@loaders.gl/parquet 3.1.0-alpha.5 → 3.1.0-beta.5

package/dist/parquetjs/compression.js

@@ -0,0 +1,168 @@
+"use strict";
+// Forked from https://github.com/kbajalc/parquets under MIT license (Copyright (c) 2017 ironSource Ltd.)
+/* eslint-disable camelcase */
+// Forked from https://github.com/ironSource/parquetjs under MIT license
+var __importDefault = (this && this.__importDefault) || function (mod) {
+    return (mod && mod.__esModule) ? mod : { "default": mod };
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.inflate = exports.decompress = exports.deflate = exports.preloadCompressions = exports.PARQUET_COMPRESSION_METHODS = void 0;
+const compression_1 = require("@loaders.gl/compression");
+const buffer_utils_1 = require("./utils/buffer-utils");
+// TODO switch to worker compression to avoid bundling...
+// import brotli from 'brotli'; - brotli has problems with decompress in browsers
+// import brotliDecompress from 'brotli/decompress';
+const lz4js_1 = __importDefault(require("lz4js"));
+const lzo_1 = __importDefault(require("lzo"));
+// import {ZstdCodec} from 'zstd-codec';
+// Inject large dependencies through Compression constructor options
+const modules = {
+    // brotli has problems with decompress in browsers
+    // brotli: {
+    //   decompress: brotliDecompress,
+    //   compress: () => {
+    //     throw new Error('brotli compress');
+    //   }
+    // },
+    lz4js: lz4js_1.default,
+    lzo: lzo_1.default
+    // 'zstd-codec': ZstdCodec
+};
+// See https://github.com/apache/parquet-format/blob/master/Compression.md
+exports.PARQUET_COMPRESSION_METHODS = {
+    UNCOMPRESSED: new compression_1.NoCompression(),
+    GZIP: new compression_1.GZipCompression(),
+    SNAPPY: new compression_1.SnappyCompression(),
+    BROTLI: new compression_1.BrotliCompression({ modules }),
+    // TODO: Understand difference between LZ4 and LZ4_RAW
+    LZ4: new compression_1.LZ4Compression({ modules }),
+    LZ4_RAW: new compression_1.LZ4Compression({ modules }),
+    LZO: new compression_1.LZOCompression({ modules }),
+    ZSTD: new compression_1.ZstdCompression({ modules })
+};
+/**
+ * Register compressions that have big external libraries
+ * @param options.modules External library dependencies
+ */
+async function preloadCompressions(options) {
+    const compressions = Object.values(exports.PARQUET_COMPRESSION_METHODS);
+    return await Promise.all(compressions.map((compression) => compression.preload()));
+}
+exports.preloadCompressions = preloadCompressions;
+/**
+ * Deflate a value using compression method `method`
+ */
+async function deflate(method, value) {
+    const compression = exports.PARQUET_COMPRESSION_METHODS[method];
+    if (!compression) {
+        throw new Error(`parquet: invalid compression method: ${method}`);
+    }
+    const inputArrayBuffer = (0, buffer_utils_1.toArrayBuffer)(value);
+    const compressedArrayBuffer = await compression.compress(inputArrayBuffer);
+    return (0, buffer_utils_1.toBuffer)(compressedArrayBuffer);
+}
+exports.deflate = deflate;
