@tokenizer/inflate - npm Package Compare versions

Comparing version 0.4.0 to 0.4.1

lib/GzipHandler.d.ts

+import type { ITokenizer } from 'strtok3';
+export declare class GzipHandler {
+    private tokenizer;
+    constructor(tokenizer: ITokenizer);
+    inflate(): ReadableStream<Uint8Array>;
+}

lib/GzipHandler.js

+export class GzipHandler {
+    constructor(tokenizer) {
+        this.tokenizer = tokenizer;
+    }
+    inflate() {
+        const tokenizer = this.tokenizer;
+        return new ReadableStream({
+            async pull(controller) {
+                const buffer = new Uint8Array(1024);
+                const size = await tokenizer.readBuffer(buffer, { mayBeLess: true });
+                if (size === 0) {
+                    controller.close();
+                    return;
+                }
+                controller.enqueue(buffer.subarray(0, size));
+            }
+        }).pipeThrough(new DecompressionStream("gzip"));
+    }
+}

lib/index.d.ts

+export { ZipHandler } from './ZipHandler.js';
+export { GzipHandler } from './GzipHandler.js';
+export type { ILocalFileHeader } from './ZipToken.js';

lib/index.js

+export { ZipHandler } from './ZipHandler.js';
+export { GzipHandler } from './GzipHandler.js';

lib/ZipHandler.d.ts

+import type { ITokenizer } from 'strtok3';
+import { type IFileHeader, type ILocalFileHeader } from "./ZipToken.js";
+export type InflateFileFilterResult = {
+    handler: InflatedDataHandler | false;
+    stop?: boolean;
+};
+export type { ILocalFileHeader } from './ZipToken.js';
+/**
+ * Return false when to ignore the file, return `InflatedDataHandler` to handle extracted data
+ */
+export type InflateFileFilter = (file: ILocalFileHeader) => InflateFileFilterResult;
+export type InflatedDataHandler = (fileData: Uint8Array) => Promise<void>;
+export declare class ZipHandler {
+    private tokenizer;
+    private syncBuffer;
+    constructor(tokenizer: ITokenizer);
+    isZip(): Promise<boolean>;
+    private peekSignature;
+    findEndOfCentralDirectoryLocator(): Promise<number>;
+    readCentralDirectory(): Promise<IFileHeader[] | undefined>;
+    unzip(fileCb: InflateFileFilter): Promise<void>;
+    private iterateOverCentralDirectory;
+    private inflate;
+    private static decompressDeflateRaw;
+    private readLocalFileHeader;
+}

