@gmod/cram 2.0.4 → 3.0.0

package/dist/cramFile/sectionParsers.js

@@ -1,126 +1,209 @@
 "use strict";
 Object.defineProperty(exports, "__esModule", { value: true });
-exports.cramFileDefinition = void 0;
+exports.cramFileDefinition = cramFileDefinition;
+exports.cramBlockHeader = cramBlockHeader;
+exports.cramBlockCrc32 = cramBlockCrc32;
+exports.cramTagDictionary = cramTagDictionary;
+exports.cramPreservationMap = cramPreservationMap;
 exports.isMappedSliceHeader = isMappedSliceHeader;
 exports.getSectionParsers = getSectionParsers;
-const binary_parser_1 = require("@gmod/binary-parser");
-const singleItf8 = new binary_parser_1.Parser().itf8();
-const cramFileDefinition = {
-    parser: new binary_parser_1.Parser()
-        .string('magic', { length: 4 })
-        .uint8('majorVersion')
-        .uint8('minorVersion')
-        .string('fileId', { length: 20, stripNull: true }),
-    maxLength: 26,
-};
-exports.cramFileDefinition = cramFileDefinition;
-const cramBlockHeader = {
-    parser: new binary_parser_1.Parser()
-        .uint8('compressionMethod', {
-        formatter: /* istanbul ignore next */ /* istanbul ignore next */ b => {
-            const method = [
-                'raw',
-                'gzip',
-                'bzip2',
-                'lzma',
-                'rans',
-                'rans4x16',
-                'arith',
-                'fqzcomp',
-                'tok3',
-            ][b];
-            if (!method) {
-                throw new Error(`compression method number ${b} not implemented`);
-            }
-            return method;
+const util_1 = require("./util");
+function cramFileDefinition() {
+    return {
+        parser: (buffer, _startOffset = 0) => {
+            const b = buffer;
+            const dataView = new DataView(b.buffer, b.byteOffset, b.length);
+            let offset = 0;
+            const magic = buffer.subarray(offset, offset + 4).toString();
+            offset += 4;
+            const majorVersion = dataView.getUint8(offset);
+            offset += 1;
+            const minorVersion = dataView.getUint8(offset);
+            offset += 1;
+            const fileId = b
+                .subarray(offset, offset + 20)
+                .toString()
+                .replaceAll('\0', '');
+            offset += 20;
+            return {
+                value: {
+                    magic,
+                    majorVersion,
+                    minorVersion,
+                    fileId,
+                },
+                offset,
+            };
         },
-    })
-        .uint8('contentType', {
-        formatter: /* istanbul ignore next */ /* istanbul ignore next */ b => {
-            const type = [
-                'FILE_HEADER',
-                'COMPRESSION_HEADER',
-                'MAPPED_SLICE_HEADER',
-                'UNMAPPED_SLICE_HEADER', // < only used in cram v1
-                'EXTERNAL_DATA',
-                'CORE_DATA',
-            ][b];
-            if (!type) {
-                throw new Error(`invalid block content type id ${b}`);
-            }
-            return type;
+        maxLength: 26,
+    };
+}
+function cramBlockHeader() {
+    const parser = (buffer, _startOffset = 0) => {
+        const b = buffer;
+        const dataView = new DataView(b.buffer, b.byteOffset, b.length);
+        let offset = 0;
+        const d = dataView.getUint8(offset);
+        const compressionMethod = [
+            'raw',
+            'gzip',
+            'bzip2',
+            'lzma',
+            'rans',
+            'rans4x16',
+            'arith',
+            'fqzcomp',
+            'tok3',
+        ][d];
+        if (!compressionMethod) {
+            throw new Error(`compression method number ${d} not implemented`);
+        }
+        offset += 1;
+        const c = dataView.getUint8(offset);
+        const contentType = [
+            'FILE_HEADER',
+            'COMPRESSION_HEADER',
+            'MAPPED_SLICE_HEADER',
+            'UNMAPPED_SLICE_HEADER', // < only used in cram v1
+            'EXTERNAL_DATA',
+            'CORE_DATA',
+        ][c];
+        if (!contentType) {
+            throw new Error(`invalid block content type id ${c}`);
+        }
+        offset += 1;
+        const [contentId, newOffset1] = (0, util_1.parseItf8)(buffer, offset);
+        offset += newOffset1;
+        const [compressedSize, newOffset2] = (0, util_1.parseItf8)(buffer, offset);
+        offset += newOffset2;