+/**
+ * Inflate a value using compression method `method`
+ */
+async function decompress(method, value, size) {
+    const compression = exports.PARQUET_COMPRESSION_METHODS[method];
+    if (!compression) {
+        throw new Error(`parquet: invalid compression method: ${method}`);
+    }
+    const inputArrayBuffer = (0, buffer_utils_1.toArrayBuffer)(value);
+    const compressedArrayBuffer = await compression.decompress(inputArrayBuffer, size);
+    return (0, buffer_utils_1.toBuffer)(compressedArrayBuffer);
+}
+exports.decompress = decompress;
+/*
+ * Inflate a value using compression method `method`
+ */
+function inflate(method, value, size) {
+    if (!(method in exports.PARQUET_COMPRESSION_METHODS)) {
+        throw new Error(`invalid compression method: ${method}`);
+    }
+    // @ts-ignore
+    return exports.PARQUET_COMPRESSION_METHODS[method].inflate(value, size);
+}
+exports.inflate = inflate;
+/*
+function deflate_identity(value: Buffer): Buffer {
+  return value;
+}
+
+function deflate_gzip(value: Buffer): Buffer {
+  return zlib.gzipSync(value);
+}
+
+function deflate_snappy(value: Buffer): Buffer {
+  return snappyjs.compress(value);
+}
+
+function deflate_lzo(value: Buffer): Buffer {
+  lzo = lzo || Util.load('lzo');
+  return lzo.compress(value);
+}
+
+function deflate_brotli(value: Buffer): Buffer {
+  brotli = brotli || Util.load('brotli');
+  const result = brotli.compress(value, {
+    mode: 0,
+    quality: 8,
+    lgwin: 22
+  });
+  return result ? Buffer.from(result) : Buffer.alloc(0);
+}
+
+function deflate_lz4(value: Buffer): Buffer {
+  lz4js = lz4js || Util.load('lz4js');
+  try {
+    // let result = Buffer.alloc(lz4js.encodeBound(value.length));
+    // const compressedSize = lz4.encodeBlock(value, result);
+    // // remove unnecessary bytes
+    // result = result.slice(0, compressedSize);
+    // return result;
+    return Buffer.from(lz4js.compress(value));
+  } catch (err) {
+    throw err;
+  }
+}
+function inflate_identity(value: Buffer): Buffer {
+  return value;
+}
+
+function inflate_gzip(value: Buffer): Buffer {
+  return zlib.gunzipSync(value);
+}
+
+function inflate_snappy(value: Buffer): Buffer {
+  return snappyjs.uncompress(value);
+}
+
+function inflate_lzo(value: Buffer, size: number): Buffer {
+  lzo = lzo || Util.load('lzo');
+  return lzo.decompress(value, size);
+}
+
+function inflate_lz4(value: Buffer, size: number): Buffer {
+  lz4js = lz4js || Util.load('lz4js');
+  try {
+    // let result = Buffer.alloc(size);
+    // const uncompressedSize = lz4js.decodeBlock(value, result);
+    // // remove unnecessary bytes
+    // result = result.slice(0, uncompressedSize);
+    // return result;
+    return Buffer.from(lz4js.decompress(value, size));
+  } catch (err) {
+    throw err;
+  }
+}
+
+function inflate_brotli(value: Buffer): Buffer {
+  brotli = brotli || Util.load('brotli');
+  if (!value.length) {
+    return Buffer.alloc(0);
+  }
+  return Buffer.from(brotli.decompress(value));
+}
+*/

package/dist/parquetjs/encoder/writer.d.ts