lib/ZipHandler.js

+import { StringType, UINT32_LE } from 'token-types';
+import initDebug from 'debug';
+import { DataDescriptor, EndOfCentralDirectoryRecordToken, FileHeader, LocalFileHeaderToken, Signature } from "./ZipToken.js";
+function signatureToArray(signature) {
+    const signatureBytes = new Uint8Array(UINT32_LE.len);
+    UINT32_LE.put(signatureBytes, 0, signature);
+    return signatureBytes;
+}
+const debug = initDebug('tokenizer:inflate');
+const syncBufferSize = 256 * 1024;
+const ddSignatureArray = signatureToArray(Signature.DataDescriptor);
+const eocdSignatureBytes = signatureToArray(Signature.EndOfCentralDirectory);
+export class ZipHandler {
+    constructor(tokenizer) {
+        this.tokenizer = tokenizer;
+        this.syncBuffer = new Uint8Array(syncBufferSize);
+    }
+    async isZip() {
+        return await this.peekSignature() === Signature.LocalFileHeader;
+    }
+    peekSignature() {
+        return this.tokenizer.peekToken(UINT32_LE);
+    }
+    async findEndOfCentralDirectoryLocator() {
+        const randomReadTokenizer = this.tokenizer;
+        const chunkLength = Math.min(16 * 1024, randomReadTokenizer.fileInfo.size);
+        const buffer = this.syncBuffer.subarray(0, chunkLength);
+        await this.tokenizer.readBuffer(buffer, { position: randomReadTokenizer.fileInfo.size - chunkLength });
+        // Search the buffer from end to beginning for EOCD signature
+        // const signature = 0x06054b50;
+        for (let i = buffer.length - 4; i >= 0; i--) {
+            // Compare 4 bytes directly without calling readUInt32LE
+            if (buffer[i] === eocdSignatureBytes[0] &&
+                buffer[i + 1] === eocdSignatureBytes[1] &&
+                buffer[i + 2] === eocdSignatureBytes[2] &&
+                buffer[i + 3] === eocdSignatureBytes[3]) {
+                return randomReadTokenizer.fileInfo.size - chunkLength + i;
+            }
+        }
+        return -1;
+    }
+    async readCentralDirectory() {
+        if (!this.tokenizer.supportsRandomAccess()) {
+            debug('Cannot reading central-directory without random-read support');
+            return;
+        }
+        debug('Reading central-directory...');
+        const pos = this.tokenizer.position;
+        const offset = await this.findEndOfCentralDirectoryLocator();
+        if (offset > 0) {
+            debug('Central-directory 32-bit signature found');
+            const eocdHeader = await this.tokenizer.readToken(EndOfCentralDirectoryRecordToken, offset);
+            const files = [];
+            this.tokenizer.setPosition(eocdHeader.offsetOfStartOfCd);
+            for (let n = 0; n < eocdHeader.nrOfEntriesOfSize; ++n) {
+                const entry = await this.tokenizer.readToken(FileHeader);
+                if (entry.signature !== Signature.CentralFileHeader) {
+                    throw new Error('Expected Central-File-Header signature');
+                }
+                entry.filename = await this.tokenizer.readToken(new StringType(entry.filenameLength, 'utf-8'));
+                await this.tokenizer.ignore(entry.extraFieldLength);
+                await this.tokenizer.ignore(entry.fileCommentLength);
+                files.push(entry);
+                debug(`Add central-directory file-entry: n=${n + 1}/${files.length}: filename=${files[n].filename}`);
+            }
+            this.tokenizer.setPosition(pos);
+            return files;
+        }
+        this.tokenizer.setPosition(pos);
+    }
+    async unzip(fileCb) {
+        const entries = await this.readCentralDirectory();
+        if (entries) {
+            // Use Central Directory to iterate over files
+            return this.iterateOverCentralDirectory(entries, fileCb);
+        }
+        // Scan Zip files for local-file-header
+        let stop = false;
+        do {
+            const zipHeader = await this.readLocalFileHeader();
+            if (!zipHeader)
+                break;
+            const next = fileCb(zipHeader);
+            stop = !!next.stop;
+            let fileData;
+            await this.tokenizer.ignore(zipHeader.extraFieldLength);
+            if (zipHeader.dataDescriptor && zipHeader.compressedSize === 0) {
+                const chunks = [];
+                let len = syncBufferSize;
+                debug('Compressed-file-size unknown, scanning for next data-descriptor-signature....');
+                let nextHeaderIndex = -1;
+                while (nextHeaderIndex < 0 && len === syncBufferSize) {
+                    len = await this.tokenizer.peekBuffer(this.syncBuffer, { mayBeLess: true });
+                    nextHeaderIndex = indexOf(this.syncBuffer.subarray(0, len), ddSignatureArray);
+                    const size = nextHeaderIndex >= 0 ? nextHeaderIndex : len;
+                    if (next.handler) {
+                        const data = new Uint8Array(size);
+                        await this.tokenizer.readBuffer(data);
+                        chunks.push(data);
+                    }
+                    else {
+                        // Move position to the next header if found, skip the whole buffer otherwise
+                        await this.tokenizer.ignore(size);
+                    }
+                }
+                debug(`Found data-descriptor-signature at pos=${this.tokenizer.position}`);