+        const [uncompressedSize, newOffset3] = (0, util_1.parseItf8)(buffer, offset);
+        offset += newOffset3;
+        return {
+            offset,
+            value: {
+                uncompressedSize,
+                compressedSize,
+                contentId,
+                contentType: contentType,
+                compressionMethod: compressionMethod,
+            },
+        };
+    };
+    return { parser, maxLength: 17 };
+}
+function cramBlockCrc32() {
+    return {
+        parser: (buffer, offset) => {
+            const b = buffer;
+            const dataView = new DataView(b.buffer, b.byteOffset, b.length);
+            const crc32 = dataView.getUint32(offset, true);
+            offset += 4;
+            return {
+                offset,
+                value: {
+                    crc32,
+                },
+            };
         },
-    })
-        .itf8('contentId')
-        .itf8('compressedSize')
-        .itf8('uncompressedSize'),
-    maxLength: 17,
-};
-const cramBlockCrc32 = {
-    parser: new binary_parser_1.Parser().uint32('crc32'),
-    maxLength: 4,
-};
-const cramTagDictionary = new binary_parser_1.Parser().itf8('size').buffer('ents', {
-    length: 'size',
-    formatter: /* istanbul ignore next */ /* istanbul ignore next */ buffer => {
-        function makeTagSet(stringStart, stringEnd) {
-            const str = buffer.toString('utf8', stringStart, stringEnd);
-            const tags = [];
-            for (let i = 0; i < str.length; i += 3) {
-                tags.push(str.slice(i, i + 3));
+        maxLength: 4,
+    };
+}
+function makeTagSet(buffer, stringStart, stringEnd) {
+    const str = buffer.toString('utf8', stringStart, stringEnd);
+    const tags = [];
+    for (let i = 0; i < str.length; i += 3) {
+        tags.push(str.slice(i, i + 3));
+    }
+    return tags;
+}
+function cramTagDictionary() {
+    return {
+        parser: (buffer, offset) => {
+            const [size, newOffset1] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset1;
+            const subbuf = buffer.subarray(offset, offset + size);
+            offset += size;
+            const tagSets = [];
+            let stringStart = 0;
+            let i = 0;
+            for (; i < subbuf.length; i++) {
+                if (!subbuf[i]) {
+                    tagSets.push(makeTagSet(subbuf, stringStart, i));
+                    stringStart = i + 1;
+                }
             }
-            return tags;
-        }
-        /* eslint-disable */
-        var tagSets = [];
-        var stringStart = 0;
-        var i;
-        /* eslint-enable */
-        for (i = 0; i < buffer.length; i += 1) {
-            if (!buffer[i]) {
-                tagSets.push(makeTagSet(stringStart, i));
-                stringStart = i + 1;
+            if (i > stringStart) {
+                tagSets.push(makeTagSet(subbuf, stringStart, i));
             }
-        }
-        if (i > stringStart) {
-            tagSets.push(makeTagSet(stringStart, i));
-        }
-        return tagSets;
-    },
-});
-// const cramPreservationMapKeys = 'XX RN AP RR SM TD'.split(' ')
-const parseByteAsBool = new binary_parser_1.Parser().uint8(null, {
-    formatter: /* istanbul ignore next */ /* istanbul ignore next */ val => !!val,
-});
-const cramPreservationMap = new binary_parser_1.Parser()
-    .itf8('mapSize')
-    .itf8('mapCount')
-    .array('ents', {
-    length: 'mapCount',
-    type: new binary_parser_1.Parser()
-        .string('key', {
-        length: 2,
-        stripNull: false,
-        // formatter: val => cramPreservationMapKeys[val] || 0,
-    })
-        .choice('value', {
-        tag: 'key',
-        choices: {
-            MI: parseByteAsBool,
-            UI: parseByteAsBool,
-            PI: parseByteAsBool,
-            RN: parseByteAsBool,
-            AP: parseByteAsBool,
-            RR: parseByteAsBool,
-            SM: new binary_parser_1.Parser().array(null, { type: 'uint8', length: 5 }),
-            TD: new binary_parser_1.Parser().nest(null, {
-                type: cramTagDictionary,
-                formatter: /* istanbul ignore next */ /* istanbul ignore next */ data => data.ents,
-            }),
+            return {
+                value: {
+                    size,
+                    ents: tagSets,
+                },
+                offset,
+            };
+        },
+    };
+}
+function cramPreservationMap() {
+    return {
+        parser: (buffer, offset) => {
+            const b = buffer;
