@gmod/bam 7.6.1 → 7.6.2

package/esm/htsget.js

@@ -1,91 +1,111 @@
 import { unzip } from '@gmod/bgzf-filehandle';
 import BamFile, { BAM_MAGIC } from "./bamFile.js";
 import Chunk from "./chunk.js";
-import { parseHeaderText } from "./sam.js";
-import { appendInRange, concatUint8Array } from "./util.js";
+import { appendInRange, concatUint8Array, parseRefSeqs } from "./util.js";
 import { VirtualOffset } from "./virtualOffset.js";
-async function fetchOk(url, opts) {
-    const res = await fetch(url, opts);
+// Errors carry a JSON body naming the error type, e.g. {"htsget":
+// {"error":"InvalidAuthentication","message":"..."}}. Surface that rather than
+// the raw payload, since 401/403 are the first thing a misconfigured token hits.
+function htsgetErrorMessage(text) {
+    let parsed;
+    try {
+        parsed = JSON.parse(text);
+    }
+    catch {
+        // not JSON, fall through to the raw body
+    }
+    const err = parsed?.htsget;
+    return err?.error === undefined
+        ? text
+        : err.message === undefined
+            ? err.error
+            : `${err.error}: ${err.message}`;
+}
+/**
+ * Offset of the first alignment record in the concatenated ticket blocks.
+ *
+ * The spec has the client concatenate the blocks in ticket order to get a
+ * complete data stream, so whether the header is its own block is up to the
+ * server: htsnexus splits it into a leading `data:` url, while htsget-rs
+ * returns header and records together in a single url. Either way the
+ * concatenation starts with the BAM header, so seek past it rather than
+ * dropping blocks — a url count or class attribute can't tell the two layouts
+ * apart. Returns 0 when the stream has no header, which is what a server
+ * sending body-only blocks produces.
+ */
+function recordsOffset(uncba) {
+    const dataView = new DataView(uncba.buffer, uncba.byteOffset, uncba.byteLength);
+    if (uncba.byteLength < 8 || dataView.getInt32(0, true) !== BAM_MAGIC) {
+        return 0;
+    }
+    const refs = parseRefSeqs(uncba, 8 + dataView.getInt32(4, true), n => n);
+    if (!refs) {
+        throw new Error('truncated BAM header in htsget response');
+    }
+    return refs.end;
+}
+async function fetchOk(fetchFn, url, opts) {
+    const res = await fetchFn(url, opts);
     if (!res.ok) {
-        throw new Error(`HTTP ${res.status} fetching ${url}: ${await res.text()}`);
+        throw new Error(`HTTP ${res.status} fetching ${url}: ${htsgetErrorMessage(await res.text())}`);
     }
     return res;
 }
-async function fetchChunk({ url, headers }, opts) {
+async function fetchChunk(fetchFn, { url, headers }, opts) {
     // pass base64 data URLs straight to fetch; otherwise apply headers (minus
     // referer, which isn't a permitted client-set header).
     // https://stackoverflow.com/a/54123275/2129219
     const { referer: _referer, ...rest } = headers ?? {};
     const res = url.startsWith('data:')
-        ? await fetchOk(url)
-        : await fetchOk(url, { ...opts, headers: rest });
+        ? await fetchOk(fetchFn, url)
+        : await fetchOk(fetchFn, url, { ...opts, headers: rest });
     return new Uint8Array(await res.arrayBuffer());
 }
-async function fetchAndConcat(arr, opts) {
+async function fetchAndConcat(fetchFn, arr, opts) {
     // Pipeline unzip after each fetch so decompression overlaps later fetches.
-    return concatUint8Array(await Promise.all(arr.map(async (c) => unzip(await fetchChunk(c, opts)))));
+    return concatUint8Array(await Promise.all(arr.map(async (c) => unzip(await fetchChunk(fetchFn, c, opts)))));
 }
 export default class HtsgetFile extends BamFile {
     baseUrl;
     trackId;
+    fetchFn;
     constructor(args) {
         super({ htsget: true, recordClass: args.recordClass });
         this.baseUrl = args.baseUrl;
         this.trackId = args.trackId;
+        this.fetchFn = args.fetch ?? ((input, init) => fetch(input, init));
+    }
+    /**
+     * Requests a ticket and returns its data blocks decompressed and
+     * concatenated, which per the spec is a complete BAM stream.
+     */
+    async fetchTicket(query, opts) {
+        const url = `${this.baseUrl}/${this.trackId}?${query}`;
+        const res = await fetchOk(this.fetchFn, url, { signal: opts?.signal });
+        const ticket = await res.json();
+        return fetchAndConcat(this.fetchFn, ticket.htsget.urls, {
+            signal: opts?.signal,
+        });
     }
     async getRecordsForRange(chr, min, max, opts) {
         await this.getHeader(opts);
-        const base = `${this.baseUrl}/${this.trackId}`;
-        const url = `${base}?referenceName=${chr}&start=${min}&end=${max}&format=BAM`;
         const chrId = this.chrToIndex?.[chr];
         if (chrId === undefined) {
             return [];
         }
-        const result = await fetchOk(url, opts);
-        const data = await result.json();
-        const uncba = await fetchAndConcat(data.htsget.urls.slice(1), {
-            signal: opts?.signal,
-        });
+        const uncba = await this.fetchTicket(`referenceName=${chr}&start=${min}&end=${max}&format=BAM`, opts);
         const zero = new VirtualOffset(0, 0);
-        const allRecords = this.readBamFeatures(uncba, [], [], new Chunk(zero, zero, 0));
-        return appendInRange(allRecords, chrId, min, max);
+        const records = this.readBamFeatures(uncba.subarray(recordsOffset(uncba)), [], [], new Chunk(zero, zero, 0));
+        return appendInRange(records, chrId, min, max);
     }
     async getHeaderPre(opts = {}) {
-        const url = `${this.baseUrl}/${this.trackId}?referenceName=na&class=header`;
-        const result = await fetchOk(url, opts);
-        const data = await result.json();
-        const uncba = await fetchAndConcat(data.htsget.urls, {
-            signal: opts.signal,
-        });
-        const dataView = new DataView(uncba.buffer, uncba.byteOffset, uncba.byteLength);
-        if (dataView.getInt32(0, true) !== BAM_MAGIC) {
-            throw new Error('Not a BAM file');
-        }
-        const headLen = dataView.getInt32(4, true);
-        const decoder = new TextDecoder();
-        const headerText = decoder.decode(uncba.subarray(8, 8 + headLen));
-        const samHeader = parseHeaderText(headerText);
-        // use the @SQ lines in the header to figure out the
-        // mapping between ref ref ID numbers and names
-        const idToName = [];
-        const nameToId = {};
-        const sqLines = samHeader.filter(l => l.tag === 'SQ');
-        for (const [refId, sqLine] of sqLines.entries()) {
-            let refName = '';
-            let length = 0;
-            for (const item of sqLine.data) {
-                if (item.tag === 'SN') {
-                    refName = item.value;
-                }
-                else if (item.tag === 'LN') {
-                    length = +item.value;
-                }
-            }
-            nameToId[refName] = refId;
-            idToName[refId] = { refName, length };
+        // format is the only parameter the spec permits alongside class=header;
+        // servers SHOULD reject anything else with InvalidInput
+        const uncba = await this.fetchTicket('class=header&format=BAM', opts);
+        const samHeader = this.applyHeader(uncba);
+        if (!samHeader) {
+            throw new Error('Insufficient data for reference sequences');
         }
-        this.chrToIndex = nameToId;
-        this.indexToChr = idToName;
         return samHeader;
     }
 }