@@ -120,3 +120,4 @@ export declare class ParquetTransformer<T> extends Transform {
     _transform(row: any, encoding: string, callback: (val?: any) => void): Promise<void>;
     _flush(callback: (val?: any) => void): Promise<void>;
 }
+//# sourceMappingURL=writer.d.ts.map

package/dist/parquetjs/encoder/writer.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"writer.d.ts","sourceRoot":"","sources":["../../../src/parquetjs/encoder/writer.ts"],"names":[],"mappings":";AAEA,OAAO,EAAC,SAAS,EAAE,QAAQ,EAAC,MAAM,QAAQ,CAAC;AAG3C,OAAO,EACL,aAAa,EAKd,MAAM,mBAAmB,CAAC;AAC3B,OAAO,EAAC,aAAa,EAAC,MAAM,kBAAkB,CAAC;AAE/C,OAAO,EAaL,QAAQ,EAGT,MAAM,mBAAmB,CAAC;AA2B3B,MAAM,WAAW,oBAAoB;IACnC,UAAU,CAAC,EAAE,MAAM,CAAC;IACpB,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,aAAa,CAAC,EAAE,OAAO,CAAC;IAGxB,KAAK,CAAC,EAAE,MAAM,CAAC;IACf,QAAQ,CAAC,EAAE,MAAM,CAAC;IAClB,EAAE,CAAC,EAAE,MAAM,CAAC;IACZ,IAAI,CAAC,EAAE,MAAM,CAAC;IACd,SAAS,CAAC,EAAE,OAAO,CAAC;IACpB,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB;AAED;;;;GAIG;AAEH,qBAAa,aAAa,CAAC,CAAC;IAC1B;;;OAGG;WACU,QAAQ,CAAC,CAAC,EACrB,MAAM,EAAE,aAAa,EACrB,IAAI,EAAE,MAAM,EACZ,IAAI,CAAC,EAAE,oBAAoB,GAC1B,OAAO,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC;IAK5B;;;OAGG;WACU,UAAU,CAAC,CAAC,EACvB,MAAM,EAAE,aAAa,EACrB,YAAY,EAAE,QAAQ,EACtB,IAAI,CAAC,EAAE,oBAAoB,GAC1B,OAAO,CAAC,aAAa,CAAC,CAAC,CAAC,CAAC;IAWrB,MAAM,EAAE,aAAa,CAAC;IACtB,cAAc,EAAE,qBAAqB,CAAC;IACtC,SAAS,EAAE,aAAa,CAAC;IACzB,YAAY,EAAE,MAAM,CAAC;IACrB,MAAM,EAAE,OAAO,CAAC;IAChB,YAAY,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;IAE5C;;OAEG;gBAED,MAAM,EAAE,aAAa,EACrB,cAAc,EAAE,qBAAqB,EACrC,IAAI,EAAE,oBAAoB;IActB,WAAW,IAAI,OAAO,CAAC,IAAI,CAAC;IAUlC;;;OAGG;IACG,SAAS,CAAC,CAAC,EAAE,GAAG,EAAE,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC;IAWzC;;;;;OAKG;IACG,KAAK,CAAC,QAAQ,CAAC,EAAE,MAAM,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC;IAqBjD;;OAEG;IACH,WAAW,CAAC,GAAG,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,GAAG,IAAI;IAK7C;;;;;OAKG;IACH,eAAe,CAAC,GAAG,EAAE,MAAM,GAAG,IAAI;IAIlC;;;OAGG;IACH,WAAW,CAAC,GAAG,EAAE,MAAM,GAAG,IAAI;CAG/B;AAED;;;;;GAKG;AACH,qBAAa,qBAAqB;IAChC;;OAEG;WACU,UAAU,CACrB,MAAM,EAAE,aAAa,EACrB,YAAY,EAAE,QAAQ,EACtB,IAAI,EAAE,oBAAoB,GACzB,OAAO,CAAC,qBAAqB,CAAC;IAM1B,MAAM,EAAE,aAAa,CAAC;IACtB,KAAK,EAAE,CAAC,GAAG,EAAE,MAAM,KAAK,OAAO,CAAC,IAAI,CAAC,CAAC;IACtC,KAAK,EAAE,MAAM,OAAO,CAAC,IAAI,CAAC,CAAC;IAC3B,MAAM,EAAE,MAAM,CAAC;IACf,QAAQ,EAAE,MAAM,CAAC;IACjB,SAAS,EAAE,QAAQ,EAAE,CAAC;IACtB,QAAQ,EAAE,MAAM,CAAC;IACjB,aAAa,EAAE,OAAO,CAAC;gBAG5B,MAAM,EAAE,aAAa,EACrB,OAAO,EAAE,CAAC,GAAG,EAAE,MAAM,KAAK,OAAO,CAAC,IAAI,CAAC,EACvC,OAAO,EAAE,MAAM,OAAO,CAAC,IAAI,CAAC,EAC5B,UAAU,EAAE,MAAM,EAClB,IAAI,EAAE,oBAAoB;IAY5B,YAAY,CAAC,GAAG,EAAE,MAAM,GAAG,OAAO,CAAC,IAAI,CAAC;IAKxC;;OAEG;IACH,WAAW,IAAI,OAAO,CAAC,IAAI,CAAC;IAI5B;;;OAGG;IACG,aAAa,CAAC,OAAO,EAAE,aAAa,GAAG,OAAO,CAAC,IAAI,CAAC;IAY1D;;OAEG;IACH,WAAW,CAAC,YAAY,EAAE,MAAM,CAAC,MAAM,EAAE,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC;IAWhE;;;OAGG;IACH,WAAW,CAAC,GAAG,EAAE,MAAM,GAAG,IAAI;CAG/B;AAED;;GAEG;AACH,qBAAa,kBAAkB,CAAC,CAAC,CAAE,SAAQ,SAAS;IAC3C,MAAM,EAAE,aAAa,CAAC,CAAC,CAAC,CAAC;gBAEpB,MAAM,EAAE,aAAa,EAAE,IAAI,GAAE,oBAAyB;IAiBlE,UAAU,CAAC,GAAG,EAAE,GAAG,EAAE,QAAQ,EAAE,MAAM,EAAE,QAAQ,EAAE,CAAC,GAAG,CAAC,EAAE,GAAG,KAAK,IAAI,GAAG,OAAO,CAAC,IAAI,CAAC;IAS9E,MAAM,CAAC,QAAQ,EAAE,CAAC,GAAG,CAAC,EAAE,GAAG,KAAK,IAAI;CAG3C"}