+            const dataView = new DataView(b.buffer, b.byteOffset, b.length);
+            const [mapSize, newOffset1] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset1;
+            const [mapCount, newOffset2] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset2;
+            const ents = [];
+            for (let i = 0; i < mapCount; i++) {
+                const key = String.fromCharCode(buffer[offset]) +
+                    String.fromCharCode(buffer[offset + 1]);
+                offset += 2;
+                if (key === 'MI' ||
+                    key === 'UI' ||
+                    key === 'PI' ||
+                    key === 'RN' ||
+                    key === 'AP' ||
+                    key === 'RR') {
+                    ents.push({
+                        key,
+                        value: !!dataView.getUint8(offset),
+                    });
+                    offset += 1;
+                }
+                else if (key === 'SM') {
+                    ents.push({
+                        key,
+                        value: [
+                            dataView.getUint8(offset),
+                            dataView.getUint8(offset + 1),
+                            dataView.getUint8(offset + 2),
+                            dataView.getUint8(offset + 3),
+                            dataView.getUint8(offset + 4),
+                        ],
+                    });
+                    offset += 5;
+                }
+                else if (key === 'TD') {
+                    const { offset: offsetRet, value } = cramTagDictionary().parser(buffer, offset);
+                    ents.push({ key, value: value.ents });
+                    offset = offsetRet;
+                }
+                else {
+                    throw new Error(`unknown key ${key}`);
+                }
+            }
+            return {
+                value: {
+                    mapSize,
+                    mapCount,
+                    ents,
+                },
+                offset,
+            };
         },
-    }),
-});
-/* istanbul ignore next */
+    };
+}
 function formatMap(data) {
     const map = {};
     for (const { key, value } of data.ents) {
@@ -131,211 +214,446 @@ function formatMap(data) {
     }
     return map;
 }
-const unversionedParsers = {
-    cramFileDefinition,
-    cramBlockHeader,
-    cramBlockCrc32,
-};
 function isMappedSliceHeader(header) {
     return typeof header.refSeqId === 'number';
 }
-// each of these is a function of the major and minor version
-const versionedParsers = {
-    // assemble a section parser for the unmapped slice header, with slight
-    // variations depending on the major version of the cram file
-    cramUnmappedSliceHeader(majorVersion) {
-        let maxLength = 0;
-        let parser = new binary_parser_1.Parser().itf8('numRecords');
-        maxLength += 5;
+// assemble a section parser for the unmapped slice header, with slight
+// variations depending on the major version of the cram file
+function cramUnmappedSliceHeader(majorVersion) {
+    let maxLength = 0;
+    maxLength += 5;
+    maxLength += 9;
+    maxLength += 5 * 2;
+    maxLength += 16;
+    const parser = (buffer, offset) => {
+        const [numRecords, newOffset1] = (0, util_1.parseItf8)(buffer, offset);
+        offset += newOffset1;
+        let recordCounter = 0;
         // recordCounter is itf8 in a CRAM v2 file, absent in CRAM v1
         if (majorVersion >= 3) {
-            parser = parser.ltf8('recordCounter');
-            maxLength += 9;
+            const [rc, newOffset2] = (0, util_1.parseLtf8)(buffer, offset);
+            offset += newOffset2;
+            recordCounter = rc;
         }
         else if (majorVersion === 2) {
-            parser = parser.itf8('recordCounter');
-            maxLength += 5;
-        }
-        parser = parser
-            .itf8('numBlocks')
-            .itf8('numContentIds')
-            .array('contentIds', {
-            type: singleItf8,
-            length: 'numContentIds',
-        });
-        maxLength += 5 * 2; // + numContentIds*5
-        // the md5 sum is missing in cram v1
-        if (majorVersion >= 2) {
-            parser = parser.array('md5', { type: 'uint8', length: 16 });
-            maxLength += 16;
+            const [rc, newOffset2] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset2;
+            recordCounter = rc;
         }
-        const maxLengthFunc = (numContentIds) => maxLength + numContentIds * 5;