package/esm/htsget.js.map

@@ -1 +1 @@
-{"version":3,"file":"htsget.js","sourceRoot":"","sources":["../src/htsget.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,EAAE,MAAM,uBAAuB,CAAA;AAE7C,OAAO,OAAO,EAAE,EAAE,SAAS,EAAE,MAAM,cAAc,CAAA;AACjD,OAAO,KAAK,MAAM,YAAY,CAAA;AAC9B,OAAO,EAAE,eAAe,EAAE,MAAM,UAAU,CAAA;AAC1C,OAAO,EAAE,aAAa,EAAE,gBAAgB,EAAE,MAAM,WAAW,CAAA;AAC3D,OAAO,EAAE,aAAa,EAAE,MAAM,oBAAoB,CAAA;AAWlD,KAAK,UAAU,OAAO,CAAC,GAAW,EAAE,IAAkB;IACpD,MAAM,GAAG,GAAG,MAAM,KAAK,CAAC,GAAG,EAAE,IAAI,CAAC,CAAA;IAClC,IAAI,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC;QACZ,MAAM,IAAI,KAAK,CAAC,QAAQ,GAAG,CAAC,MAAM,aAAa,GAAG,KAAK,MAAM,GAAG,CAAC,IAAI,EAAE,EAAE,CAAC,CAAA;IAC5E,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC;AAED,KAAK,UAAU,UAAU,CAAC,EAAE,GAAG,EAAE,OAAO,EAAe,EAAE,IAAkB;IACzE,0EAA0E;IAC1E,uDAAuD;IACvD,+CAA+C;IAC/C,MAAM,EAAE,OAAO,EAAE,QAAQ,EAAE,GAAG,IAAI,EAAE,GAAG,OAAO,IAAI,EAAE,CAAA;IACpD,MAAM,GAAG,GAAG,GAAG,CAAC,UAAU,CAAC,OAAO,CAAC;QACjC,CAAC,CAAC,MAAM,OAAO,CAAC,GAAG,CAAC;QACpB,CAAC,CAAC,MAAM,OAAO,CAAC,GAAG,EAAE,EAAE,GAAG,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAA;IAClD,OAAO,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,WAAW,EAAE,CAAC,CAAA;AAChD,CAAC;AAED,KAAK,UAAU,cAAc,CAAC,GAAkB,EAAE,IAAkB;IAClE,2EAA2E;IAC3E,OAAO,gBAAgB,CACrB,MAAM,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,GAAG,CAAC,KAAK,EAAC,CAAC,EAAC,EAAE,CAAC,KAAK,CAAC,MAAM,UAAU,CAAC,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC,CAAC,CACxE,CAAA;AACH,CAAC;AAED,MAAM,CAAC,OAAO,OAAO,UAEnB,SAAQ,OAAU;IACV,OAAO,CAAQ;IAEf,OAAO,CAAQ;IAEvB,YAAY,IAIX;QACC,KAAK,CAAC,EAAE,MAAM,EAAE,IAAI,EAAE,WAAW,EAAE,IAAI,CAAC,WAAW,EAAE,CAAC,CAAA;QACtD,IAAI,CAAC,OAAO,GAAG,IAAI,CAAC,OAAO,CAAA;QAC3B,IAAI,CAAC,OAAO,GAAG,IAAI,CAAC,OAAO,CAAA;IAC7B,CAAC;IAED,KAAK,CAAC,kBAAkB,CACtB,GAAW,EACX,GAAW,EACX,GAAW,EACX,IAAc;QAEd,MAAM,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,CAAA;QAC1B,MAAM,IAAI,GAAG,GAAG,IAAI,CAAC,OAAO,IAAI,IAAI,CAAC,OAAO,EAAE,CAAA;QAC9C,MAAM,GAAG,GAAG,GAAG,IAAI,kBAAkB,GAAG,UAAU,GAAG,QAAQ,GAAG,aAAa,CAAA;QAC7E,MAAM,KAAK,GAAG,IAAI,CAAC,UAAU,EAAE,CAAC,GAAG,CAAC,CAAA;QACpC,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YACxB,OAAO,EAAE,CAAA;QACX,CAAC;QACD,MAAM,MAAM,GAAG,MAAM,OAAO,CAAC,GAAG,EAAE,IAAI,CAAC,CAAA;QACvC,MAAM,IAAI,GAAG,MAAM,MAAM,CAAC,IAAI,EAAE,CAAA;QAChC,MAAM,KAAK,GAAG,MAAM,cAAc,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,EAAE;YAC5D,MAAM,EAAE,IAAI,EAAE,MAAM;SACrB,CAAC,CAAA;QAEF,MAAM,IAAI,GAAG,IAAI,aAAa,CAAC,CAAC,EAAE,CAAC,CAAC,CAAA;QACpC,MAAM,UAAU,GAAG,IAAI,CAAC,eAAe,CACrC,KAAK,EACL,EAAE,EACF,EAAE,EACF,IAAI,KAAK,CAAC,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC,CACzB,CAAA;QAED,OAAO,aAAa,CAAC,UAAU,EAAE,KAAK,EAAE,GAAG,EAAE,GAAG,CAAC,CAAA;IACnD,CAAC;IAED,KAAK,CAAC,YAAY,CAAC,OAAiB,EAAE;QACpC,MAAM,GAAG,GAAG,GAAG,IAAI,CAAC,OAAO,IAAI,IAAI,CAAC,OAAO,gCAAgC,CAAA;QAC3E,MAAM,MAAM,GAAG,MAAM,OAAO,CAAC,GAAG,EAAE,IAAI,CAAC,CAAA;QACvC,MAAM,IAAI,GAAG,MAAM,MAAM,CAAC,IAAI,EAAE,CAAA;QAChC,MAAM,KAAK,GAAG,MAAM,cAAc,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,EAAE;YACnD,MAAM,EAAE,IAAI,CAAC,MAAM;SACpB,CAAC,CAAA;QACF,MAAM,QAAQ,GAAG,IAAI,QAAQ,CAC3B,KAAK,CAAC,MAAM,EACZ,KAAK,CAAC,UAAU,EAChB,KAAK,CAAC,UAAU,CACjB,CAAA;QAED,IAAI,QAAQ,CAAC,QAAQ,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;YAC7C,MAAM,IAAI,KAAK,CAAC,gBAAgB,CAAC,CAAA;QACnC,CAAC;QACD,MAAM,OAAO,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAC,EAAE,IAAI,CAAC,CAAA;QAE1C,MAAM,OAAO,GAAG,IAAI,WAAW,EAAE,CAAA;QACjC,MAAM,UAAU,GAAG,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,CAAC,CAAA;QACjE,MAAM,SAAS,GAAG,eAAe,CAAC,UAAU,CAAC,CAAA;QAE7C,oDAAoD;QACpD,+CAA+C;QAC/C,MAAM,QAAQ,GAA0C,EAAE,CAAA;QAC1D,MAAM,QAAQ,GAA2B,EAAE,CAAA;QAC3C,MAAM,OAAO,GAAG,SAAS,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,KAAK,IAAI,CAAC,CAAA;QACrD,KAAK,MAAM,CAAC,KAAK,EAAE,MAAM,CAAC,IAAI,OAAO,CAAC,OAAO,EAAE,EAAE,CAAC;YAChD,IAAI,OAAO,GAAG,EAAE,CAAA;YAChB,IAAI,MAAM,GAAG,CAAC,CAAA;YACd,KAAK,MAAM,IAAI,IAAI,MAAM,CAAC,IAAI,EAAE,CAAC;gBAC/B,IAAI,IAAI,CAAC,GAAG,KAAK,IAAI,EAAE,CAAC;oBACtB,OAAO,GAAG,IAAI,CAAC,KAAK,CAAA;gBACtB,CAAC;qBAAM,IAAI,IAAI,CAAC,GAAG,KAAK,IAAI,EAAE,CAAC;oBAC7B,MAAM,GAAG,CAAC,IAAI,CAAC,KAAK,CAAA;gBACtB,CAAC;YACH,CAAC;YACD,QAAQ,CAAC,OAAO,CAAC,GAAG,KAAK,CAAA;YACzB,QAAQ,CAAC,KAAK,CAAC,GAAG,EAAE,OAAO,EAAE,MAAM,EAAE,CAAA;QACvC,CAAC;QACD,IAAI,CAAC,UAAU,GAAG,QAAQ,CAAA;QAC1B,IAAI,CAAC,UAAU,GAAG,QAAQ,CAAA;QAC1B,OAAO,SAAS,CAAA;IAClB,CAAC;CACF"}