package/dist/parquetjs/encoder/writer.js

@@ -0,0 +1,478 @@
+"use strict";
+var __createBinding = (this && this.__createBinding) || (Object.create ? (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    Object.defineProperty(o, k2, { enumerable: true, get: function() { return m[k]; } });
+}) : (function(o, m, k, k2) {
+    if (k2 === undefined) k2 = k;
+    o[k2] = m[k];
+}));
+var __setModuleDefault = (this && this.__setModuleDefault) || (Object.create ? (function(o, v) {
+    Object.defineProperty(o, "default", { enumerable: true, value: v });
+}) : function(o, v) {
+    o["default"] = v;
+});
+var __importStar = (this && this.__importStar) || function (mod) {
+    if (mod && mod.__esModule) return mod;
+    var result = {};
+    if (mod != null) for (var k in mod) if (k !== "default" && Object.prototype.hasOwnProperty.call(mod, k)) __createBinding(result, mod, k);
+    __setModuleDefault(result, mod);
+    return result;
+};
+var __importDefault = (this && this.__importDefault) || function (mod) {
+    return (mod && mod.__esModule) ? mod : { "default": mod };
+};
+Object.defineProperty(exports, "__esModule", { value: true });
+exports.ParquetTransformer = exports.ParquetEnvelopeWriter = exports.ParquetWriter = void 0;
+// Forked from https://github.com/kbajalc/parquets under MIT license (Copyright (c) 2017 ironSource Ltd.)
+/* eslint-disable camelcase */
+const stream_1 = require("stream");
+const codecs_1 = require("../codecs");
+const Compression = __importStar(require("../compression"));
+const Shred = __importStar(require("../schema/shred"));
+const parquet_thrift_1 = require("../parquet-thrift");
+const file_utils_1 = require("../utils/file-utils");
+const read_utils_1 = require("../utils/read-utils");
+const node_int64_1 = __importDefault(require("node-int64"));
+/**
+ * Parquet File Magic String
+ */
+const PARQUET_MAGIC = 'PAR1';
+/**
+ * Parquet File Format Version
+ */
+const PARQUET_VERSION = 1;
+/**
+ * Default Page and Row Group sizes
+ */
+const PARQUET_DEFAULT_PAGE_SIZE = 8192;
+const PARQUET_DEFAULT_ROW_GROUP_SIZE = 4096;
+/**
+ * Repetition and Definition Level Encoding
+ */
+const PARQUET_RDLVL_TYPE = 'INT32';
+const PARQUET_RDLVL_ENCODING = 'RLE';
+/**
+ * Write a parquet file to an output stream. The ParquetWriter will perform
+ * buffering/batching for performance, so close() must be called after all rows
+ * are written.
+ */
+// eslint-disable-next-line @typescript-eslint/no-unused-vars
+class ParquetWriter {
+    /**
+     * Create a new buffered parquet writer for a given envelope writer
+     */
+    constructor(schema, envelopeWriter, opts) {
+        this.schema = schema;
+        this.envelopeWriter = envelopeWriter;
+        // @ts-ignore Row buffer typings...