+                if (next.handler) {
+                    await this.inflate(zipHeader, mergeArrays(chunks), next.handler);
+                }
+            }
+            else {
+                if (next.handler) {
+                    debug(`Reading compressed-file-data: ${zipHeader.compressedSize} bytes`);
+                    fileData = new Uint8Array(zipHeader.compressedSize);
+                    await this.tokenizer.readBuffer(fileData);
+                    await this.inflate(zipHeader, fileData, next.handler);
+                }
+                else {
+                    debug(`Ignoring compressed-file-data: ${zipHeader.compressedSize} bytes`);
+                    await this.tokenizer.ignore(zipHeader.compressedSize);
+                }
+            }
+            debug(`Reading data-descriptor at pos=${this.tokenizer.position}`);
+            if (zipHeader.dataDescriptor) {
+                // await this.tokenizer.ignore(DataDescriptor.len);
+                const dataDescriptor = await this.tokenizer.readToken(DataDescriptor);
+                if (dataDescriptor.signature !== 0x08074b50) {
+                    throw new Error(`Expected data-descriptor-signature at position ${this.tokenizer.position - DataDescriptor.len}`);
+                }
+            }
+        } while (!stop);
+    }
+    async iterateOverCentralDirectory(entries, fileCb) {
+        for (const fileHeader of entries) {
+            const next = fileCb(fileHeader);
+            if (next.handler) {
+                this.tokenizer.setPosition(fileHeader.relativeOffsetOfLocalHeader);
+                const zipHeader = await this.readLocalFileHeader();
+                if (zipHeader) {
+                    await this.tokenizer.ignore(zipHeader.extraFieldLength);
+                    const fileData = new Uint8Array(fileHeader.compressedSize);
+                    await this.tokenizer.readBuffer(fileData);
+                    await this.inflate(zipHeader, fileData, next.handler);
+                }
+            }
+            if (next.stop)
+                break;
+        }
+    }
+    async inflate(zipHeader, fileData, cb) {
+        if (zipHeader.compressedMethod === 0) {
+            // Stored (uncompressed)
+            return cb(fileData);
+        }
+        if (zipHeader.compressedMethod !== 8) {
+            throw new Error(`Unsupported ZIP compression method: ${zipHeader.compressedMethod}`);
+        }
+        debug(`Decompress filename=${zipHeader.filename}, compressed-size=${fileData.length}`);
+        const uncompressedData = await ZipHandler.decompressDeflateRaw(fileData);
+        return cb(uncompressedData);
+    }
+    static async decompressDeflateRaw(data) {
+        // Wrap Uint8Array in a ReadableStream without copying
+        const input = new ReadableStream({
+            start(controller) {
+                controller.enqueue(data);
+                controller.close();
+            }
+        });
+        const ds = new DecompressionStream("deflate-raw");
+        const output = input.pipeThrough(ds);
+        try {
+            // Collect decompressed bytes from the output stream
+            const response = new Response(output);
+            const buffer = await response.arrayBuffer();
+            return new Uint8Array(buffer);
+        }
+        catch (err) {
+            // Provide ZIP-specific error context
+            const message = err instanceof Error
+                ? `Failed to deflate ZIP entry: ${err.message}`
+                : "Unknown decompression error in ZIP entry";
+            throw new TypeError(message);
+        }
+    }
+    async readLocalFileHeader() {
+        const signature = await this.tokenizer.peekToken(UINT32_LE);
+        if (signature === Signature.LocalFileHeader) {
+            const header = await this.tokenizer.readToken(LocalFileHeaderToken);
+            header.filename = await this.tokenizer.readToken(new StringType(header.filenameLength, 'utf-8'));
+            return header;
+        }
+        if (signature === Signature.CentralFileHeader) {
+            return false;
+        }
+        if (signature === 0xE011CFD0) {
+            throw new Error('Encrypted ZIP');
+        }
+        throw new Error('Unexpected signature');
+    }
+}
+function indexOf(buffer, portion) {
+    const bufferLength = buffer.length;
+    const portionLength = portion.length;
+    // Return -1 if the portion is longer than the buffer
+    if (portionLength > bufferLength)
+        return -1;
+    // Search for the portion in the buffer
+    for (let i = 0; i <= bufferLength - portionLength; i++) {
+        let found = true;
+        for (let j = 0; j < portionLength; j++) {
+            if (buffer[i + j] !== portion[j]) {
+                found = false;
+                break;
+            }
+        }
+        if (found) {
+            return i; // Return the starting offset
+        }
+    }
+    return -1; // Not found
+}
+function mergeArrays(chunks) {
+    // Concatenate chunks into a single Uint8Array
+    const totalLength = chunks.reduce((acc, curr) => acc + curr.length, 0);
+    const mergedArray = new Uint8Array(totalLength);
+    let offset = 0;
+    for (const chunk of chunks) {
+        mergedArray.set(chunk, offset);
+        offset += chunk.length;
+    }
+    return mergedArray;
+}