+{"version":3,"file":"htsget.js","sourceRoot":"","sources":["../src/htsget.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,EAAE,MAAM,uBAAuB,CAAA;AAE7C,OAAO,OAAO,EAAE,EAAE,SAAS,EAAE,MAAM,cAAc,CAAA;AACjD,OAAO,KAAK,MAAM,YAAY,CAAA;AAC9B,OAAO,EAAE,aAAa,EAAE,gBAAgB,EAAE,YAAY,EAAE,MAAM,WAAW,CAAA;AACzE,OAAO,EAAE,aAAa,EAAE,MAAM,oBAAoB,CAAA;AAuBlD,kEAAkE;AAClE,+EAA+E;AAC/E,iFAAiF;AACjF,SAAS,kBAAkB,CAAC,IAAY;IACtC,IAAI,MAAmC,CAAA;IACvC,IAAI,CAAC;QACH,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;IAC3B,CAAC;IAAC,MAAM,CAAC;QACP,yCAAyC;IAC3C,CAAC;IACD,MAAM,GAAG,GAAG,MAAM,EAAE,MAAM,CAAA;IAC1B,OAAO,GAAG,EAAE,KAAK,KAAK,SAAS;QAC7B,CAAC,CAAC,IAAI;QACN,CAAC,CAAC,GAAG,CAAC,OAAO,KAAK,SAAS;YACzB,CAAC,CAAC,GAAG,CAAC,KAAK;YACX,CAAC,CAAC,GAAG,GAAG,CAAC,KAAK,KAAK,GAAG,CAAC,OAAO,EAAE,CAAA;AACtC,CAAC;AAED;;;;;;;;;;;GAWG;AACH,SAAS,aAAa,CAAC,KAAiB;IACtC,MAAM,QAAQ,GAAG,IAAI,QAAQ,CAC3B,KAAK,CAAC,MAAM,EACZ,KAAK,CAAC,UAAU,EAChB,KAAK,CAAC,UAAU,CACjB,CAAA;IACD,IAAI,KAAK,CAAC,UAAU,GAAG,CAAC,IAAI,QAAQ,CAAC,QAAQ,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,SAAS,EAAE,CAAC;QACrE,OAAO,CAAC,CAAA;IACV,CAAC;IACD,MAAM,IAAI,GAAG,YAAY,CAAC,KAAK,EAAE,CAAC,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAC,EAAE,IAAI,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAA;IACxE,IAAI,CAAC,IAAI,EAAE,CAAC;QACV,MAAM,IAAI,KAAK,CAAC,yCAAyC,CAAC,CAAA;IAC5D,CAAC;IACD,OAAO,IAAI,CAAC,GAAG,CAAA;AACjB,CAAC;AAED,KAAK,UAAU,OAAO,CAAC,OAAgB,EAAE,GAAW,EAAE,IAAkB;IACtE,MAAM,GAAG,GAAG,MAAM,OAAO,CAAC,GAAG,EAAE,IAAI,CAAC,CAAA;IACpC,IAAI,CAAC,GAAG,CAAC,EAAE,EAAE,CAAC;QACZ,MAAM,IAAI,KAAK,CACb,QAAQ,GAAG,CAAC,MAAM,aAAa,GAAG,KAAK,kBAAkB,CAAC,MAAM,GAAG,CAAC,IAAI,EAAE,CAAC,EAAE,CAC9E,CAAA;IACH,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC;AAED,KAAK,UAAU,UAAU,CACvB,OAAgB,EAChB,EAAE,GAAG,EAAE,OAAO,EAAe,EAC7B,IAAkB;IAElB,0EAA0E;IAC1E,uDAAuD;IACvD,+CAA+C;IAC/C,MAAM,EAAE,OAAO,EAAE,QAAQ,EAAE,GAAG,IAAI,EAAE,GAAG,OAAO,IAAI,EAAE,CAAA;IACpD,MAAM,GAAG,GAAG,GAAG,CAAC,UAAU,CAAC,OAAO,CAAC;QACjC,CAAC,CAAC,MAAM,OAAO,CAAC,OAAO,EAAE,GAAG,CAAC;QAC7B,CAAC,CAAC,MAAM,OAAO,CAAC,OAAO,EAAE,GAAG,EAAE,EAAE,GAAG,IAAI,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAA;IAC3D,OAAO,IAAI,UAAU,CAAC,MAAM,GAAG,CAAC,WAAW,EAAE,CAAC,CAAA;AAChD,CAAC;AAED,KAAK,UAAU,cAAc,CAC3B,OAAgB,EAChB,GAAkB,EAClB,IAAkB;IAElB,2EAA2E;IAC3E,OAAO,gBAAgB,CACrB,MAAM,OAAO,CAAC,GAAG,CACf,GAAG,CAAC,GAAG,CAAC,KAAK,EAAC,CAAC,EAAC,EAAE,CAAC,KAAK,CAAC,MAAM,UAAU,CAAC,OAAO,EAAE,CAAC,EAAE,IAAI,CAAC,CAAC,CAAC,CAC9D,CACF,CAAA;AACH,CAAC;AAED,MAAM,CAAC,OAAO,OAAO,UAEnB,SAAQ,OAAU;IACV,OAAO,CAAQ;IAEf,OAAO,CAAQ;IAEf,OAAO,CAAS;IAExB,YAAY,IAaX;QACC,KAAK,CAAC,EAAE,MAAM,EAAE,IAAI,EAAE,WAAW,EAAE,IAAI,CAAC,WAAW,EAAE,CAAC,CAAA;QACtD,IAAI,CAAC,OAAO,GAAG,IAAI,CAAC,OAAO,CAAA;QAC3B,IAAI,CAAC,OAAO,GAAG,IAAI,CAAC,OAAO,CAAA;QAC3B,IAAI,CAAC,OAAO,GAAG,IAAI,CAAC,KAAK,IAAI,CAAC,CAAC,KAAK,EAAE,IAAI,EAAE,EAAE,CAAC,KAAK,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC,CAAA;IACpE,CAAC;IAED;;;OAGG;IACK,KAAK,CAAC,WAAW,CAAC,KAAa,EAAE,IAAe;QACtD,MAAM,GAAG,GAAG,GAAG,IAAI,CAAC,OAAO,IAAI,IAAI,CAAC,OAAO,IAAI,KAAK,EAAE,CAAA;QACtD,MAAM,GAAG,GAAG,MAAM,OAAO,CAAC,IAAI,CAAC,OAAO,EAAE,GAAG,EAAE,EAAE,MAAM,EAAE,IAAI,EAAE,MAAM,EAAE,CAAC,CAAA;QACtE,MAAM,MAAM,GAAiB,MAAM,GAAG,CAAC,IAAI,EAAE,CAAA;QAC7C,OAAO,cAAc,CAAC,IAAI,CAAC,OAAO,EAAE,MAAM,CAAC,MAAM,CAAC,IAAI,EAAE;YACtD,MAAM,EAAE,IAAI,EAAE,MAAM;SACrB,CAAC,CAAA;IACJ,CAAC;IAED,KAAK,CAAC,kBAAkB,CACtB,GAAW,EACX,GAAW,EACX,GAAW,EACX,IAAc;QAEd,MAAM,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,CAAA;QAC1B,MAAM,KAAK,GAAG,IAAI,CAAC,UAAU,EAAE,CAAC,GAAG,CAAC,CAAA;QACpC,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YACxB,OAAO,EAAE,CAAA;QACX,CAAC;QACD,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,WAAW,CAClC,iBAAiB,GAAG,UAAU,GAAG,QAAQ,GAAG,aAAa,EACzD,IAAI,CACL,CAAA;QACD,MAAM,IAAI,GAAG,IAAI,aAAa,CAAC,CAAC,EAAE,CAAC,CAAC,CAAA;QACpC,MAAM,OAAO,GAAG,IAAI,CAAC,eAAe,CAClC,KAAK,CAAC,QAAQ,CAAC,aAAa,CAAC,KAAK,CAAC,CAAC,EACpC,EAAE,EACF,EAAE,EACF,IAAI,KAAK,CAAC,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC,CACzB,CAAA;QACD,OAAO,aAAa,CAAC,OAAO,EAAE,KAAK,EAAE,GAAG,EAAE,GAAG,CAAC,CAAA;IAChD,CAAC;IAED,KAAK,CAAC,YAAY,CAAC,OAAiB,EAAE;QACpC,wEAAwE;QACxE,wDAAwD;QACxD,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC,yBAAyB,EAAE,IAAI,CAAC,CAAA;QACrE,MAAM,SAAS,GAAG,IAAI,CAAC,WAAW,CAAC,KAAK,CAAC,CAAA;QACzC,IAAI,CAAC,SAAS,EAAE,CAAC;YACf,MAAM,IAAI,KAAK,CAAC,2CAA2C,CAAC,CAAA;QAC9D,CAAC;QACD,OAAO,SAAS,CAAA;IAClB,CAAC;CACF"}