-        return { parser, maxLength: maxLengthFunc }; // : p, maxLength: numContentIds => 5 + 9 + 5 * 2 + 5 * numContentIds + 16 }
-    },
-    // assembles a section parser for the unmapped slice header, with slight
-    // variations depending on the major version of the cram file
-    cramMappedSliceHeader(majorVersion) {
-        let parser = new binary_parser_1.Parser()
-            .itf8('refSeqId')
-            .itf8('refSeqStart')
-            .itf8('refSeqSpan')
-            .itf8('numRecords');
-        let maxLength = 5 * 4;
-        if (majorVersion >= 3) {
-            parser = parser.ltf8('recordCounter');
-            maxLength += 9;
+        else {
+            console.warn('recordCounter=0');
         }
-        else if (majorVersion === 2) {
-            parser = parser.itf8('recordCounter');
-            maxLength += 5;
+        const [numBlocks, newOffset3] = (0, util_1.parseItf8)(buffer, offset);
+        offset += newOffset3;
+        const [numContentIds, newOffset4] = (0, util_1.parseItf8)(buffer, offset);
+        offset += newOffset4;
+        const contentIds = [];
+        for (let i = 0; i < numContentIds; i++) {
+            const [id, newOffset5] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset5;
+            contentIds.push(id);
         }
-        parser = parser
-            .itf8('numBlocks')
-            .itf8('numContentIds')
-            .array('contentIds', {
-            type: singleItf8,
-            length: 'numContentIds',
-        })
-            .itf8('refBaseBlockId');
-        maxLength += 5 * 3;
         // the md5 sum is missing in cram v1
+        let md5;
         if (majorVersion >= 2) {
-            parser = parser.array('md5', { type: 'uint8', length: 16 });
-            maxLength += 16;
+            md5 = [...buffer.subarray(offset, offset + 16)];
+            offset += 16;
         }
-        const maxLengthFunc = (numContentIds) => maxLength + numContentIds * 5;
-        return { parser, maxLength: maxLengthFunc };
-    },
-    cramEncoding(_majorVersion) {
-        const parser = new binary_parser_1.Parser()
-            .namely('cramEncoding')
-            .itf8('codecId')
-            .itf8('parametersBytes')
-            .choice('parameters', {
-            tag: 'codecId',
-            choices: {
-                0: new binary_parser_1.Parser(), // NULL
-                1: new binary_parser_1.Parser().itf8('blockContentId'), // EXTERNAL
-                2: new binary_parser_1.Parser().itf8('offset').itf8('M'), // GOLOMB,
-                // HUFFMAN_INT
-                3: binary_parser_1.Parser.start()
-                    .itf8('numCodes')
-                    .array('symbols', { length: 'numCodes', type: singleItf8 })
-                    .itf8('numLengths')
-                    .array('bitLengths', { length: 'numLengths', type: singleItf8 }),
-                4: binary_parser_1.Parser.start() // BYTE_ARRAY_LEN
-                    .nest('lengthsEncoding', { type: 'cramEncoding' })
-                    .nest('valuesEncoding', { type: 'cramEncoding' }),
-                // BYTE_ARRAY_STOP is a little different for CRAM v1
-                5: new binary_parser_1.Parser().uint8('stopByte').itf8('blockContentId'),
-                6: new binary_parser_1.Parser().itf8('offset').itf8('length'), // BETA
-                7: new binary_parser_1.Parser().itf8('offset').itf8('K'), // SUBEXP
-                8: new binary_parser_1.Parser().itf8('offset').itf8('log2m'), // GOLOMB_RICE
-                9: new binary_parser_1.Parser().itf8('offset'), // GAMMA
+        return {
+            value: {
+                recordCounter,
+                md5,
+                contentIds,
+                numContentIds,
+                numBlocks,
+                numRecords,
             },
-        });
-        return { parser };
-    },
-    cramDataSeriesEncodingMap(majorVersion) {
-        return new binary_parser_1.Parser()
-            .itf8('mapSize')
-            .itf8('mapCount')
-            .array('ents', {
-            length: 'mapCount',
-            type: new binary_parser_1.Parser()
-                .string('key', { length: 2, stripNull: false })
-                .nest('value', { type: this.cramEncoding(majorVersion).parser }),
-        });
-    },
-    cramTagEncodingMap(majorVersion) {
-        return new binary_parser_1.Parser()
-            .itf8('mapSize')
-            .itf8('mapCount')
-            .array('ents', {
-            length: 'mapCount',
