compressionStream
compressionStream returns a TransformStream<u8[], u8[]>: bytes written to its writable side leave its readable side compressed in the format you name. It never holds the whole input or output, so a file of any size is compressed with constant memory, and a slow reader holds the writer back instead of letting data pile up.
Signature
compressionStream(format: CompressFormat, options?: CompressOptions): TransformStream<u8[], u8[]>Minimal example
tyrus
import { CompressFormat, compressionStream } from "tyrus:compress";
import { readFileStream, writeFileStream } from "tyrus:fs";
import { TransformStream } from "tyrus:stream";
async function compressLog(): Promise<void> {
try {
const gzip: TransformStream<u8[], u8[]> = compressionStream(CompressFormat.Gzip);
const feeding = readFileStream("./app.log").pipeTo(gzip.writable);
const draining = gzip.readable.pipeTo(writeFileStream("./app.log.gz"));
await feeding;
await draining;
print("compressed"); // compressed
} catch (error: Error) {
printError(error.message);
}
}Every option
tyrus
import { CompressFormat, CompressOptions, compressionStream } from "tyrus:compress";
import { TransformStream } from "tyrus:stream";
async function compressChunks(): Promise<void> {
const brotli: TransformStream<u8[], u8[]> = compressionStream(CompressFormat.Brotli, {
level: 9, // 0 to 11 for Brotli, 0 to 9 for the others; default 6
});
await brotli.writable.write([116, 121, 114, 117, 115]);
await brotli.writable.close();
const first: u8[]? = await brotli.readable.read();
print((first ?? []).length > 0); // true
}Errors
A failure reaches both sides as a CompressError: read() on the output and write() or close() on the input reject with it, and catch (error: CompressError) receives it. A level outside the range of the format makes the stream fail on its first use with ERR_COMPRESS_INVALID_LEVEL; a literal one is refused when the program is compiled (E3312).