package/esm/index.d.ts

@@ -5,3 +5,4 @@ export { default as BamRecord } from './record.ts';
 export { default as HtsgetFile } from './htsget.ts';
 export type { Bytes } from './record.ts';
 export type { BamRecordClass, BamRecordLike } from './bamFile.ts';
+export type { Fetcher } from 'generic-filehandle2';

package/esm/util.d.ts

@@ -36,11 +36,19 @@ export declare function parseRefSeqs(uncba: Uint8Array, start: number, renameRef
         refName: string;
         length: number;
     }[];
+    end: number;
 } | undefined;
 /**
  * The smallest of `current` and the `count` packed virtual offsets starting at
  * `offset`, allocating at most one VirtualOffset rather than one per entry.
  *
+ * 0:0 is skipped rather than treated as the minimum. No BAM record can live
+ * there — the magic and header occupy the start of the file — so it is the
+ * "unset" placeholder htslib leaves in linear-index windows ahead of a
+ * reference's first read (see test/data/HG00096_illumina_lowcov.bam.bai, whose
+ * first three windows are 0). Counting it collapses firstDataLine to 0:0 and
+ * makes callers size a header read from nothing.
+ *
  * The index first pass exists only to find this minimum, and it visits every
  * linear-index entry in the file to do it. Building a VirtualOffset per entry
  * to compare and discard it is the bulk of that pass.

package/esm/util.js

@@ -126,13 +126,23 @@ export function parseRefSeqs(uncba, start, renameRefSeq) {
         indexToChr.push({ refName, length: lRef });
         p += 8 + lName;
     }
-    return { chrToIndex, indexToChr };
+    // end is the offset just past the header, i.e. where alignment records start
+    return { chrToIndex, indexToChr, end: p };
 }
-// SYNC: ~/src/gmod/tabix-js/src/util.ts minVirtualOffset
+// SYNC: ~/src/gmod/tabix-js/src/util.ts minVirtualOffset — but NOT the 0:0
+// skip below, which is only sound for BAM. A tabix'd file with no header lines
+// really does have its first record at 0:0.
 /**
  * The smallest of `current` and the `count` packed virtual offsets starting at
  * `offset`, allocating at most one VirtualOffset rather than one per entry.
  *
+ * 0:0 is skipped rather than treated as the minimum. No BAM record can live
+ * there — the magic and header occupy the start of the file — so it is the
+ * "unset" placeholder htslib leaves in linear-index windows ahead of a
+ * reference's first read (see test/data/HG00096_illumina_lowcov.bam.bai, whose
+ * first three windows are 0). Counting it collapses firstDataLine to 0:0 and
+ * makes callers size a header read from nothing.
+ *
  * The index first pass exists only to find this minimum, and it visits every
  * linear-index entry in the file to do it. Building a VirtualOffset per entry
  * to compare and discard it is the bulk of that pass.
@@ -150,7 +160,8 @@ export function minVirtualOffset(bytes, offset, count, current) {
             bytes[p + 3] * 0x100 +
             bytes[p + 2];
         const data = (bytes[p + 1] << 8) | bytes[p];
-        if (block < minBlock || (block === minBlock && data < minData)) {
+        if ((block !== 0 || data !== 0) &&
+            (block < minBlock || (block === minBlock && data < minData))) {
             minBlock = block;
             minData = data;
             found = true;