-            type: new binary_parser_1.Parser()
-                .itf8('key', {
-                formatter: /* istanbul ignore next */ /* istanbul ignore next */ integerRepresentation => 
-                /* istanbul ignore next */
-                String.fromCharCode((integerRepresentation >> 16) & 0xff) +
-                    String.fromCharCode((integerRepresentation >> 8) & 0xff) +
-                    String.fromCharCode(integerRepresentation & 0xff),
-            })
-                .nest('value', { type: this.cramEncoding(majorVersion).parser }),
-        });
-    },
-    cramCompressionHeader(majorVersion) {
-        let parser = new binary_parser_1.Parser();
-        // TODO: if we want to support CRAM v1, we will need to refactor
-        // compression header into 2 parts to parse the landmarks,
-        // like the container header
-        parser = parser
-            .nest('preservation', {
-            type: cramPreservationMap,
-            formatter: formatMap,
-        })
-            .nest('dataSeriesEncoding', {
-            type: this.cramDataSeriesEncodingMap(majorVersion),
-            formatter: formatMap,
-        })
-            .nest('tagEncoding', {
-            type: this.cramTagEncodingMap(majorVersion),
-            formatter: formatMap,
-        });
-        return { parser };
-    },
-    cramContainerHeader1(majorVersion) {
-        let parser = new binary_parser_1.Parser()
-            .int32('length') // byte size of the container data (blocks)
-            .itf8('refSeqId') // reference sequence identifier, -1 for unmapped reads, -2 for multiple reference sequences
-            .itf8('refSeqStart') // the alignment start position or 0 for unmapped reads
-            .itf8('alignmentSpan') // the length of the alignment or 0 for unmapped reads
-            .itf8('numRecords'); // number of records in the container
-        let maxLength = 4 + 5 * 4;
-        if (majorVersion >= 3) {
-            parser = parser.ltf8('recordCounter'); // 1-based sequential index of records in the file/stream.
-            maxLength += 9;
-        }
-        else if (majorVersion === 2) {
-            parser = parser.itf8('recordCounter');
-            maxLength += 5;
-        }
-        if (majorVersion > 1) {
-            parser = parser.ltf8('numBases'); // number of read bases
-            maxLength += 9;
+            offset,
+        };
+    };
+    return {
+        parser,
+        maxLength: (numContentIds) => maxLength + numContentIds * 5,
+    };
+}
+// assembles a section parser for the unmapped slice header, with slight
+// variations depending on the major version of the cram file
+function cramMappedSliceHeader(majorVersion) {
+    let maxLength = 0;
+    maxLength += 5 * 4; // EL0
+    maxLength += 9; // EL1
+    maxLength += 5 * 3; // EL2 ITF8s
+    maxLength += 16; // MD5
+    return {
+        parser: (buffer, offset) => {
+            // L0
+            const [refSeqId, newOffset1] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset1;
+            const [refSeqStart, newOffset2] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset2;
+            const [refSeqSpan, newOffset3] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset3;
+            const [numRecords, newOffset4] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset4;
+            // EL0
+            // L1
+            let recordCounter = 0;
+            if (majorVersion >= 3) {
+                const [rc, newOffset5] = (0, util_1.parseLtf8)(buffer, offset);
+                offset += newOffset5;
+                recordCounter = rc;
+            }
+            else if (majorVersion === 2) {
+                const [rc, newOffset5] = (0, util_1.parseItf8)(buffer, offset);
+                offset += newOffset5;
+                recordCounter = rc;
+            }
+            else {
+                console.warn('majorVersion is <2, recordCounter set to 0');
+            }
+            // EL1
+            // L2
+            const [numBlocks, newOffset6] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset6;
+            const [numContentIds, newOffset7] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset7;
+            const contentIds = [];
+            for (let i = 0; i < numContentIds; i++) {
+                const [id, newOffset5] = (0, util_1.parseItf8)(buffer, offset);
+                offset += newOffset5;
+                contentIds.push(id);
+            }
+            const [refBaseBlockId, newOffset8] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset8;