+        this.rowBuffer = {};
+        this.rowGroupSize = opts.rowGroupSize || PARQUET_DEFAULT_ROW_GROUP_SIZE;
+        this.closed = false;
+        this.userMetadata = {};
+        // eslint-disable-next-line @typescript-eslint/no-floating-promises
+        this.writeHeader();
+    }
+    /**
+     * Convenience method to create a new buffered parquet writer that writes to
+     * the specified file
+     */
+    static async openFile(schema, path, opts) {
+        const outputStream = await (0, file_utils_1.osopen)(path, opts);
+        return ParquetWriter.openStream(schema, outputStream, opts);
+    }
+    /**
+     * Convenience method to create a new buffered parquet writer that writes to
+     * the specified stream
+     */
+    static async openStream(schema, outputStream, opts) {
+        if (!opts) {
+            // tslint:disable-next-line:no-parameter-reassignment
+            opts = {};
+        }
+        const envelopeWriter = await ParquetEnvelopeWriter.openStream(schema, outputStream, opts);
+        return new ParquetWriter(schema, envelopeWriter, opts);
+    }
+    async writeHeader() {
+        // TODO - better not mess with promises in the constructor
+        try {
+            await this.envelopeWriter.writeHeader();
+        }
+        catch (err) {
+            await this.envelopeWriter.close();
+            throw err;
+        }
+    }
+    /**
+     * Append a single row to the parquet file. Rows are buffered in memory until
+     * rowGroupSize rows are in the buffer or close() is called
+     */
+    async appendRow(row) {
+        if (this.closed) {
+            throw new Error('writer was closed');
+        }
+        Shred.shredRecord(this.schema, row, this.rowBuffer);
+        if (this.rowBuffer.rowCount >= this.rowGroupSize) {
+            // @ts-ignore
+            this.rowBuffer = {};
+        }
+    }
+    /**
+     * Finish writing the parquet file and commit the footer to disk. This method
+     * MUST be called after you are finished adding rows. You must not call this
+     * method twice on the same object or add any rows after the close() method has
+     * been called
+     */
+    async close(callback) {
+        if (this.closed) {
+            throw new Error('writer was closed');
+        }
+        this.closed = true;
+        if (this.rowBuffer.rowCount > 0 || this.rowBuffer.rowCount >= this.rowGroupSize) {
+            // @ts-ignore
+            this.rowBuffer = {};
+        }
+        await this.envelopeWriter.writeFooter(this.userMetadata);
+        await this.envelopeWriter.close();
+        // this.envelopeWriter = null;
+        if (callback) {
+            callback();
+        }
+    }
+    /**
+     * Add key<>value metadata to the file
+     */
+    setMetadata(key, value) {
+        // TODO: value to be any, obj -> JSON
+        this.userMetadata[String(key)] = String(value);
+    }
+    /**
+     * Set the parquet row group size. This values controls the maximum number
+     * of rows that are buffered in memory at any given time as well as the number
+     * of rows that are co-located on disk. A higher value is generally better for
+     * read-time I/O performance at the tradeoff of write-time memory usage.
+     */
+    setRowGroupSize(cnt) {
+        this.rowGroupSize = cnt;
+    }
+    /**
+     * Set the parquet data page size. The data page size controls the maximum
+     * number of column values that are written to disk as a consecutive array
+     */
+    setPageSize(cnt) {
+        this.envelopeWriter.setPageSize(cnt);
+    }
+}
+exports.ParquetWriter = ParquetWriter;
+/**
+ * Create a parquet file from a schema and a number of row groups. This class
+ * performs direct, unbuffered writes to the underlying output stream and is
+ * intendend for advanced and internal users; the writeXXX methods must be
+ * called in the correct order to produce a valid file.
+ */
+class ParquetEnvelopeWriter {
+    constructor(schema, writeFn, closeFn, fileOffset, opts) {
+        this.schema = schema;
+        this.write = writeFn;
+        this.close = closeFn;
+        this.offset = fileOffset;
+        this.rowCount = 0;
+        this.rowGroups = [];
+        this.pageSize = opts.pageSize || PARQUET_DEFAULT_PAGE_SIZE;
+        this.useDataPageV2 = 'useDataPageV2' in opts ? Boolean(opts.useDataPageV2) : false;
+    }
+    /**