package/esm/util.js.map

@@ -1 +1 @@
-{"version":3,"file":"util.js","sourceRoot":"","sources":["../src/util.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,MAAM,YAAY,CAAA;AAC9B,OAAO,EAAE,uBAAuB,EAAE,MAAM,WAAW,CAAA;AACnD,OAAO,EAAE,aAAa,EAAE,MAAM,oBAAoB,CAAA;AA8BlD,MAAM,UAAU,cAAc,CAAC,MAAe,EAAE,MAAe;IAC7D,MAAM,CAAC,GAAG,MAAM,CAAC,MAAM,CAAA;IACvB,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;QACZ,OAAO,MAAM,CAAA;IACf,CAAC;IAED,0DAA0D;IAC1D,IAAI,QAAiB,CAAA;IACrB,IAAI,MAAM,EAAE,CAAC;QACX,MAAM,WAAW,GAAG,MAAM,CAAC,aAAa,CAAA;QACxC,MAAM,UAAU,GAAG,MAAM,CAAC,YAAY,CAAA;QACtC,QAAQ,GAAG,EAAE,CAAA;QACb,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;YAC3B,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAE,CAAA;YACxB,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAA;YACvB,MAAM,GAAG,GACP,IAAI,CAAC,aAAa,GAAG,WAAW,IAAI,IAAI,CAAC,YAAY,GAAG,UAAU,CAAA;YACpE,IAAI,GAAG,GAAG,CAAC,EAAE,CAAC;gBACZ,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;YACtB,CAAC;QACH,CAAC;QACD,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAC1B,OAAO,QAAQ,CAAA;QACjB,CAAC;IACH,CAAC;SAAM,CAAC;QACN,QAAQ,GAAG,MAAM,CAAA;IACnB,CAAC;IAED,QAAQ,CAAC,IAAI,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE,EAAE;QACvB,MAAM,GAAG,GAAG,EAAE,CAAC,IAAI,CAAC,aAAa,GAAG,EAAE,CAAC,IAAI,CAAC,aAAa,CAAA;QACzD,OAAO,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,CAAC,YAAY,GAAG,EAAE,CAAC,IAAI,CAAC,YAAY,CAAA;IACtE,CAAC,CAAC,CAAA;IAEF,yEAAyE;IACzE,uEAAuE;IACvE,MAAM,YAAY,GAAY,CAAC,QAAQ,CAAC,CAAC,CAAE,CAAC,CAAA;IAC5C,IAAI,YAAY,GAAG,QAAQ,CAAC,CAAC,CAAE,CAAC,IAAI,CAAC,aAAa,CAAA;IAClD,IAAI,YAAY,GAAG,QAAQ,CAAC,CAAC,CAAE,CAAC,IAAI,CAAC,aAAa,CAAA;IAElD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,QAAQ,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACzC,MAAM,KAAK,GAAG,QAAQ,CAAC,CAAC,CAAE,CAAA;QAC1B,MAAM,aAAa,GAAG,KAAK,CAAC,IAAI,CAAC,aAAa,CAAA;QAC9C,MAAM,aAAa,GAAG,KAAK,CAAC,IAAI,CAAC,aAAa,CAAA;QAC9C,oEAAoE;QACpE,yEAAyE;QACzE,IACE,aAAa,GAAG,YAAY,GAAG,KAAK;YACpC,aAAa,GAAG,YAAY,GAAG,OAAO,EACtC,CAAC;YACD,MAAM,SAAS,GAAG,YAAY,CAAC,YAAY,CAAC,MAAM,GAAG,CAAC,CAAE,CAAA;YACxD,MAAM,GAAG,GACP,aAAa,GAAG,YAAY;gBAC5B,KAAK,CAAC,IAAI,CAAC,YAAY,GAAG,SAAS,CAAC,IAAI,CAAC,YAAY,CAAA;YACvD,IAAI,GAAG,GAAG,CAAC,EAAE,CAAC;gBACZ,YAAY,CAAC,YAAY,CAAC,MAAM,GAAG,CAAC,CAAC,GAAG,IAAI,KAAK,CAC/C,SAAS,CAAC,IAAI,EACd,KAAK,CAAC,IAAI,EACV,SAAS,CAAC,GAAG,EACb,KAAK,CAAC,WAAW,CAClB,CAAA;gBACD,YAAY,GAAG,aAAa,CAAA;YAC9B,CAAC;QACH,CAAC;aAAM,CAAC;YACN,YAAY,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;YACxB,YAAY,GAAG,aAAa,CAAA;YAC5B,YAAY,GAAG,aAAa,CAAA;QAC9B,CAAC;IACH,CAAC;IAED,OAAO,YAAY,CAAA;AACrB,CAAC;AAED,MAAM,UAAU,cAAc,CAAC,KAAiB,EAAE,MAAc;IAC9D,OAAO;QACL,SAAS,EAAE,uBAAuB,CAAC,KAAK,EAAE,MAAM,CAAC;KAClD,CAAA;AACH,CAAC;AAED,6EAA6E;AAC7E,kFAAkF;AAClF,kFAAkF;AAClF,iFAAiF;AACjF,kFAAkF;AAClF,iFAAiF;AACjF,yEAAyE;AACzE,MAAM,UAAU,cAAc,CAC5B,MAAe,EACf,kBAA4B,EAAE;IAE9B,MAAM,UAAU,GAAG,CAAC,GAAG,eAAe,CAAC,CAAA;IACvC,KAAK,MAAM,CAAC,IAAI,MAAM,EAAE,CAAC;QACvB,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,aAAa,EAAE,CAAC,CAAC,IAAI,CAAC,aAAa,CAAC,CAAA;IAC7D,CAAC;IACD,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;IAEhC,KAAK,MAAM,CAAC,IAAI,MAAM,EAAE,CAAC;QACvB,MAAM,GAAG,GAAG,CAAC,CAAC,IAAI,CAAC,aAAa,CAAA;QAChC,2CAA2C;QAC3C,IAAI,EAAE,GAAG,CAAC,CAAA;QACV,IAAI,EAAE,GAAG,UAAU,CAAC,MAAM,CAAA;QAC1B,OAAO,EAAE,GAAG,EAAE,EAAE,CAAC;YACf,MAAM,GAAG,GAAG,CAAC,EAAE,GAAG,EAAE,CAAC,KAAK,CAAC,CAAA;YAC3B,IAAI,UAAU,CAAC,GAAG,CAAE,GAAG,GAAG,EAAE,CAAC;gBAC3B,EAAE,GAAG,GAAG,CAAA;YACV,CAAC;iBAAM,CAAC;gBACN,EAAE,GAAG,GAAG,GAAG,CAAC,CAAA;YACd,CAAC;QACH,CAAC;QACD,IAAI,EAAE,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC;YAC3B,CAAC,CAAC,WAAW,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,WAAW,EAAE,UAAU,CAAC,EAAE,CAAE,CAAC,CAAA;QAC1D,CAAC;IACH,CAAC;AACH,CAAC;AAED,6EAA6E;AAC7E,4EAA4E;AAC5E,eAAe;AACf,MAAM,UAAU,YAAY,CAC1B,KAAiB,EACjB,KAAa,EACb,YAAmC;IAEnC,IAAI,KAAK,GAAG,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;QAC7B,OAAO,SAAS,CAAA;IAClB,CAAC;IACD,MAAM,QAAQ,GAAG,IAAI,QAAQ,CAC3B,KAAK,CAAC,MAAM,EACZ,KAAK,CAAC,UAAU,EAChB,KAAK,CAAC,UAAU,CACjB,CAAA;IACD,MAAM,IAAI,GAAG,QAAQ,CAAC,QAAQ,CAAC,KAAK,EAAE,IAAI,CAAC,CAAA;IAC3C,kEAAkE;IAClE,uDAAuD;IACvD,MAAM,UAAU,GAA2B,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;IAC9D,MAAM,UAAU,GAA0C,EAAE,CAAA;IAC5D,MAAM,OAAO,GAAG,IAAI,WAAW,CAAC,MAAM,CAAC,CAAA;IAEvC,IAAI,CAAC,GAAG,KAAK,GAAG,CAAC,CAAA;IACjB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,EAAE,CAAC,EAAE,EAAE,CAAC;QAC9B,IAAI,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;YACzB,OAAO,SAAS,CAAA;QAClB,CAAC;QACD,MAAM,KAAK,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAC,EAAE,IAAI,CAAC,CAAA;QACxC,IAAI,CAAC,GAAG,CAAC,GAAG,KAAK,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;YACjC,OAAO,SAAS,CAAA;QAClB,CAAC;QACD,MAAM,OAAO,GAAG,YAAY,CAC1B,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,GAAG,KAAK,GAAG,CAAC,CAAC,CAAC,CACzD,CAAA;QACD,MAAM,IAAI,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAC,GAAG,KAAK,GAAG,CAAC,EAAE,IAAI,CAAC,CAAA;QACnD,UAAU,CAAC,OAAO,CAAC,GAAG,CAAC,CAAA;QACvB,UAAU,CAAC,IAAI,CAAC,EAAE,OAAO,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,CAAA;QAC1C,CAAC,IAAI,CAAC,GAAG,KAAK,CAAA;IAChB,CAAC;IACD,OAAO,EAAE,UAAU,EAAE,UAAU,EAAE,CAAA;AACnC,CAAC;AAED,yDAAyD;AACzD;;;;;;;GAOG;AACH,MAAM,UAAU,gBAAgB,CAC9B,KAAiB,EACjB,MAAc,EACd,KAAa,EACb,OAAkC;IAElC,IAAI,QAAQ,GAAG,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,aAAa,CAAC,CAAC,CAAC,QAAQ,CAAA;IACzD,IAAI,OAAO,GAAG,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC,CAAA;IAChD,IAAI,KAAK,GAAG,KAAK,CAAA;IACjB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,EAAE,CAAC,EAAE,EAAE,CAAC;QAC/B,MAAM,CAAC,GAAG,MAAM,GAAG,CAAC,GAAG,CAAC,CAAA;QACxB,MAAM,KAAK,GACT,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,GAAG,aAAa;YAC7B,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,GAAG,WAAW;YAC3B,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,GAAG,SAAS;YACzB,