@gmod/tabix 3.8.2 → 3.9.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +24 -11
- package/dist/csi.js +5 -2
- package/dist/csi.js.map +1 -1
- package/dist/indexFile.d.ts +16 -0
- package/dist/indexFile.js +68 -3
- package/dist/indexFile.js.map +1 -1
- package/dist/tabix-bundle.js +1 -1
- package/dist/tabixIndexedFile.js +152 -65
- package/dist/tabixIndexedFile.js.map +1 -1
- package/dist/util.js +1 -0
- package/dist/util.js.map +1 -1
- package/esm/csi.js +5 -2
- package/esm/csi.js.map +1 -1
- package/esm/indexFile.d.ts +16 -0
- package/esm/indexFile.js +67 -3
- package/esm/indexFile.js.map +1 -1
- package/esm/tabixIndexedFile.js +152 -65
- package/esm/tabixIndexedFile.js.map +1 -1
- package/esm/util.js +1 -0
- package/esm/util.js.map +1 -1
- package/package.json +2 -2
- package/src/csi.ts +5 -2
- package/src/indexFile.ts +79 -3
- package/src/tabixIndexedFile.ts +215 -89
- package/src/util.ts +1 -0
package/README.md
CHANGED
|
@@ -40,11 +40,13 @@ const custom = new TabixIndexedFile({
|
|
|
40
40
|
})
|
|
41
41
|
```
|
|
42
42
|
|
|
43
|
-
Over HTTP
|
|
44
|
-
[`@gmod/range-cache-filehandle`](https://github.com/GMOD/range-cache-filehandle)
|
|
45
|
-
|
|
46
|
-
|
|
47
|
-
|
|
43
|
+
Over HTTP it is worth swapping in
|
|
44
|
+
[`@gmod/range-cache-filehandle`](https://github.com/GMOD/range-cache-filehandle).
|
|
45
|
+
A query fetches the index once, then reads the BGZF blocks it points at as byte
|
|
46
|
+
ranges spread through the file. Overlapping queries re-read the same blocks:
|
|
47
|
+
panning twenty half-overlapping windows across the 3.4 MB test BED file reads 11
|
|
48
|
+
MB. The cache serves those reads out of 256 KiB chunks, so neighboring blocks
|
|
49
|
+
share a request and each byte is fetched once.
|
|
48
50
|
|
|
49
51
|
```typescript
|
|
50
52
|
import { RemoteFileWithRangeCache } from '@gmod/range-cache-filehandle'
|
|
@@ -105,6 +107,17 @@ Notes:
|
|
|
105
107
|
if a query comes back unexpectedly empty
|
|
106
108
|
- `start > end` throws a `TypeError`; `start === end` returns without reading
|
|
107
109
|
|
|
110
|
+
### GAF
|
|
111
|
+
|
|
112
|
+
Reads `tabix -p gaf` indexes of sorted GAF (e.g. from `vg gamsort`):
|
|
113
|
+
|
|
114
|
+
- `refName` is ignored; `getReferenceSequenceNames()` returns `[]`
|
|
115
|
+
- `start`/`end` are node ids: a read matches when its lowest-to-highest node
|
|
116
|
+
range `[min, max]` touches any of `start..end-1`, the same reads as
|
|
117
|
+
`tabix file.gaf.gz "{x}:start-end"` (no `+1`)
|
|
118
|
+
- The callback gets `min` and `max + 1`
|
|
119
|
+
- Reads whose path is a stable sequence name (no `>`/`<`) are skipped
|
|
120
|
+
|
|
108
121
|
### Without NPM (CDN)
|
|
109
122
|
|
|
110
123
|
```html
|
|
@@ -147,7 +160,7 @@ const file = new TabixIndexedFile({
|
|
|
147
160
|
Safe to pass unconditionally: `getSharedWorkerPool()` returns `undefined` under
|
|
148
161
|
node, or anywhere the host forbids Workers, which keeps the in-process path. No
|
|
149
162
|
cross-origin isolation needed. tabix-js never creates a pool on its own — the
|
|
150
|
-
|
|
163
|
+
consumer controls the thread budget.