lib/ZipToken.d.ts

+/**
+ * Ref https://pkware.cachefly.net/webdocs/casestudies/APPNOTE.TXT
+ */
+import type { IGetToken } from "strtok3";
+export declare const Signature: {
+    LocalFileHeader: number;
+    DataDescriptor: number;
+    CentralFileHeader: number;
+    EndOfCentralDirectory: number;
+};
+interface IDataDescriptor {
+    signature: number;
+    compressedSize: number;
+    uncompressedSize: number;
+}
+export interface ILocalFileHeader extends IDataDescriptor {
+    minVersion: number;
+    dataDescriptor: boolean;
+    compressedMethod: number;
+    compressedSize: number;
+    uncompressedSize: number;
+    filenameLength: number;
+    extraFieldLength: number;
+    filename: string;
+}
+export declare const DataDescriptor: IGetToken<IDataDescriptor>;
+/**
+ * First part of the ZIP Local File Header
+ * Offset | Bytes| Description
+ * -------|------+-------------------------------------------------------------------
+ *      0 |    4 | Signature (0x04034b50)
+ *      4 |    2 | Minimum version needed to extract
+ *      6 |    2 | Bit flag
+ *      8 |    2 | Compression method
+ *     10 |    2 | File last modification time (MS-DOS format)
+ *     12 |    2 | File last modification date (MS-DOS format)
+ *     14 |    4 | CRC-32 of uncompressed data
+ *     18 |    4 | Compressed size
+ *     22 |    4 | Uncompressed size
+ *     26 |    2 | File name length (n)
+ *     28 |    2 | Extra field length (m)
+ *     30 |    n | File name
+ * 30 + n |    m | Extra field
+ */
+export declare const LocalFileHeaderToken: IGetToken<ILocalFileHeader>;
+interface IEndOfCentralDirectoryRecord {
+    signature: number;
+    nrOfThisDisk: number;
+    nrOfThisDiskWithTheStart: number;
+    nrOfEntriesOnThisDisk: number;
+    nrOfEntriesOfSize: number;
+    sizeOfCd: number;
+    offsetOfStartOfCd: number;
+    zipFileCommentLength: number;
+}
+/**
+ * 4.3.16  End of central directory record:
+ *  end of central dir signature (0x06064b50)                                      4 bytes
+ *  number of this disk                                                            2 bytes
+ *  number of the disk with the start of the central directory                     2 bytes
+ *  total number of entries in the central directory on this disk                  2 bytes
+ *  total number of entries in the size of the central directory                   2 bytes
+ *  sizeOfTheCentralDirectory                                                      4 bytes
+ *  offset of start of central directory with respect to the starting disk number  4 bytes
+ *  .ZIP file comment length                                                       2 bytes
+ *  .ZIP file comment       (variable size)
+ */
+export declare const EndOfCentralDirectoryRecordToken: IGetToken<IEndOfCentralDirectoryRecord>;
+export interface IFileHeader extends ILocalFileHeader {
+    fileCommentLength: number;
+    relativeOffsetOfLocalHeader: number;
+}
+/**
+ * File header:
+ *    central file header signature   4 bytes   0 (0x02014b50)
+ *    version made by                 2 bytes   4
+ *    version needed to extract       2 bytes   6
+ *    general purpose bit flag        2 bytes   8
+ *    compression method              2 bytes  10
+ *    last mod file time              2 bytes  12
+ *    last mod file date              2 bytes  14
+ *    crc-32                          4 bytes  16
+ *    compressed size                 4 bytes  20
+ *    uncompressed size               4 bytes  24
+ *    file name length                2 bytes  28
+ *    extra field length              2 bytes  30
+ *    file comment length             2 bytes  32
+ *    disk number start               2 bytes  34
+ *    internal file attributes        2 bytes  36
+ *    external file attributes        4 bytes  38
+ *    relative offset of local header 4 bytes  42
+ */
+export declare const FileHeader: IGetToken<IFileHeader>;
+export {};

