("recursiveCharacterTextSplitter", (0, $fe4dcef142601b8c$export$e600492876ee595b));
1561
1581
  expression.registerFunction("getDataUri", (0, $15c432f5f036a88a$export$2e2bcd8739ae039));
1582
+ expression.registerFunction("teeStream", (0, $9a2529096849a04f$export$2e2bcd8739ae039));
1583
+ expression.registerFunction("bufferToString", (0, $0f748d6318103cdf$export$2e2bcd8739ae039));
1562
1584
  return expression;
1563
1585
  }
1564
1586