|
|
151
164
|
|
|
152
165
|
**Worth about 1.4x here, against the 1.95x a BAM reader reports.** Measured in
|
|
153
166
|
jbrowse-components on `test/data/1kg.chr1.subset.vcf.gz` — 213MB of 1000
|
|
@@ -155,12 +168,12 @@ Genomes, headless Chrome, real HTTP, four workers, arms interleaved, both
|
|
|
155
168
|
returning the same record count: **1.34-1.46x** across five window sizes and a
|
|
156
169
|
twelve-step pan.
|
|
157
170
|
|
|
158
|
-
The decompression itself moves **1.83x**.
|
|
159
|
-
|
|
160
|
-
|
|
171
|
+
The decompression itself moves **1.83x**. A **28% floor of per-line byte
|
|
172
|
+
scanning and string decoding** holds the end-to-end figure below that; no worker
|
|
173
|
+
count reaches it, and the floor is at its worst on multi-sample VCF, whose
|
|
161
174
|
records carry a genotype field per sample and run to ~60KB a line. A format with
|
|
162
|
-
narrower lines sits closer to BAM.
|
|
163
|
-
|
|
175
|
+
narrower lines sits closer to BAM. Getting more than ~1.5x on a multi-sample VCF
|
|
176
|
+
means attacking the scan, not the decompression.
|
|
164
177
|
|
|
165
178
|
Worker counts, lifecycle and benchmarks:
|
|
166
179
|
[bgzf-filehandle's worker pool docs](https://github.com/GMOD/bgzf-filehandle/blob/main/docs/worker-pool.md);
|
package/dist/csi.js
CHANGED
|
@@ -34,10 +34,13 @@ class CSI extends indexFile_ts_1.default {
|
|
|
34
34
|
}
|
|
35
35
|
const minShift = dataView.getInt32(4, true);
|
|
36
36
|
const depth = dataView.getInt32(8, true);
|
|
37
|
-
const maxBinNumber = (
|
|
37
|
+
const maxBinNumber = (8 ** (depth + 1) - 1) / 7;
|
|
38
38
|
const maxRefLength = 2 ** (minShift + depth * 3);
|
|
39
39
|
const auxLength = dataView.getInt32(12, true);
|
|
40
|
-
|
|
40
|
+
// htslib names no references in a GAF index, so its aux is the bare
|
|
41
|
+
// 28-byte header block
|
|
42
|
+
const isGaf = auxLength >= 28 && (dataView.getInt32(16, true) & 0xf) === 3;
|
|
43
|
+
const aux = auxLength >= 30 || isGaf
|
|
41
44
|
? (0, util_ts_1.parseAuxData)(bytes, 16)
|
|
42
45
|
: {
|
|
43
46
|
refIdToName: [],
|