+     * Create a new parquet envelope writer that writes to the specified stream
+     */
+    static async openStream(schema, outputStream, opts) {
+        const writeFn = file_utils_1.oswrite.bind(undefined, outputStream);
+        const closeFn = file_utils_1.osclose.bind(undefined, outputStream);
+        return new ParquetEnvelopeWriter(schema, writeFn, closeFn, 0, opts);
+    }
+    writeSection(buf) {
+        this.offset += buf.length;
+        return this.write(buf);
+    }
+    /**
+     * Encode the parquet file header
+     */
+    writeHeader() {
+        return this.writeSection(Buffer.from(PARQUET_MAGIC));
+    }
+    /**
+     * Encode a parquet row group. The records object should be created using the
+     * shredRecord method
+     */
+    async writeRowGroup(records) {
+        const rgroup = await encodeRowGroup(this.schema, records, {
+            baseOffset: this.offset,
+            pageSize: this.pageSize,
+            useDataPageV2: this.useDataPageV2
+        });
+        this.rowCount += records.rowCount;
+        this.rowGroups.push(rgroup.metadata);
+        return await this.writeSection(rgroup.body);
+    }
+    /**
+     * Write the parquet file footer
+     */
+    writeFooter(userMetadata) {
+        if (!userMetadata) {
+            // tslint:disable-next-line:no-parameter-reassignment
+            userMetadata = {};
+        }
+        return this.writeSection(encodeFooter(this.schema, this.rowCount, this.rowGroups, userMetadata));
+    }
+    /**
+     * Set the parquet data page size. The data page size controls the maximum
+     * number of column values that are written to disk as a consecutive array
+     */
+    setPageSize(cnt) {
+        this.pageSize = cnt;
+    }
+}
+exports.ParquetEnvelopeWriter = ParquetEnvelopeWriter;
+/**
+ * Create a parquet transform stream
+ */
+class ParquetTransformer extends stream_1.Transform {
+    constructor(schema, opts = {}) {
+        super({ objectMode: true });
+        const writeProxy = (function (t) {
+            return async function (b) {
+                t.push(b);
+            };
+        })(this);
+        this.writer = new ParquetWriter(schema, new ParquetEnvelopeWriter(schema, writeProxy, async () => { }, 0, opts), opts);
+    }
+    // tslint:disable-next-line:function-name
+    _transform(row, encoding, callback) {
+        if (row) {
+            return this.writer.appendRow(row).then(callback);
+        }
+        callback();
+        return Promise.resolve();
+    }
+    // tslint:disable-next-line:function-name
+    async _flush(callback) {
+        await this.writer.close(callback);
+    }
+}
+exports.ParquetTransformer = ParquetTransformer;
+/**
+ * Encode a consecutive array of data using one of the parquet encodings
+ */
+function encodeValues(type, encoding, values, opts) {
+    if (!(encoding in codecs_1.PARQUET_CODECS)) {
+        throw new Error(`invalid encoding: ${encoding}`);
+    }
+    return codecs_1.PARQUET_CODECS[encoding].encodeValues(type, values, opts);
+}
+/**
+ * Encode a parquet data page
+ */
+async function encodeDataPage(column, data) {
+    /* encode repetition and definition levels */
+    let rLevelsBuf = Buffer.alloc(0);
+    if (column.rLevelMax > 0) {
+        rLevelsBuf = encodeValues(PARQUET_RDLVL_TYPE, PARQUET_RDLVL_ENCODING, data.rlevels, {
+            bitWidth: (0, read_utils_1.getBitWidth)(column.rLevelMax)
+            // disableEnvelope: false
+        });
+    }
+    let dLevelsBuf = Buffer.alloc(0);
+    if (column.dLevelMax > 0) {
+        dLevelsBuf = encodeValues(PARQUET_RDLVL_TYPE, PARQUET_RDLVL_ENCODING, data.dlevels, {
+            bitWidth: (0, read_utils_1.getBitWidth)(column.dLevelMax)
+            // disableEnvelope: false
+        });
+    }
+    /* encode values */
+    const valuesBuf = encodeValues(column.primitiveType, column.encoding, data.values, {
+        typeLength: column.typeLength,
+        bitWidth: column.typeLength
+    });
+    const dataBuf = Buffer.concat([rLevelsBuf, dLevelsBuf, valuesBuf]);
+    // compression = column.compression === 'UNCOMPRESSED' ? (compression || 'UNCOMPRESSED') : column.compression;
+    const compressedBuf = await Compression.deflate(column.compression, dataBuf);
+    /* build page header */
+    const header = new parquet_thrift_1.PageHeader({
+        type: parquet_thrift_1.PageType.DATA_PAGE,
