@office-open/xlsx 0.10.15 → 0.11.0

package/README.md

@@ -4,7 +4,7 @@
 ![npm downloads](https://img.shields.io/npm/dw/@office-open/xlsx)
 ![npm license](https://img.shields.io/npm/l/@office-open/xlsx)
 
-> Generate and parse .xlsx spreadsheets with a declarative TypeScript API. Works in Node.js and browsers.
+> Create Excel spreadsheets (.xlsx) in TypeScript and JavaScript — generate, parse, and patch from plain JSON, no Microsoft Office required. Built for AI agents and hand-written code alike; runs in Node.js, browsers, Deno, and Bun.
 
 ## Features
 
@@ -78,29 +78,33 @@ Performance comparison against [hucre](https://github.com/nicolo-ribaudo/hucre)
 await generateWorkbook(options);
 // All STORE (no compression)
 await generateWorkbook(options, { compression: { xml: 0, media: 0 } });
+// Stream as ReadableStream<Uint8Array> (pipe to a file / HTTP response)
+generateWorkbookStream(options);
 ```
 
-**Create + toBuffer (end-to-end)**
+**Create + toBuffer / toStream (end-to-end)**
 
-| Scenario         | Default sync | Default async | All STORE sync | All STORE async |     hucre |
-| ---------------- | -----------: | ------------: | -------------: | --------------: | --------: |
-| Simple (3 rows)  |  2,040 ops/s |   1,096 ops/s |   21,165 ops/s |    21,851 ops/s | 854 ops/s |
-| Styled rows (20) |  1,894 ops/s |   1,107 ops/s |   18,327 ops/s |    18,581 ops/s | 815 ops/s |
-| Table (10x5)     |  2,017 ops/s |   1,148 ops/s |   17,836 ops/s |    16,761 ops/s | 865 ops/s |
+| Scenario         | Default sync | Default async | All STORE sync | All STORE async | Default stream |     hucre |
+| ---------------- | -----------: | ------------: | -------------: | --------------: | -------------: | --------: |
+| Simple (3 rows)  |  1,923 ops/s |   1,115 ops/s |   18,126 ops/s |    16,147 ops/s |     22.0 ops/s | 782 ops/s |
+| Styled rows (20) |  2,120 ops/s |   1,199 ops/s |   16,476 ops/s |    16,039 ops/s |     21.5 ops/s | 797 ops/s |
+| Table (10x5)     |  1,847 ops/s |   1,105 ops/s |   16,181 ops/s |    16,204 ops/s |     21.9 ops/s | 875 ops/s |
 
-**Large Files — Create + toBuffer**
+**Large Files — Create + toBuffer / toStream**
 
-| Scenario                      | Default sync | Default async | All STORE sync | All STORE async |      hucre |
-| ----------------------------- | -----------: | ------------: | -------------: | --------------: | ---------: |
-| 2000 rows + 10 images         |    110 ops/s |     127 ops/s |      127 ops/s |       134 ops/s | 43.7 ops/s |
-| 200x10 table                  |    821 ops/s |     610 ops/s |    1,253 ops/s |     1,172 ops/s |  251 ops/s |
-| 20 sheets × 100 rows + 20 img |   73.8 ops/s |    56.9 ops/s |       96 ops/s |        99 ops/s | 22.8 ops/s |
+| Scenario                      | Default sync | Default async | All STORE sync | All STORE async | Default stream |      hucre |
+| ----------------------------- | -----------: | ------------: | -------------: | --------------: | -------------: | ---------: |
+| 2000 rows + 10 images         |    131 ops/s |     138 ops/s |      135 ops/s |       140 ops/s |     12.1 ops/s | 42.2 ops/s |
+| 200x10 table                  |    834 ops/s |     610 ops/s |    1,357 ops/s |     1,337 ops/s |     19.9 ops/s |  242 ops/s |
+| 20 sheets × 100 rows + 20 img |   71.1 ops/s |    53.3 ops/s |     97.3 ops/s |       101 ops/s |     2.26 ops/s | 24.3 ops/s |
 
 **Large Data — 100,000 rows × 20 columns (2M cells)**
 
-| Scenario  | Default sync | Default async | All STORE sync | All STORE async |      hucre |
-| --------- | -----------: | ------------: | -------------: | --------------: | ---------: |
-| 100k × 20 |   0.86 ops/s |    0.84 ops/s |     1.01 ops/s |      0.92 ops/s | 0.40 ops/s |
+| Scenario  | Default sync | Default async | All STORE sync | All STORE async | Default stream |      hucre |
+| --------- | -----------: | ------------: | -------------: | --------------: | -------------: | ---------: |
+| 100k × 20 |   0.92 ops/s |    0.90 ops/s |     1.16 ops/s |      1.03 ops/s |     0.96 ops/s | 0.38 ops/s |
+
+**Stream column** = `generateWorkbookStream` (default compression, fully drained). Streaming trades throughput for pipeability: each part is compressed in a Web Worker as it is produced, so a fixed per-part worker handoff dominates small workbooks. Plain-data workbooks stream through a constant-memory path — worksheet rows serialize chunk-wise with inline strings, so the full worksheet XML and the archive never coexist (at 1M rows × 3 cols: peak RSS +177 MB streamed vs +810 MB buffered, and ~17% faster; at 100k × 20 it is also the fastest mode). Scenarios with images fall back to the full-memory stream.
 
 ## License