+            // EL2
+            // the md5 sum is missing in cram v1
+            let md5;
+            if (majorVersion >= 2) {
+                md5 = [...buffer.subarray(offset, offset + 16)];
+                offset += 16;
+            }
+            return {
+                value: {
+                    md5,
+                    numBlocks,
+                    numRecords,
+                    numContentIds,
+                    refSeqSpan,
+                    refSeqId,
+                    refSeqStart,
+                    recordCounter,
+                    refBaseBlockId,
+                    contentIds,
+                },
+                offset,
+            };
+        },
+        maxLength: (numContentIds) => maxLength + numContentIds * 5,
+    };
+}
+function cramEncoding() {
+    return {
+        parser: (buffer, offset) => cramEncodingSub(buffer, offset),
+    };
+}
+function cramEncodingSub(buffer, offset) {
+    const b = buffer;
+    const dataView = new DataView(b.buffer, b.byteOffset, b.length);
+    const [codecId, newOffset1] = (0, util_1.parseItf8)(buffer, offset);
+    offset += newOffset1;
+    const [parametersBytes, newOffset2] = (0, util_1.parseItf8)(buffer, offset);
+    offset += newOffset2;
+    const parameters = {};
+    if (codecId === 0) {
+        // NULL
+    }
+    else if (codecId === 1) {
+        // EXTERNAL
+        const [bc, newOffset3] = (0, util_1.parseItf8)(buffer, offset);
+        parameters.blockContentId = bc;
+        offset += newOffset3;
+    }
+    else if (codecId === 2) {
+        // GOLUMB
+        const [off, newOffset3] = (0, util_1.parseItf8)(buffer, offset);
+        parameters.offset = off;
+        offset += newOffset3;
+        const [M2, newOffset4] = (0, util_1.parseItf8)(buffer, offset);
+        parameters.M = M2;
+        offset += newOffset4;
+    }
+    else if (codecId === 3) {
+        // HUFFMAN_INT
+        const val = (0, util_1.parseItf8)(buffer, offset);
+        const numCodes = val[0];
+        offset += val[1];
+        const symbols = [];
+        for (let i = 0; i < numCodes; i++) {
+            const code = (0, util_1.parseItf8)(buffer, offset);
+            symbols.push(code[0]);
+            offset += code[1];
         }
-        parser = parser
-            .itf8('numBlocks') // the number of blocks
-            .itf8('numLandmarks'); // the number of landmarks
-        maxLength += 5 + 5;
-        return { parser, maxLength };
-    },
-    cramContainerHeader2(majorVersion) {
-        let parser = new binary_parser_1.Parser()
-            .itf8('numLandmarks') // the number of blocks
-            // Each integer value of this array is a byte offset
-            // into the blocks byte array. Landmarks are used for
-            // random access indexing.
-            .array('landmarks', {
-            type: new binary_parser_1.Parser().itf8(),
-            length: 'numLandmarks',
-        });
-        let crcLength = 0;
-        if (majorVersion >= 3) {
-            parser = parser.uint32('crc32');
-            crcLength = 4;
+        parameters.symbols = symbols;
+        const val2 = (0, util_1.parseItf8)(buffer, offset);
+        const numLengths = val[0];
+        parameters.numLengths = numLengths;
+        parameters.numCodes = numCodes;
+        parameters.numLengths = numLengths;
+        offset += val2[1];
+        const bitLengths = [];
+        for (let i = 0; i < numLengths; i++) {
+            const len = (0, util_1.parseItf8)(buffer, offset);
+            offset += len[1];
+            bitLengths.push(len[0]);
         }
-        return {
-            parser,
-            maxLength: (numLandmarks) => 5 + numLandmarks * 5 + crcLength,
-        };
-    },
-};
+        parameters.bitLengths = bitLengths;
+    }
+    else if (codecId === 4) {
+        // BYTE_ARRAY_LEN
+        const { value: lengthsEncoding, offset: newOffset1 } = cramEncodingSub(buffer, offset);
+        parameters.lengthsEncoding = lengthsEncoding;
+        offset = newOffset1;
+        const { value: valuesEncoding, offset: newOffset2 } = cramEncodingSub(buffer, offset);
+        parameters.valuesEncoding = valuesEncoding;
+        offset = newOffset2;
+    }
+    else if (codecId === 5) {
+        // BYTE_ARRAY_STOP
+        parameters.stopByte = dataView.getUint8(offset);
+        offset += 1;
+        const [blockContentId, newOffset1] = (0, util_1.parseItf8)(buffer, offset);
+        parameters.blockContentId = blockContentId;