package/dist/csi.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"csi.js","sourceRoot":"","sources":["../src/csi.ts"],"names":[],"mappings":";;;;;AAAA,0DAA8B;AAC9B,kEAAsC;AACtC,uCAMkB;AAClB,yDAA8C;AAK9C,MAAM,UAAU,GAAG,UAAU,CAAA,CAAC,QAAQ;AACtC,MAAM,UAAU,GAAG,UAAU,CAAA,CAAC,QAAQ;AAEtC,gFAAgF;AAChF,wDAAwD;AACxD,SAAS,MAAM,CAAC,GAAW,EAAE,IAAY;IACvC,OAAO,GAAG,GAAG,CAAC,IAAI,IAAI,CAAA;AACxB,CAAC;AACD,SAAS,MAAM,CAAC,GAAW,EAAE,IAAY;IACvC,OAAO,IAAI,CAAC,KAAK,CAAC,GAAG,GAAG,CAAC,IAAI,IAAI,CAAC,CAAA;AACpC,CAAC;AAED,MAAqB,GAAI,SAAQ,sBAAS;IACxC,gBAAgB;IAChB,KAAK,CAAC,MAAM,CAAC,OAAgB,EAAE;QAC7B,MAAM,EAAE,KAAK,EAAE,QAAQ,EAAE,GAAG,MAAM,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,CAAA;QAE3D,MAAM,KAAK,GAAG,QAAQ,CAAC,SAAS,CAAC,CAAC,EAAE,IAAI,CAAC,CAAA;QACzC,IAAI,UAAU,CAAA;QACd,IAAI,KAAK,KAAK,UAAU,EAAE,CAAC;YACzB,UAAU,GAAG,CAAC,CAAA;QAChB,CAAC;aAAM,IAAI,KAAK,KAAK,UAAU,EAAE,CAAC;YAChC,UAAU,GAAG,CAAC,CAAA;QAChB,CAAC;aAAM,CAAC;YACN,MAAM,IAAI,KAAK,CAAC,yBAAyB,KAAK,GAAG,CAAC,CAAA;QACpD,CAAC;QAED,MAAM,QAAQ,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAC,EAAE,IAAI,CAAC,CAAA;QAC3C,MAAM,KAAK,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAC,EAAE,IAAI,CAAC,CAAA;QACxC,MAAM,YAAY,GAAG,CAAC,CAAC,
|
|
1
|
+
{"version":3,"file":"csi.js","sourceRoot":"","sources":["../src/csi.ts"],"names":[],"mappings":";;;;;AAAA,0DAA8B;AAC9B,kEAAsC;AACtC,uCAMkB;AAClB,yDAA8C;AAK9C,MAAM,UAAU,GAAG,UAAU,CAAA,CAAC,QAAQ;AACtC,MAAM,UAAU,GAAG,UAAU,CAAA,CAAC,QAAQ;AAEtC,gFAAgF;AAChF,wDAAwD;AACxD,SAAS,MAAM,CAAC,GAAW,EAAE,IAAY;IACvC,OAAO,GAAG,GAAG,CAAC,IAAI,IAAI,CAAA;AACxB,CAAC;AACD,SAAS,MAAM,CAAC,GAAW,EAAE,IAAY;IACvC,OAAO,IAAI,CAAC,KAAK,CAAC,GAAG,GAAG,CAAC,IAAI,IAAI,CAAC,CAAA;AACpC,CAAC;AAED,MAAqB,GAAI,SAAQ,sBAAS;IACxC,gBAAgB;IAChB,KAAK,CAAC,MAAM,CAAC,OAAgB,EAAE;QAC7B,MAAM,EAAE,KAAK,EAAE,QAAQ,EAAE,GAAG,MAAM,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,CAAA;QAE3D,MAAM,KAAK,GAAG,QAAQ,CAAC,SAAS,CAAC,CAAC,EAAE,IAAI,CAAC,CAAA;QACzC,IAAI,UAAU,CAAA;QACd,IAAI,KAAK,KAAK,UAAU,EAAE,CAAC;YACzB,UAAU,GAAG,CAAC,CAAA;QAChB,CAAC;aAAM,IAAI,KAAK,KAAK,UAAU,EAAE,CAAC;YAChC,UAAU,GAAG,CAAC,CAAA;QAChB,CAAC;aAAM,CAAC;YACN,MAAM,IAAI,KAAK,CAAC,yBAAyB,KAAK,GAAG,CAAC,CAAA;QACpD,CAAC;QAED,MAAM,QAAQ,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAC,EAAE,IAAI,CAAC,CAAA;QAC3C,MAAM,KAAK,GAAG,QAAQ,CAAC,QAAQ,CAAC,CAAC,EAAE,IAAI,CAAC,CAAA;QACxC,MAAM,YAAY,GAAG,CAAC,CAAC,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAA;QAC/C,MAAM,YAAY,GAAG,CAAC,IAAI,CAAC,QAAQ,GAAG,KAAK,GAAG,CAAC,CAAC,CAAA;QAChD,MAAM,SAAS,GAAG,QAAQ,CAAC,QAAQ,CAAC,EAAE,EAAE,IAAI,CAAC,CAAA;QAC7C,oEAAoE;QACpE,uBAAuB;QACvB,MAAM,KAAK,GAAG,SAAS,IAAI,EAAE,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,EAAE,EAAE,IAAI,CAAC,GAAG,GAAG,CAAC,KAAK,CAAC,CAAA;QAC1E,MAAM,GAAG,GACP,SAAS,IAAI,EAAE,IAAI,KAAK;YACtB,CAAC,CAAC,IAAA,sBAAY,EAAC,KAAK,EAAE,EAAE,CAAC;YACzB,CAAC,CAAC;gBACE,WAAW,EAAE,EAAc;gBAC3B,WAAW,EAAE,EAA4B;gBACzC,QAAQ,EAAE,SAAS;gBACnB,aAAa,EAAE,EAAE,GAAG,EAAE,CAAC,EAAE,KAAK,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE;gBAC3C,cAAc,EAAE,sBAAsB;gBACtC,MAAM,EAAE,SAAS;aAClB,CAAA;QACP,MAAM,QAAQ,GAAG,QAAQ,CAAC,QAAQ,CAAC,EAAE,GAAG,SAAS,EAAE,IAAI,CAAC,CAAA;QAExD,iEAAiE;QACjE,gFAAgF;QAChF,IAAI,IAAI,GAAG,EAAE,GAAG,SAAS,GAAG,CAAC,CAAA;QAC7B,IAAI,aAAwC,CAAA;QAC5C,MAAM,OAAO,GAAa,EAAE,CAAA;QAE5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,QAAQ,EAAE,CAAC,EAAE,EAAE,CAAC;YAClC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAA;YAClB,MAAM,QAAQ,GAAG,QAAQ,CAAC,QAAQ,CAAC,IAAI,EAAE,IAAI,CAAC,CAAA;YAC9C,IAAI,IAAI,CAAC,CAAA;YACT,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,QAAQ,EAAE,CAAC,EAAE,EAAE,CAAC;gBAClC,MAAM,GAAG,GAAG,QAAQ,CAAC,SAAS,CAAC,IAAI,EAAE,IAAI,CAAC,CAAA;gBAC1C,IAAI,IAAI,CAAC,CAAA;gBACT,IAAI,GAAG,GAAG,YAAY,EAAE,CAAC;oBACvB,IAAI,IAAI,EAAE,GAAG,EAAE,CAAA,CAAC,gDAAgD;gBAClE,CAAC;qBAAM,CAAC;oBACN,aAAa,GAAG,IAAA,0BAAgB,EAAC,KAAK,EAAE,IAAI,EAAE,CAAC,EAAE,aAAa,CAAC,CAAA;oBAC/D,IAAI,IAAI,CAAC,CAAA,CAAC,UAAU;oBACpB,MAAM,UAAU,GAAG,QAAQ,CAAC,QAAQ,CAAC,IAAI,EAAE,IAAI,CAAC,CAAA;oBAChD,IAAI,IAAI,CAAC,GAAG,EAAE,GAAG,UAAU,CAAA;gBAC7B,CAAC;YACH,CAAC;QACH,CAAC;QAED,SAAS,UAAU,CAAC,KAAa;YAC/B,MAAM,KAAK,GAAG,OAAO,CAAC,KAAK,CAAC,CAAA;YAC5B,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;gBACxB,OAAO,SAAS,CAAA;YAClB,CAAC;YACD,IAAI,GAAG,GAAG,KAAK,CAAA;YACf,MAAM,QAAQ,GAAG,QAAQ,CAAC,QAAQ,CAAC,GAAG,EAAE,IAAI,CAAC,CAAA;YAC7C,GAAG,IAAI,CAAC,CAAA;YACR,MAAM,QAAQ,GAA4B,EAAE,CAAA;YAC5C,MAAM,QAAQ,GAAkC,EAAE,CAAA;YAClD,IAAI,KAAK,CAAA;YACT,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,QAAQ,EAAE,CAAC,EAAE,EAAE,CAAC;gBAClC,MAAM,GAAG,GAAG,QAAQ,CAAC,SAAS,CAAC,GAAG,EAAE,IAAI,CAAC,CAAA;gBACzC,GAAG,IAAI,CAAC,CAAA;gBACR,IAAI,GAAG,GAAG,YAAY,EAAE,CAAC;oBACvB,KAAK,GAAG,IAAA,wBAAc,EAAC,KAAK,EAAE,GAAG,GAAG,EAAE,CAAC,CAAA;oBACvC,GAAG,IAAI,EAAE,GAAG,EAAE,CAAA;gBAChB,CAAC;qBAAM,CAAC;oBACN,QAAQ,CAAC,GAAG,CAAC,GAAG,IAAA,4BAAS,EAAC,KAAK,EAAE,GAAG,CAAC,CAAA;oBACrC,GAAG,IAAI,CAAC,CAAA;oBACR,MAAM,UAAU,GAAG,QAAQ,CAAC,QAAQ,CAAC,GAAG,EAAE,IAAI,CAAC,CAAA;oBAC/C,GAAG,IAAI,CAAC,CAAA;oBACR,MAAM,MAAM,GAAG,KAAK,CAAC,IAAI,CAAQ,EAAE,MAAM,EAAE,UAAU,EAAE,CAAC,CAAA;oBACxD,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,UAAU,EAAE,CAAC,EAAE,EAAE,CAAC;wBACpC,MAAM,CAAC