lib/ZipToken.js

+/**
+ * Ref https://pkware.cachefly.net/webdocs/casestudies/APPNOTE.TXT
+ */
+import { UINT16_LE, UINT32_LE } from "token-types";
+export const Signature = {
+    LocalFileHeader: 0x04034b50,
+    DataDescriptor: 0x08074b50,
+    CentralFileHeader: 0x02014b50,
+    EndOfCentralDirectory: 0x06054b50
+};
+export const DataDescriptor = {
+    get(array) {
+        return {
+            signature: UINT32_LE.get(array, 0),
+            compressedSize: UINT32_LE.get(array, 8),
+            uncompressedSize: UINT32_LE.get(array, 12),
+        };
+    }, len: 16
+};
+/**
+ * First part of the ZIP Local File Header
+ * Offset | Bytes| Description
+ * -------|------+-------------------------------------------------------------------
+ *      0 |    4 | Signature (0x04034b50)
+ *      4 |    2 | Minimum version needed to extract
+ *      6 |    2 | Bit flag
+ *      8 |    2 | Compression method
+ *     10 |    2 | File last modification time (MS-DOS format)
+ *     12 |    2 | File last modification date (MS-DOS format)
+ *     14 |    4 | CRC-32 of uncompressed data
+ *     18 |    4 | Compressed size
+ *     22 |    4 | Uncompressed size
+ *     26 |    2 | File name length (n)
+ *     28 |    2 | Extra field length (m)
+ *     30 |    n | File name
+ * 30 + n |    m | Extra field
+ */
+export const LocalFileHeaderToken = {
+    get(array) {
+        const flags = UINT16_LE.get(array, 6);
+        return {
+            signature: UINT32_LE.get(array, 0),
+            minVersion: UINT16_LE.get(array, 4),
+            dataDescriptor: !!(flags & 0x0008),
+            compressedMethod: UINT16_LE.get(array, 8),
+            compressedSize: UINT32_LE.get(array, 18),
+            uncompressedSize: UINT32_LE.get(array, 22),
+            filenameLength: UINT16_LE.get(array, 26),
+            extraFieldLength: UINT16_LE.get(array, 28),
+            filename: null
+        };
+    }, len: 30
+};
+/**
+ * 4.3.16  End of central directory record:
+ *  end of central dir signature (0x06064b50)                                      4 bytes
+ *  number of this disk                                                            2 bytes
+ *  number of the disk with the start of the central directory                     2 bytes
+ *  total number of entries in the central directory on this disk                  2 bytes
+ *  total number of entries in the size of the central directory                   2 bytes
+ *  sizeOfTheCentralDirectory                                                      4 bytes
+ *  offset of start of central directory with respect to the starting disk number  4 bytes
+ *  .ZIP file comment length                                                       2 bytes
+ *  .ZIP file comment       (variable size)
+ */
+export const EndOfCentralDirectoryRecordToken = {
+    get(array) {
+        return {
+            signature: UINT32_LE.get(array, 0),
+            nrOfThisDisk: UINT16_LE.get(array, 4),
+            nrOfThisDiskWithTheStart: UINT16_LE.get(array, 6),
+            nrOfEntriesOnThisDisk: UINT16_LE.get(array, 8),
+            nrOfEntriesOfSize: UINT16_LE.get(array, 10),
+            sizeOfCd: UINT32_LE.get(array, 12),
+            offsetOfStartOfCd: UINT32_LE.get(array, 16),
+            zipFileCommentLength: UINT16_LE.get(array, 20),
+        };
+    }, len: 22
+};
+/**
+ * File header:
+ *    central file header signature   4 bytes   0 (0x02014b50)
+ *    version made by                 2 bytes   4
+ *    version needed to extract       2 bytes   6
+ *    general purpose bit flag        2 bytes   8
+ *    compression method              2 bytes  10
+ *    last mod file time              2 bytes  12
+ *    last mod file date              2 bytes  14
+ *    crc-32                          4 bytes  16
+ *    compressed size                 4 bytes  20
+ *    uncompressed size               4 bytes  24
+ *    file name length                2 bytes  28
+ *    extra field length              2 bytes  30
+ *    file comment length             2 bytes  32
+ *    disk number start               2 bytes  34
+ *    internal file attributes        2 bytes  36
+ *    external file attributes        4 bytes  38
+ *    relative offset of local header 4 bytes  42
+ */
+export const FileHeader = {
+    get(array) {
+        const flags = UINT16_LE.get(array, 8);
+        return {
+            signature: UINT32_LE.get(array, 0),
+            minVersion: UINT16_LE.get(array, 6),
+            dataDescriptor: !!(flags & 0x0008),
+            compressedMethod: UINT16_LE.get(array, 10),
+            compressedSize: UINT32_LE.get(array, 20),
+            uncompressedSize: UINT32_LE.get(array, 24),
+            filenameLength: UINT16_LE.get(array, 28),
+            extraFieldLength: UINT16_LE.get(array, 30),
+            fileCommentLength: UINT16_LE.get(array, 32),
+            relativeOffsetOfLocalHeader: UINT32_LE.get(array, 42),
+            filename: null
+        };
+    }, len: 46
+};