+        offset += newOffset1;
+    }
+    else if (codecId === 6) {
+        // BETA
+        const [off, newOffset1] = (0, util_1.parseItf8)(buffer, offset);
+        parameters.offset = off;
+        offset += newOffset1;
+        const [len, newOffset2] = (0, util_1.parseItf8)(buffer, offset);
+        parameters.length = len;
+        offset += newOffset2;
+    }
+    else if (codecId === 7) {
+        // SUBEXP
+        const [off, newOffset1] = (0, util_1.parseItf8)(buffer, offset);
+        parameters.offset = off;
+        offset += newOffset1;
+        const [K, newOffset2] = (0, util_1.parseItf8)(buffer, offset);
+        parameters.K = K;
+        offset += newOffset2;
+    }
+    else if (codecId === 8) {
+        // GOLOMB_RICE
+        const [off, newOffset1] = (0, util_1.parseItf8)(buffer, offset);
+        parameters.offset = off;
+        offset += newOffset1;
+        const [l2m, newOffset2] = (0, util_1.parseItf8)(buffer, offset);
+        parameters.log2m = l2m;
+        offset += newOffset2;
+    }
+    else if (codecId === 9) {
+        // GAMMA
+        const [off, newOffset1] = (0, util_1.parseItf8)(buffer, offset);
+        parameters.offset = off;
+        offset += newOffset1;
+    }
+    else {
+        throw new Error(`unknown codecId ${codecId}`);
+    }
+    return {
+        value: {
+            codecId,
+            parametersBytes,
+            parameters,
+        },
+        offset,
+    };
+}
+function cramDataSeriesEncodingMap() {
+    return {
+        parser: (buffer, offset) => {
+            const [mapSize, newOffset1] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset1;
+            const [mapCount, newOffset2] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset2;
+            const ents = [];
+            for (let i = 0; i < mapCount; i++) {
+                const key = String.fromCharCode(buffer[offset]) +
+                    String.fromCharCode(buffer[offset + 1]);
+                offset += 2;
+                const { value, offset: newOffset4 } = cramEncodingSub(buffer, offset);
+                offset = newOffset4;
+                ents.push({ key, value });
+            }
+            return {
+                value: {
+                    mapSize,
+                    ents,
+                    mapCount,
+                },
+                offset,
+            };
+        },
+    };
+}
+function cramTagEncodingMap() {
+    return {
+        parser: (buffer, offset) => {
+            const [mapSize, newOffset1] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset1;
+            const [mapCount, newOffset2] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset2;
+            const ents = [];
+            for (let i = 0; i < mapCount; i++) {
+                const [k0, newOffset3] = (0, util_1.parseItf8)(buffer, offset);
+                offset += newOffset3;
+                const key = String.fromCharCode((k0 >> 16) & 0xff) +
+                    String.fromCharCode((k0 >> 8) & 0xff) +
+                    String.fromCharCode(k0 & 0xff);
+                const { value, offset: newOffset4 } = cramEncodingSub(buffer, offset);
+                offset = newOffset4;
+                ents.push({ key, value });
+            }
+            return {
+                value: {
+                    mapSize,
+                    ents,
+                    mapCount,
+                },
+                offset,
+            };
+        },
+    };
+}
+function cramCompressionHeader() {
+    return {
+        parser: (buffer, offset) => {
+            // TODO: if we want to support CRAM v1, we will need to refactor
+            // compression header into 2 parts to parse the landmarks, like the
+            // container header
+            const { value: preservation, offset: newOffset1 } = cramPreservationMap().parser(buffer, offset);
+            offset = newOffset1;
+            const { value: dataSeriesEncoding, offset: newOffset2 } = cramDataSeriesEncodingMap().parser(buffer, offset);
+            offset = newOffset2;
+            const { value: tagEncoding, offset: newOffset3 } = cramTagEncodingMap().parser(buffer, offset);
+            offset = newOffset3;
+            return {
+                value: {
+                    dataSeriesEncoding: formatMap(dataSeriesEncoding),
+                    preservation: formatMap(preservation),
+                    tagEncoding: formatMap(tagEncoding),
+                },
+                offset,
