@gmod/bam 7.1.18 → 7.1.20

package/package.json

@@ -1,16 +1,17 @@
 {
   "name": "@gmod/bam",
-  "version": "7.1.18",
+  "version": "7.1.20",
   "description": "Parser for BAM and BAM index (bai) files",
   "license": "MIT",
-  "repository": "GMOD/bam-js",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/GMOD/bam-js.git"
+  },
   "type": "module",
-  "types": "./dist/index.d.ts",
+  "types": "./esm/index.d.ts",
   "exports": {
-    "import": {
-      "import": "./esm/index.js"
-    },
-    "require": {
+    ".": {
+      "import": "./esm/index.js",
       "require": "./dist/index.js"
     }
   },
@@ -20,7 +21,7 @@
     "url": "https://github.com/cmdcolin"
   },
   "engines": {
-    "node": ">=6"
+    "node": ">=20"
   },
   "files": [
     "dist",
@@ -35,11 +36,11 @@
     "clean": "rimraf dist esm",
     "format": "prettier --write .",
     "build:esm": "tsc --outDir esm",
-    "build:es5": "tsc --module commonjs --outDir dist",
-    "build": "yarn build:esm && yarn build:es5",
-    "prebuild": "yarn clean",
+    "build:es5": "tsc --module commonjs --moduleResolution bundler --outDir dist",
+    "build": "pnpm build:esm && pnpm build:es5",
+    "prebuild": "pnpm clean",
     "postbuild:es5": "echo '{\"type\": \"commonjs\"}' > dist/package.json",
-    "preversion": "yarn lint && yarn test --run && yarn build",
+    "preversion": "pnpm lint && pnpm test --run && pnpm build",
     "version": "standard-changelog && git add CHANGELOG.md",
     "postversion": "git push --follow-tags"
   },
@@ -53,21 +54,21 @@
     "@gmod/bgzf-filehandle": "^6.0.12",
     "@jbrowse/quick-lru": "^7.3.5",
     "crc": "^4.3.2",
-    "generic-filehandle2": "^2.0.18"
+    "generic-filehandle2": "^2.1.4"
   },
   "devDependencies": {
     "@eslint/js": "^10.0.1",
-    "@types/node": "^25.3.3",
-    "@vitest/coverage-v8": "^4.0.18",
-    "eslint": "^10.0.2",
-    "eslint-plugin-import": "^2.31.0",
-    "eslint-plugin-unicorn": "^63.0.0",
+    "@types/node": "^25.5.0",
+    "@vitest/coverage-v8": "^4.1.2",
+    "eslint": "^9.39.4",
+    "eslint-plugin-import": "^2.32.0",
+    "eslint-plugin-unicorn": "^64.0.0",
     "prettier": "^3.8.1",
     "rimraf": "^6.1.3",
     "standard-changelog": "^7.0.1",
-    "typescript": "^5.0.4",
-    "typescript-eslint": "^8.56.1",
-    "vitest": "^4.0.18"
+    "typescript": "^6.0.2",
+    "typescript-eslint": "^8.57.2",
+    "vitest": "^4.1.2"
   },
   "publishConfig": {
     "access": "public"

package/src/bamFile.ts

@@ -139,32 +139,49 @@ export default class BamFile<T extends BamRecordLike = BAMFeature> {
 
     // firstDataLine is not defined in cases where there is no data in the file
     // (just bam header and nothing else)
-    const buffer =
+    let readLen =
       indexData.firstDataLine === undefined
+        ? undefined
+        : indexData.firstDataLine.blockPosition + blockLen
+
+    const buffer =
+      readLen === undefined
         ? await this.bam.readFile()
-        : // the logic indexData.firstDataLine is a virtualOffset telling us
-          // where the data is. It is in the middle of a virtualOffset
-          // (provided by the bgzip block offset at blockPosition + the
-          // virtualOffset dataPosition, so we add one extra blockLen to make
-          // sure we consume the full header)
-          await this.bam.read(
-            indexData.firstDataLine.blockPosition + blockLen,
-            0,
-          )
-    const uncba = await unzip(buffer)
+        : await this.bam.read(readLen, 0)
+    let uncba = await unzip(buffer)
     const dataView = new DataView(uncba.buffer)
 