+        data_page_header: new parquet_thrift_1.DataPageHeader({
+            num_values: data.count,
+            encoding: parquet_thrift_1.Encoding[column.encoding],
+            definition_level_encoding: parquet_thrift_1.Encoding[PARQUET_RDLVL_ENCODING],
+            repetition_level_encoding: parquet_thrift_1.Encoding[PARQUET_RDLVL_ENCODING] // [PARQUET_RDLVL_ENCODING]
+        }),
+        uncompressed_page_size: dataBuf.length,
+        compressed_page_size: compressedBuf.length
+    });
+    /* concat page header, repetition and definition levels and values */
+    const headerBuf = (0, read_utils_1.serializeThrift)(header);
+    const page = Buffer.concat([headerBuf, compressedBuf]);
+    return { header, headerSize: headerBuf.length, page };
+}
+/**
+ * Encode a parquet data page (v2)
+ */
+async function encodeDataPageV2(column, data, rowCount) {
+    /* encode values */
+    const valuesBuf = encodeValues(column.primitiveType, column.encoding, data.values, {
+        typeLength: column.typeLength,
+        bitWidth: column.typeLength
+    });
+    // compression = column.compression === 'UNCOMPRESSED' ? (compression || 'UNCOMPRESSED') : column.compression;
+    const compressedBuf = await Compression.deflate(column.compression, valuesBuf);
+    /* encode repetition and definition levels */
+    let rLevelsBuf = Buffer.alloc(0);
+    if (column.rLevelMax > 0) {
+        rLevelsBuf = encodeValues(PARQUET_RDLVL_TYPE, PARQUET_RDLVL_ENCODING, data.rlevels, {
+            bitWidth: (0, read_utils_1.getBitWidth)(column.rLevelMax),
+            disableEnvelope: true
+        });
+    }
+    let dLevelsBuf = Buffer.alloc(0);
+    if (column.dLevelMax > 0) {
+        dLevelsBuf = encodeValues(PARQUET_RDLVL_TYPE, PARQUET_RDLVL_ENCODING, data.dlevels, {
+            bitWidth: (0, read_utils_1.getBitWidth)(column.dLevelMax),
+            disableEnvelope: true
+        });
+    }
+    /* build page header */
+    const header = new parquet_thrift_1.PageHeader({
+        type: parquet_thrift_1.PageType.DATA_PAGE_V2,
+        data_page_header_v2: new parquet_thrift_1.DataPageHeaderV2({
+            num_values: data.count,
+            num_nulls: data.count - data.values.length,
+            num_rows: rowCount,
+            encoding: parquet_thrift_1.Encoding[column.encoding],
+            definition_levels_byte_length: dLevelsBuf.length,
+            repetition_levels_byte_length: rLevelsBuf.length,
+            is_compressed: column.compression !== 'UNCOMPRESSED'
+        }),
+        uncompressed_page_size: rLevelsBuf.length + dLevelsBuf.length + valuesBuf.length,
+        compressed_page_size: rLevelsBuf.length + dLevelsBuf.length + compressedBuf.length
+    });
+    /* concat page header, repetition and definition levels and values */
+    const headerBuf = (0, read_utils_1.serializeThrift)(header);
+    const page = Buffer.concat([headerBuf, rLevelsBuf, dLevelsBuf, compressedBuf]);
+    return { header, headerSize: headerBuf.length, page };
+}
+/**
+ * Encode an array of values into a parquet column chunk
+ */
+async function encodeColumnChunk(column, buffer, offset, opts) {
+    const data = buffer.columnData[column.path.join()];
+    const baseOffset = (opts.baseOffset || 0) + offset;
+    /* encode data page(s) */
+    // const pages: Buffer[] = [];
+    let pageBuf;
+    // tslint:disable-next-line:variable-name
+    let total_uncompressed_size = 0;
+    // tslint:disable-next-line:variable-name
+    let total_compressed_size = 0;
+    {
+        const result = opts.useDataPageV2
+            ? await encodeDataPageV2(column, data, buffer.rowCount)
+            : await encodeDataPage(column, data);
+        // pages.push(result.page);
+        pageBuf = result.page;
+        total_uncompressed_size += result.header.uncompressed_page_size + result.headerSize;
+        total_compressed_size += result.header.compressed_page_size + result.headerSize;
+    }
+    // const pagesBuf = Buffer.concat(pages);
+    // const compression = column.compression === 'UNCOMPRESSED' ? (opts.compression || 'UNCOMPRESSED') : column.compression;
+    /* prepare metadata header */
+    const metadata = new parquet_thrift_1.ColumnMetaData({
+        path_in_schema: column.path,
+        num_values: data.count,
+        data_page_offset: baseOffset,
+        encodings: [],
+        total_uncompressed_size,