+            };
+        },
+    };
+}
+function cramContainerHeader1(majorVersion) {
+    let maxLength = 4;
+    maxLength += 5 * 4;
+    maxLength += 9;
+    maxLength += 9;
+    maxLength += 5 + 5;
+    return {
+        maxLength,
+        parser: (buffer, offset) => {
+            const b = buffer;
+            const dataView = new DataView(b.buffer, b.byteOffset, b.length);
+            // byte size of the container data (blocks)
+            const length = dataView.getInt32(offset, true);
+            offset += 4;
+            // reference sequence identifier, -1 for unmapped reads, -2 for multiple
+            // reference sequences
+            const [refSeqId, newOffset1] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset1;
+            const [refSeqStart, newOffset2] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset2;
+            const [alignmentSpan, newOffset3] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset3;
+            const [numRecords, newOffset4] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset4;
+            let recordCounter = 0;
+            if (majorVersion >= 3) {
+                const [rc, newOffset5] = (0, util_1.parseLtf8)(buffer, offset);
+                recordCounter = rc;
+                offset += newOffset5;
+            }
+            else if (majorVersion === 2) {
+                const [rc, newOffset5] = (0, util_1.parseItf8)(buffer, offset);
+                recordCounter = rc;
+                offset += newOffset5;
+            }
+            else {
+                console.warn('setting recordCounter=0');
+            }
+            let numBases;
+            if (majorVersion > 1) {
+                const [n, newOffset5] = (0, util_1.parseLtf8)(buffer, offset);
+                numBases = n;
+                offset += newOffset5;
+            }
+            const [numBlocks, newOffset6] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset6;
+            const [numLandmarks, newOffset7] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset7;
+            return {
+                value: {
+                    length,
+                    refSeqId,
+                    refSeqStart,
+                    alignmentSpan,
+                    numBlocks,
+                    numLandmarks,
+                    numBases,
+                    recordCounter,
+                    numRecords,
+                },
+                offset,
+            };
+        },
+    };
+}
+function cramContainerHeader2(majorVersion) {
+    return {
+        parser: (buffer, offset) => {
+            const b = buffer;
+            const dataView = new DataView(b.buffer, b.byteOffset, b.length);
+            const [numLandmarks, newOffset1] = (0, util_1.parseItf8)(buffer, offset);
+            offset += newOffset1;
+            const landmarks = [];
+            for (let i = 0; i < numLandmarks; i++) {
+                const [landmark, newOffset2] = (0, util_1.parseItf8)(buffer, offset);
+                offset += newOffset2;
+                landmarks.push(landmark);
+            }
+            let crc32;
+            if (majorVersion >= 3) {
+                crc32 = dataView.getUint32(offset, true);
+                offset += 4;
+            }
+            return {
+                value: Object.assign(Object.assign({}, (crc32 === undefined ? {} : { crc32 })), { numLandmarks,
+                    landmarks }),
+                offset,
+            };
+        },
+        maxLength: (numLandmarks) => 5 + 5 * numLandmarks + 4,
+    };
+}
 function getSectionParsers(majorVersion) {
-    const parsers = Object.assign({}, unversionedParsers);
-    Object.keys(versionedParsers).forEach(parserName => {
-        parsers[parserName] = versionedParsers[parserName](majorVersion);
-    });
-    return parsers;
+    return {
+        cramFileDefinition: cramFileDefinition(),
+        cramBlockHeader: cramBlockHeader(),
+        cramBlockCrc32: cramBlockCrc32(),
+        cramDataSeriesEncodingMap: cramDataSeriesEncodingMap(),
+        cramTagEncodingMap: cramTagEncodingMap(),
+        cramCompressionHeader: cramCompressionHeader(),
+        cramEncoding: cramEncoding(),
+        cramUnmappedSliceHeader: cramUnmappedSliceHeader(majorVersion),
+        cramMappedSliceHeader: cramMappedSliceHeader(majorVersion),
+        cramContainerHeader1: cramContainerHeader1(majorVersion),
+        cramContainerHeader2: cramContainerHeader2(majorVersion),
+    };
 }
 //# sourceMappingURL=sectionParsers.js.map