     if (dataView.getInt32(0, true) !== BAM_MAGIC) {
       throw new Error('Not a BAM file')
     }
     const headLen = dataView.getInt32(4, true)
-    const decoder = new TextDecoder('utf8')
-    this.header = decoder.decode(uncba.subarray(8, 8 + headLen))
-
-    const { chrToIndex, indexToChr } = this._parseRefSeqs(uncba, headLen + 8)
-    this.chrToIndex = chrToIndex
-    this.indexToChr = indexToChr
-    return parseHeaderText(this.header)
+    this.header = new TextDecoder('utf8').decode(uncba.subarray(8, 8 + headLen))
+
+    // BAM files with many reference sequences may need more data than the
+    // initial read provides, so re-fetch with doubled size until it fits
+    const refSeqStart = headLen + 8
+    const maxRetries = 5
+    for (let attempt = 0; attempt < maxRetries; attempt++) {
+      if (this._hasEnoughRefSeqData(uncba, refSeqStart)) {
+        const { chrToIndex, indexToChr } = this._parseRefSeqs(
+          uncba,
+          refSeqStart,
+        )
+        this.chrToIndex = chrToIndex
+        this.indexToChr = indexToChr
+        return parseHeaderText(this.header)
+      }
+      if (readLen === undefined) {
+        throw new Error(
+          `Insufficient data for reference sequences in ${uncba.length} bytes`,
+        )
+      }
+      readLen *= 2
+      uncba = await unzip(await this.bam.read(readLen, 0))
+    }
+    throw new Error(
+      `Insufficient data for reference sequences after ${maxRetries} retries`,
+    )
   }
 
   getHeader(opts?: BaseOpts) {
@@ -182,6 +199,23 @@ export default class BamFile<T extends BamRecordLike = BAMFeature> {
     return this.header
   }
 
+  _hasEnoughRefSeqData(uncba: Uint8Array, start: number) {
+    if (start + 4 > uncba.length) {
+      return false
+    }
+    const dataView = new DataView(uncba.buffer)
+    const nRef = dataView.getInt32(start, true)
+    let p = start + 4
+    for (let i = 0; i < nRef; i += 1) {
+      if (p + 8 > uncba.length) {
+        return false
+      }
+      const lName = dataView.getInt32(p, true)
+      p = p + 8 + lName
+    }
+    return true
+  }
+
   _parseRefSeqs(
     uncba: Uint8Array,
     start: number,
@@ -198,12 +232,6 @@ export default class BamFile<T extends BamRecordLike = BAMFeature> {
     const decoder = new TextDecoder('utf8')
 
     for (let i = 0; i < nRef; i += 1) {
-      if (p + 8 > uncba.length) {
-        throw new Error(
-          `Insufficient data for reference sequences: need more than ${uncba.length} bytes`,
-        )
-      }
-
       const lName = dataView.getInt32(p, true)
       const refName = this.renameRefSeq(
         decoder.decode(uncba.subarray(p + 4, p + 4 + lName - 1)),

package/src/nullFilehandle.ts

@@ -1,16 +1,16 @@
 export default class NullFilehandle {
-  public read(): Promise<any> {
+  public read(): Promise<never> {
     throw new Error('never called')
   }
-  public stat(): Promise<any> {
+  public stat(): Promise<never> {
     throw new Error('never called')
   }
 
-  public readFile(): Promise<any> {
+  public readFile(): Promise<never> {
     throw new Error('never called')
   }
 
-  public close(): Promise<any> {
+  public close(): Promise<never> {
     throw new Error('never called')
   }
 }

package/src/nullIndex.ts

@@ -1,18 +1,18 @@
 import IndexFile from './indexFile.ts'
 
 export default class NullIndex extends IndexFile {
-  public lineCount(): Promise<any> {
+  public lineCount(): Promise<never> {
     throw new Error('never called')
   }
-  protected _parse(): Promise<any> {
+  protected _parse(): Promise<never> {
     throw new Error('never called')
   }
 
-  public async indexCov(): Promise<any> {
+  public async indexCov(): Promise<never> {
     throw new Error('never called')
   }
 
-  public blocksForRange(): Promise<any> {
+  public blocksForRange(): Promise<never> {
     throw new Error('never called')
   }
 }

package/src/record.ts

@@ -674,10 +674,14 @@ export default class BamRecord {
   }
 
   // adapted from igv.js
-  // uses precomputed lookup table indexed by flag bits + isize sign
+  // uses precomputed lookup table indexed by flag bits + isize sign.
+  // the BAM spec defines tlen as positive for the leftmost segment and
+  // negative for the rightmost, so tlen > 0 reliably indicates which
+  // read comes first without needing position-based correction
+  // (see also: gmod/cram-js src/cramFile/record.ts getPairOrientation)
   get pair_orientation() {
     const f = this.flags
-    // combined check: unmapped (0x4) clear, mate unmapped (0x8) clear
+    // unmapped (0x4) or mate unmapped (0x8) -> undefined
     if (f & 0xc || this.ref_id !== this.next_refid) {
       return undefined
     }
@@ -727,7 +731,7 @@ export default class BamRecord {
   }
 
   toJSON() {
-    const data: Record<string, any> = {}
+    const data: Record<string, unknown> = {}
     for (const k of Object.keys(this)) {
       if (k.startsWith('_') || k === 'bytes') {
         continue