package.json

 {
   "name": "@tokenizer/inflate",
-  "version": "0.4.0",
+  "version": "0.4.1",
   "description": "Tokenized zip support",
@@ -17,2 +17,3 @@ "type": "module",
     "build": "yarn run clean && yarn run compile",
+    "prepublishOnly": "yarn run build",
     "eslint": "eslint lib test",
@@ -19,0 +20,0 @@ "lint:md": "remark -u preset-lint-recommended .",

New alerts

Major refactor

Supply chain risk

Package has recently undergone a major refactor. It may be unstable or indicate significant internal changes. Use caution when updating to versions that include significant changes.

Found 1 instance in 1 package

Fixed alerts

Deprecated

Maintenance

The maintainer of the package marked it as deprecated. This could indicate that a single version should not be used, or that the package is no longer maintained and any new vulnerabilities will not be fixed.

Found 1 instance in 1 package

Empty package

Supply chain risk

Package does not contain any code. It may be removed, is name squatting, or the result of a faulty package publish.

Found 1 instance in 1 package

Major refactor

Supply chain risk

Package has recently undergone a major refactor. It may be unstable or indicate significant internal changes. Use caution when updating to versions that include significant changes.

Found 1 instance in 1 package

Improved metrics

Total package byte prevSize

28506

290.6%

Number of package files

11

266.67%

Lines of code

500

Infinity%

Number of medium maintenance alerts

0

-100%

Number of medium supply chain risk alerts

1

-50%

No dependency changes