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,GAAG,OAAO;YACvB,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,GAAG,KAAK;YACrB,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,CAAA;QACf,MAAM,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAE,CAAA;QAC7C,IAAI,KAAK,GAAG,QAAQ,IAAI,CAAC,KAAK,KAAK,QAAQ,IAAI,IAAI,GAAG,OAAO,CAAC,EAAE,CAAC;YAC/D,QAAQ,GAAG,KAAK,CAAA;YAChB,OAAO,GAAG,IAAI,CAAA;YACd,KAAK,GAAG,IAAI,CAAA;QACd,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC,CAAC,CAAC,IAAI,aAAa,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC,CAAC,CAAC,OAAO,CAAA;AAC/D,CAAC;AAED,4FAA4F;AAC5F,MAAM,UAAU,cAAc,CAC5B,UAAsB,EACtB,eAAwC,CAAC,CAAC,EAAE,CAAC,CAAC;IAE9C,MAAM,OAAO,GAAG,IAAI,WAAW,EAAE,CAAA;IACjC,IAAI,SAAS,GAAG,CAAC,CAAA;IACjB,IAAI,aAAa,GAAG,CAAC,CAAA;IACrB,MAAM,WAAW,GAAa,EAAE,CAAA;IAChC,MAAM,WAAW,GAA2B,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;IAC/D,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC3C,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,EAAE,CAAC;YACnB,IAAI,aAAa,GAAG,CAAC,EAAE,CAAC;gBACtB,MAAM,OAAO,GAAG,YAAY,CAC1B,OAAO,CAAC,MAAM,CAAC,UAAU,CAAC,QAAQ,CAAC,aAAa,EAAE,CAAC,CAAC,CAAC,CACtD,CAAA;gBACD,WAAW,CAAC,SAAS,CAAC,GAAG,OAAO,CAAA;gBAChC,WAAW,CAAC,OAAO,CAAC,GAAG,SAAS,CAAA;YAClC,CAAC;YACD,aAAa,GAAG,CAAC,GAAG,CAAC,CAAA;YACrB,SAAS,EAAE,CAAA;QACb,CAAC;IACH,CAAC;IACD,OAAO,EAAE,WAAW,EAAE,WAAW,EAAE,CAAA;AACrC,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,IAAkB;IACjD,IAAI,WAAW,GAAG,CAAC,CAAA;IACnB,KAAK,MAAM,KAAK,IAAI,IAAI,EAAE,CAAC;QACzB,WAAW,IAAI,KAAK,CAAC,MAAM,CAAA;IAC7B,CAAC;IACD,MAAM,WAAW,GAAG,IAAI,UAAU,CAAC,WAAW,CAAC,CAAA;IAC/C,IAAI,MAAM,GAAG,CAAC,CAAA;IACd,KAAK,MAAM,KAAK,IAAI,IAAI,EAAE,CAAC;QACzB,WAAW,CAAC,GAAG,CAAC,KAAK,EAAE,MAAM,CAAC,CAAA;QAC9B,MAAM,IAAI,KAAK,CAAC,MAAM,CAAA;IACxB,CAAC;IACD,OAAO,WAAW,CAAA;AACpB,CAAC;AAQD,0EAA0E;AAC1E,4EAA4E;AAC5E,6EAA6E;AAC7E,iDAAiD;AACjD,MAAM,UAAU,aAAa,CAC3B,OAAY,EACZ,KAAa,EACb,GAAW,EACX,GAAW,EACX,MAAW,EAAE;IAEb,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QAC/C,MAAM,CAAC,GAAG,OAAO,CAAC,CAAC,CAAE,CAAA;QACrB,IAAI,CAAC,CAAC,MAAM,KAAK,KAAK,EAAE,CAAC;YACvB,IAAI,CAAC,CAAC,KAAK,IAAI,GAAG,EAAE,CAAC;gBACnB,MAAK;YACP,CAAC;iBAAM,IAAI,CAAC,CAAC,GAAG,IAAI,GAAG,EAAE,CAAC;gBACxB,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;YACb,CAAC;QACH,CAAC;aAAM,IAAI,CAAC,CAAC,MAAM,GAAG,KAAK,EAAE,CAAC;YAC5B,MAAK;QACP,CAAC;IACH,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC"}
+{"version":3,"file":"util.js","sourceRoot":"","sources":["../src/util.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,MAAM,YAAY,CAAA;AAC9B,OAAO,EAAE,uBAAuB,EAAE,MAAM,WAAW,CAAA;AACnD,OAAO,EAAE,aAAa,EAAE,MAAM,oBAAoB,CAAA;AA8BlD,MAAM,UAAU,cAAc,CAAC,MAAe,EAAE,MAAe;IAC7D,MAAM,CAAC,GAAG,MAAM,CAAC,MAAM,CAAA;IACvB,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;QACZ,OAAO,MAAM,CAAA;IACf,CAAC;IAED,0DAA0D;IAC1D,IAAI,QAAiB,CAAA;IACrB,IAAI,MAAM,EAAE,CAAC;QACX,MAAM,WAAW,GAAG,MAAM,CAAC,aAAa,CAAA;QACxC,MAAM,UAAU,GAAG,MAAM,CAAC,YAAY,CAAA;QACtC,QAAQ,GAAG,EAAE,CAAA;QACb,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;YAC3B,MAAM,KAAK,GAAG,MAAM,CAAC,CAAC,CAAE,CAAA;YACxB,MAAM,IAAI,GAAG,KAAK,CAAC,IAAI,CAAA;YACvB,MAAM,GAAG,GACP,IAAI,CAAC,aAAa,GAAG,WAAW,IAAI,IAAI,CAAC,YAAY,GAAG,UAAU,CAAA;YACpE,IAAI,GAAG,GAAG,CAAC,EAAE,CAAC;gBACZ,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;YACtB,CAAC;QACH,CAAC;QACD,IAAI,QAAQ,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAC1B,OAAO,QAAQ,CAAA;QACjB,CAAC;IACH,CAAC;SAAM,CAAC;QACN,QAAQ,GAAG,MAAM,CAAA;IACnB,CAAC;IAED,QAAQ,CAAC,IAAI,CAAC,CAAC,EAAE,EAAE,EAAE,EAAE,EAAE;QACvB,MAAM,GAAG,GAAG,EAAE,CAAC,IAAI,CAAC,aAAa,GAAG,EAAE,CAAC,IAAI,CAAC,aAAa,CAAA;QACzD,OAAO,GAAG,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,CAAC,YAAY,GAAG,EAAE,CAAC,IAAI,CAAC,YAAY,CAAA;IACtE,CAAC,CAAC,CAAA;IAEF,yEAAyE;IACzE,uEAAuE;IACvE,MAAM,YAAY,GAAY,CAAC,QAAQ,CAAC,CAAC,CAAE,CAAC,CAAA;IAC5C,IAAI,YAAY,GAAG,QAAQ,CAAC,CAAC,CAAE,CAAC,IAAI,CAAC,aAAa,CAAA;IAClD,IAAI,YAAY,GAAG,QAAQ,CAAC,CAAC,CAAE,CAAC,IAAI,CAAC,aAAa,CAAA;IAElD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,QAAQ,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QACzC,MAAM,KAAK,GAAG,QAAQ,CAAC,CAAC,CAAE,CAAA;QAC1B,MAAM,aAAa,GAAG,KAAK,CAAC,IAAI,CAAC,aAAa,CAAA;QAC9C,MAAM,aAAa,GAAG,KAAK,CAAC,IAAI,CAAC,aAAa,CAAA;QAC9C,oEAAoE;QACpE,yEAAyE;QACzE,IACE,aAAa,GAAG,YAAY,GAAG,KAAK;YACpC,aAAa,GAAG,YAAY,GAAG,OAAO,EACtC,CAAC;YACD,MAAM,SAAS,GAAG,YAAY,CAAC,YAAY,CAAC,MAAM,GAAG,CAAC,CAAE,CAAA;YACxD,MAAM,GAAG,GACP,aAAa,GAAG,YAAY;gBAC5B,KAAK,CAAC,IAAI,CAAC,YAAY,GAAG,SAAS,CAAC,IAAI,CAAC,YAAY,CAAA;YACvD,IAAI,GAAG,GAAG,CAAC,EAAE,CAAC;gBACZ,YAAY,CAAC,YAAY,CAAC,MAAM,GAAG,CAAC,CAAC,GAAG,IAAI,KAAK,CAC/C,SAAS,CAAC,IAAI,EACd,KAAK,CAAC,IAAI,EACV,SAAS,CAAC,GAAG,EACb,KAAK,CAAC,WAAW,CAClB,CAAA;gBACD,YAAY,GAAG,aAAa,CAAA;YAC9B,CAAC;QACH,CAAC;aAAM,CAAC;YACN,YAAY,CAAC,IAAI,CAAC,KAAK,CAAC,CAAA;YACxB,YAAY,GAAG,aAAa,CAAA;YAC5B,YAAY,GAAG,aAAa,CAAA;QAC9B,CAAC;IACH,CAAC;IAED,OAAO,YAAY,CAAA;AACrB,CAAC;AAED,MAAM,UAAU,cAAc,CAAC,KAAiB,EAAE,MAAc;IAC9D,OAAO;QACL,SAAS,EAAE,uBAAuB,CAAC,KAAK,EAAE,MAAM,CAAC;KAClD,CAAA;AACH,CAAC;AAED,6EAA6E;AAC7E,kFAAkF;AAClF,kFAAkF;AAClF,iFAAiF;AACjF,kFAAkF;AAClF,iFAAiF;AACjF,yEAAyE;AACzE,MAAM,UAAU,cAAc,CAC5B,MAAe,EACf,kBAA4B,EAAE;IAE9B,MAAM,UAAU,GAAG,CAAC,GAAG,eAAe,CAAC,CAAA;IACvC,KAAK,MAAM,CAAC,IAAI,MAAM,EAAE,CAAC;QACvB,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,aAAa,EAAE,CAAC,CAAC,IAAI,CAAC,aAAa,CAAC,CAAA;IAC7D,CAAC;IACD,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;IAEhC,KAAK,MAAM,CAAC,IAAI,MAAM,EAAE,CAAC;QACvB,MAAM,GAAG,GAAG,CAAC,CAAC,IAAI,CAAC,aAAa,CAAA;QAChC,2CAA2C;QAC3C,IAAI,EAAE,GAAG,CAAC,CAAA;QACV,IAAI,EAAE,GAAG,UAAU,CAAC,MAAM,CAAA;QAC1B,OAAO,EAAE,GAAG,EAAE,EAAE,CAAC;YACf,MAAM,GAAG,GAAG,CAAC,EAAE,GAAG,EAAE,CAAC,KAAK,CAAC,CAAA;YAC3B,IAAI,UAAU,CAAC,GAAG,CAAE,GAAG,GAAG,EAAE,CAAC;gBAC3B,EAAE,GAAG,GAAG,CAAA;YACV,CAAC;iBAAM,CAAC;gBACN,EAAE,GAAG,GAAG,GAAG,CAAC,CAAA;YACd,CAAC;QACH,CAAC;QACD,IAAI,EAAE,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC;YAC3B,CAAC,CAAC,WAAW,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,CAAC,WAAW,EAAE,UAAU,CAAC,EAAE,CAAE,CAAC,CAAA;QAC1D,CAAC;IACH,CAAC;AACH,CAAC;AAED,6EAA6E;AAC7E,4EAA4E;AAC5E,eAAe;AACf,MAAM,UAAU,YAAY,CAC1B,KAAiB,EACjB,KAAa,EACb,YAAmC;IAEnC,IAAI,KAAK,GAAG,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;QAC7B,OAAO,SAAS,CAAA;IAClB,CAAC;IACD,MAAM,QAAQ,GAAG,IAAI,QAAQ,CAC3B,KAAK,CAAC,MAAM,EACZ,KAAK,CAAC,UAAU,EAChB,KAAK,CAAC,UAAU,CACjB,CAAA;IACD,MAAM,IAAI,GAAG,QAAQ,CAAC,QAAQ,CAAC,KAAK,EAAE,IAAI,CAAC,CAAA;IAC3C,kEAAkE;IAClE,uDAAuD;IACvD,MAAM,UAAU,GAA2B,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;IAC9D,MAAM,UAAU,GAA0C,EAAE,CAAA;IAC5D,MAAM,OAAO,GAAG,IAAI,WAAW,CAAC,MAAM,CAAC,CAAA;IAEvC,IAAI,CAAC,GAAG,KAAK,GAAG,CAAC,CAAA;IACjB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,IAAI,EAAE,CAAC,EAAE,EAAE,CAAC;QAC9B,IAAI,CAAC,GAAG,CAAC,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;YACzB,OAAO,SAAS,CAAA;QAClB,CAAC;QACD,MAAM,KAAK,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAC,EAAE,IAAI,CAAC,CAAA;QACxC,IAAI,CAAC,GAAG,CAAC,GAAG,KAAK,GAAG,KAAK,CAAC,MAAM,EAAE,CAAC;YACjC,OAAO,SAAS,CAAA;QAClB,CAAC;QACD,MAAM,OAAO,GAAG,YAAY,CAC1B,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,CAAC,GAAG,KAAK,GAAG,CAAC,CAAC,CAAC,CACzD,CAAA;QACD,MAAM,IAAI,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAC,GAAG,KAAK,GAAG,CAAC,EAAE,IAAI,CAAC,CAAA;QACnD,UAAU,CAAC,OAAO,CAAC,GAAG,CAAC,CAAA;QACvB,UAAU,CAAC,IAAI,CAAC,EAAE,OAAO,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,CAAA;QAC1C,CAAC,IAAI,CAAC,GAAG,KAAK,CAAA;IAChB,CAAC;IACD,6EAA6E;IAC7E,OAAO,EAAE,UAAU,EAAE,UAAU,EAAE,GAAG,EAAE,CAAC,EAAE,CAAA;AAC3C,CAAC;AAED,2EAA2E;AAC3E,+EAA+E;AAC/E,4CAA4C;AAC5C;;;;;;;;;;;;;;GAcG;AACH,MAAM,UAAU,gBAAgB,CAC9B,KAAiB,EACjB,MAAc,EACd,KAAa,EACb,OAAkC;IAElC,IAAI,QAAQ,GAAG,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,aAAa,CAAC,CAAC,CAAC,QAAQ,CAAA;IACzD,IAAI,OAAO,GAAG,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC,CAAA;IAChD,IAAI,KAAK,GAAG,KAAK,CAAA;IACjB,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,KAAK,EAAE,CAAC,EAAE,EAAE,CAAC;QAC/B,MAAM,CAAC,GAAG,MAAM,GAAG,CAAC,GAAG,CAAC,CAAA;QACxB,MAAM,KAAK,GACT,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,GAAG,aAAa;YAC7B,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,GAAG,WAAW;YAC3B,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,GAAG,SAAS;YACzB,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,GAAG,OAAO;YACvB,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,GAAG,KAAK;YACrB,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,CAAA;QACf,MAAM,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC,GAAG,CAAC,CAAE,IAAI,CAAC,CAAC,GAAG,KAAK,CAAC,CAAC,CAAE,CAAA;QAC7C,IACE,CAAC,KAAK,KAAK,CAAC,IAAI,IAAI,KAAK,CAAC,CAAC;YAC3B,CAAC,KAAK,GAAG,QAAQ,IAAI,CAAC,KAAK,KAAK,QAAQ,IAAI,IAAI,GAAG,OAAO,CAAC,CAAC,EAC5D,CAAC;YACD,QAAQ,GAAG,KAAK,CAAA;YAChB,OAAO,GAAG,IAAI,CAAA;YACd,KAAK,GAAG,IAAI,CAAA;QACd,CAAC;IACH,CAAC;IACD,OAAO,KAAK,CAAC,CAAC,CAAC,IAAI,aAAa,CAAC,QAAQ,EAAE,OAAO,CAAC,CAAC,CAAC,CAAC,OAAO,CAAA;AAC/D,CAAC;AAED,4FAA4F;AAC5F,MAAM,UAAU,cAAc,CAC5B,UAAsB,EACtB,eAAwC,CAAC,CAAC,EAAE,CAAC,CAAC;IAE9C,MAAM,OAAO,GAAG,IAAI,WAAW,EAAE,CAAA;IACjC,IAAI,SAAS,GAAG,CAAC,CAAA;IACjB,IAAI,aAAa,GAAG,CAAC,CAAA;IACrB,MAAM,WAAW,GAAa,EAAE,CAAA;IAChC,MAAM,WAAW,GAA2B,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,CAAA;IAC/D,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,UAAU,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;QAC3C,IAAI,CAAC,UAAU,CAAC,CAAC,CAAC,EAAE,CAAC;YACnB,IAAI,aAAa,GAAG,CAAC,EAAE,CAAC;gBACtB,MAAM,OAAO,GAAG,YAAY,CAC1B,OAAO,CAAC,MAAM,CAAC,UAAU,CAAC,QAAQ,CAAC,aAAa,EAAE,CAAC,CAAC,CAAC,CACtD,CAAA;gBACD,WAAW,CAAC,SAAS,CAAC,GAAG,OAAO,CAAA;gBAChC,WAAW,CAAC,OAAO,CAAC,GAAG,SAAS,CAAA;YAClC,CAAC;YACD,aAAa,GAAG,CAAC,GAAG,CAAC,CAAA;YACrB,SAAS,EAAE,CAAA;QACb,CAAC;IACH,CAAC;IACD,OAAO,EAAE,WAAW,EAAE,WAAW,EAAE,CAAA;AACrC,CAAC;AAED,MAAM,UAAU,gBAAgB,CAAC,IAAkB;IACjD,IAAI,WAAW,GAAG,CAAC,CAAA;IACnB,KAAK,MAAM,KAAK,IAAI,IAAI,EAAE,CAAC;QACzB,WAAW,IAAI,KAAK,CAAC,MAAM,CAAA;IAC7B,CAAC;IACD,MAAM,WAAW,GAAG,IAAI,UAAU,CAAC,WAAW,CAAC,CAAA;IAC/C,IAAI,MAAM,GAAG,CAAC,CAAA;IACd,KAAK,MAAM,KAAK,IAAI,IAAI,EAAE,CAAC;QACzB,WAAW,CAAC,GAAG,CAAC,KAAK,EAAE,MAAM,CAAC,CAAA;QAC9B,MAAM,IAAI,KAAK,CAAC,MAAM,CAAA;IACxB,CAAC;IACD,OAAO,WAAW,CAAA;AACpB,CAAC;AAQD,0EAA0E;AAC1E,4EAA4E;AAC5E,6EAA6E;AAC7E,iDAAiD;AACjD,MAAM,UAAU,aAAa,CAC3B,OAAY,EACZ,KAAa,EACb,GAAW,EACX,GAAW,EACX,MAAW,EAAE;IAEb,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,OAAO,CAAC,MAAM,EAAE,CAAC,GAAG,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC;QAC/C,MAAM,CAAC,GAAG,OAAO,CAAC,CAAC,CAAE,CAAA;QACrB,IAAI,CAAC,CAAC,MAAM,KAAK,KAAK,EAAE,CAAC;YACvB,IAAI,CAAC,CAAC,KAAK,IAAI,GAAG,EAAE,CAAC;gBACnB,MAAK;YACP,CAAC;iBAAM,IAAI,CAAC,CAAC,GAAG,IAAI,GAAG,EAAE,CAAC;gBACxB,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;YACb,CAAC;QACH,CAAC;aAAM,IAAI,CAAC,CAAC,MAAM,GAAG,KAAK,EAAE,CAAC;YAC5B,MAAK;QACP,CAAC;IACH,CAAC;IACD,OAAO,GAAG,CAAA;AACZ,CAAC"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@gmod/bam",
-  "version": "7.6.1",
+  "version": "7.6.2",
   "packageManager": "pnpm@11.15.1",
   "description": "Parser for BAM and BAM index (bai) files",
   "license": "MIT",
@@ -39,7 +39,7 @@
     "build": "pnpm build:esm && pnpm build:es5",
     "benchonly": "vitest bench",
     "bench": "./scripts/build-both-branches.sh \"$BRANCH1\" \"$BRANCH2\" && vitest bench",
-    "version": "standard-changelog && git add CHANGELOG.md"
+    "version": "git-cliff --tag v$npm_package_version --unreleased --prepend CHANGELOG.md && git add CHANGELOG.md"
   },
   "keywords": [
     "bionode",
@@ -60,9 +60,9 @@
     "eslint": "^10.4.1",
     "eslint-plugin-import-x": "^4.16.2",
     "eslint-plugin-unicorn": "^65.0.0",
+    "git-cliff": "^2.13.1",
     "prettier": "^3.8.3",
     "rimraf": "^6.1.3",
-    "standard-changelog": "^7.0.1",
     "typescript": "^6.0.3",
     "typescript-eslint": "^8.61.0",
     "vitest": "^4.1.8"