+        total_compressed_size,
+        type: parquet_thrift_1.Type[column.primitiveType],
+        codec: parquet_thrift_1.CompressionCodec[column.compression]
+    });
+    /* list encodings */
+    metadata.encodings.push(parquet_thrift_1.Encoding[PARQUET_RDLVL_ENCODING]);
+    metadata.encodings.push(parquet_thrift_1.Encoding[column.encoding]);
+    /* concat metadata header and data pages */
+    const metadataOffset = baseOffset + pageBuf.length;
+    const body = Buffer.concat([pageBuf, (0, read_utils_1.serializeThrift)(metadata)]);
+    return { body, metadata, metadataOffset };
+}
+/**
+ * Encode a list of column values into a parquet row group
+ */
+async function encodeRowGroup(schema, data, opts) {
+    const metadata = new parquet_thrift_1.RowGroup({
+        num_rows: data.rowCount,
+        columns: [],
+        total_byte_size: 0
+    });
+    let body = Buffer.alloc(0);
+    for (const field of schema.fieldList) {
+        if (field.isNested) {
+            continue; // eslint-disable-line no-continue
+        }
+        const cchunkData = await encodeColumnChunk(field, data, body.length, opts);
+        const cchunk = new parquet_thrift_1.ColumnChunk({
+            file_offset: cchunkData.metadataOffset,
+            meta_data: cchunkData.metadata
+        });
+        metadata.columns.push(cchunk);
+        metadata.total_byte_size = new node_int64_1.default(Number(metadata.total_byte_size) + cchunkData.body.length);
+        body = Buffer.concat([body, cchunkData.body]);
+    }
+    return { body, metadata };
+}
+/**
+ * Encode a parquet file metadata footer
+ */
+function encodeFooter(schema, rowCount, rowGroups, userMetadata) {
+    const metadata = new parquet_thrift_1.FileMetaData({
+        version: PARQUET_VERSION,
+        created_by: 'parquets',
+        num_rows: rowCount,
+        row_groups: rowGroups,
+        schema: [],
+        key_value_metadata: []
+    });
+    for (const key in userMetadata) {
+        const kv = new parquet_thrift_1.KeyValue({
+            key,
+            value: userMetadata[key]
+        });
+        metadata.key_value_metadata?.push?.(kv);
+    }
+    {
+        const schemaRoot = new parquet_thrift_1.SchemaElement({
+            name: 'root',
+            num_children: Object.keys(schema.fields).length
+        });
+        metadata.schema.push(schemaRoot);
+    }
+    for (const field of schema.fieldList) {
+        const relt = parquet_thrift_1.FieldRepetitionType[field.repetitionType];
+        const schemaElem = new parquet_thrift_1.SchemaElement({
+            name: field.name,
+            repetition_type: relt
+        });
+        if (field.isNested) {
+            schemaElem.num_children = field.fieldCount;
+        }
+        else {
+            schemaElem.type = parquet_thrift_1.Type[field.primitiveType];
+        }
+        if (field.originalType) {
+            schemaElem.converted_type = parquet_thrift_1.ConvertedType[field.originalType];
+        }
+        schemaElem.type_length = field.typeLength;
+        metadata.schema.push(schemaElem);
+    }
+    const metadataEncoded = (0, read_utils_1.serializeThrift)(metadata);
+    const footerEncoded = Buffer.alloc(metadataEncoded.length + 8);
+    metadataEncoded.copy(footerEncoded);
+    footerEncoded.writeUInt32LE(metadataEncoded.length, metadataEncoded.length);
+    footerEncoded.write(PARQUET_MAGIC, metadataEncoded.length + 4);
+    return footerEncoded;
+}

package/dist/parquetjs/file.d.ts

@@ -7,3 +7,4 @@ export declare function fclose(fd: any): Promise<unknown>;
 export declare function oswrite(os: any, buf: any): Promise<void>;
 export declare function osclose(os: any): Promise<void>;
 export declare function osopen(path: any, opts: any): Promise<unknown>;
+//# sourceMappingURL=file.d.ts.map

package/dist/parquetjs/file.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"file.d.ts","sourceRoot":"","sources":["../../src/parquetjs/file.ts"],"names":[],"mappings":";AACA,OAAO,EAAE,MAAM,IAAI,CAAC;AAEpB,wBAAgB,KAAK,CAAC,QAAQ,KAAA,oBAU7B;AAED,wBAAgB,KAAK,CAAC,QAAQ,KAAA,qBAU7B;AAED,wBAAgB,KAAK,CAAC,EAAE,KAAA,EAAE,QAAQ,KAAA,EAAE,MAAM,KAAA,oBAYzC;AAED,wBAAgB,MAAM,CAAC,EAAE,KAAA,oBAUxB;AAED,wBAAgB,OAAO,CAAC,EAAE,KAAA,EAAE,GAAG,KAAA,GAAG,OAAO,CAAC,IAAI,CAAC,CAU9C;AAED,wBAAgB,OAAO,CAAC,EAAE,KAAA,GAAG,OAAO,CAAC,IAAI,CAAC,CAUzC;AAED,wBAAgB,MAAM,CAAC,IAAI,KAAA,EAAE,IAAI,KAAA,oBAYhC"}