package/src/bamFile.ts

@@ -37,6 +37,12 @@ export const BAM_MAGIC = 21840194
 
 const blockLen = 1 << 16
 
+// Ceiling on the header read. A million contigs is roughly 10MB of compressed
+// @SQ lines and ref-seq table, so anything past this is a corrupt header
+// claiming a huge n_ref rather than a real one, and growing the read further
+// just downloads the file.
+const maxHeaderReadLen = 32 * 1024 * 1024
+
 function resolveFilehandle(
   filehandle?: GenericFilehandle,
   path?: string,
@@ -230,46 +236,61 @@ export default class BamFile<T extends BamRecordLike = BAMFeature> {
     }
     const indexData = await this.index.parse(opts)
 
-    // firstDataLine is not defined in cases where there is no data in the file
-    // (just bam header and nothing else)
-    const readLen =
+    // The records start at firstDataLine, so reading up to it plus the bgzf
+    // block straddling it covers the header. It is undefined when the index
+    // records no data at all (header-only BAM), and it undershoots when the
+    // index leaves offsets unset, so grow the read until the ref-seq table
+    // parses. Never readFile() here: on a remote BAM that is a whole-file
+    // download to read a header.
+    let readLen =
       indexData.firstDataLine === undefined
-        ? undefined
+        ? blockLen
         : indexData.firstDataLine.blockPosition + blockLen
 
-    const buffer =
-      readLen === undefined
-        ? await this.bam.readFile()
-        : await this.bam.read(readLen, 0)
-    let uncba = await unzip(buffer)
+    let samHeader
+    let atEof = false
+    while (samHeader === undefined && !atEof && readLen <= maxHeaderReadLen) {
+      const buffer = await this.bam.read(readLen, 0, { signal: opts.signal })
+      // a short read means readLen ran past the end of the file, so there are
+      // no more bytes to grow into
+      atEof = buffer.length < readLen
+      samHeader = this.applyHeader(await unzip(buffer))
+      readLen *= 2
+    }
+    if (samHeader === undefined) {
+      throw new Error('Insufficient data for reference sequences')
+    }
+    return samHeader
+  }
+
+  /**
+   * Installs the header text and ref name/id maps from the start of a
+   * decompressed BAM stream, returning the parsed SAM header lines. Returns
+   * undefined if the stream is cut off partway through the ref-seq table, so
+   * the caller can retry with more data.
+   */
+  protected applyHeader(uncba: Uint8Array) {
     const dataView = new DataView(
       uncba.buffer,
       uncba.byteOffset,
       uncba.byteLength,
     )
-
     if (dataView.getInt32(0, true) !== BAM_MAGIC) {
       throw new Error('Not a BAM file')
     }
     const headLen = dataView.getInt32(4, true)
-    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 covers. If the first attempt comes up short, fall back to
-    // reading the whole file (the index's firstDataLine is just an
-    // optimization hint, not a guaranteed cap on the ref-seq table size).
-    const refSeqStart = headLen + 8
-    let parsed = parseRefSeqs(uncba, refSeqStart, this.renameRefSeq)
-    if (!parsed) {
-      uncba = await unzip(await this.bam.readFile())
-      parsed = parseRefSeqs(uncba, refSeqStart, this.renameRefSeq)
-    }
-    if (!parsed) {
-      throw new Error('Insufficient data for reference sequences')
+    const parsed = parseRefSeqs(uncba, headLen + 8, this.renameRefSeq)
+    let samHeader
+    if (parsed) {
+      const headerText = new TextDecoder('utf8').decode(
+        uncba.subarray(8, 8 + headLen),
+      )
+      this.header = headerText
+      this.chrToIndex = parsed.chrToIndex
+      this.indexToChr = parsed.indexToChr
+      samHeader = parseHeaderText(headerText)
     }
-    this.chrToIndex = parsed.chrToIndex
-    this.indexToChr = parsed.indexToChr
-    return parseHeaderText(this.header)
+    return samHeader
   }
 
   getHeader(opts?: BaseOpts) {