,CAAC,CAAC,GAAG,IAAI,kBAAK,CACnB,IAAA,4BAAS,EAAC,KAAK,EAAE,GAAG,CAAC,EACrB,IAAA,4BAAS,EAAC,KAAK,EAAE,GAAG,GAAG,CAAC,CAAC,EACzB,GAAG,CACJ,CAAA;wBACD,GAAG,IAAI,EAAE,CAAA;oBACX,CAAC;oBACD,QAAQ,CAAC,GAAG,CAAC,GAAG,MAAM,CAAA;gBACxB,CAAC;YACH,CAAC;YACD,IAAA,wBAAc,EAAC,MAAM,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC,IAAI,EAAE,CAAC,CAAA;YAC9C,OAAO,EAAE,QAAQ,EAAE,QAAQ,EAAE,KAAK,EAAE,CAAA;QACtC,CAAC;QAED,OAAO;YACL,GAAG,GAAG;YACN,GAAG,EAAE,IAAI;YACT,QAAQ;YACR,YAAY,EAAE,CAAC,IAAI,EAAE;YACrB,aAAa;YACb,UAAU;YACV,OAAO,EAAE,IAAA,wBAAc,EAAC,UAAU,CAAC;YACnC,QAAQ;YACR,KAAK;YACL,YAAY;YACZ,YAAY;SACb,CAAA;IACH,CAAC;IAED;;;;;;OAMG;IACO,YAAY,CACpB,GAAa,EACb,GAAW,EACX,EAAE,QAAQ,EAAE,KAAK,EAAa;QAE9B,MAAM,EAAE,QAAQ,EAAE,GAAG,GAAG,CAAA;QACxB,IAAI,KAAgC,CAAA;QACpC,IAAI,QAAQ,EAAE,CAAC;YACb,0DAA0D;YAC1D,IAAI,GAAG,GACL,CAAC,MAAM,CAAC,CAAC,EAAE,CAAC,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,IAAI,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC,EAAE,QAAQ,CAAC,CAAA;YACrE,OAAO,KAAK,KAAK,SAAS,IAAI,GAAG,GAAG,CAAC,EAAE,CAAC;gBACtC,KAAK,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAA;gBACrB,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;oBACxB,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,CAAA;oBACxC,MAAM,UAAU,GAAG,MAAM,GAAG,CAAC,GAAG,CAAC,CAAA;oBACjC,GAAG,GAAG,GAAG,GAAG,UAAU,CAAC,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAA;gBAC3C,CAAC;YACH,CAAC;YACD,KAAK,KAAK,QAAQ,CAAC,CAAC,CAAC,CAAA;QACvB,CAAC;QACD,OAAO,KAAK,CAAA;IACd,CAAC;IAES,QAAQ,CAChB,GAAW,EACX,GAAW,EACX,EAAE,QAAQ,EAAE,KAAK,EAAE,YAAY,EAAa;QAE5C,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC,CAAA;QACtB,MAAM,MAAM,GAAG,CAAC,IAAI,CAAC,QAAQ,GAAG,KAAK,GAAG,CAAC,CAAC,CAAA;QAC1C,IAAI,GAAG,GAAG,MAAM,EAAE,CAAC;YACjB,GAAG,GAAG,MAAM,CAAA;QACd,CAAC;QACD,GAAG,IAAI,CAAC,CAAA;QACR,IAAI,CAAC,GAAG,CAAC,CAAA;QACT,IAAI,CAAC,GAAG,CAAC,CAAA;QACT,IAAI,CAAC,GAAG,QAAQ,GAAG,KAAK,GAAG,CAAC,CAAA;QAC5B,MAAM,IAAI,GAAkC,EAAE,CAAA;QAC9C,OAAO,CAAC,IAAI,KAAK,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC,IAAI,MAAM,CAAC,CAAC,EAAE,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,IAAI,CAAC,EAAE,CAAC;YACzD,MAAM,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,GAAG,EAAE,CAAC,CAAC,CAAA;YAC5B,MAAM,CAAC,GAAG,CAAC,GAAG,MAAM,CAAC,GAAG,EAAE,CAAC,CAAC,CAAA;YAC5B,IAAI,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,MAAM,GAAG,YAAY,EAAE,CAAC;gBACvC,MAAM,IAAI,KAAK,CACb,SAAS,GAAG,IAAI,GAAG,mDAAmD,QAAQ,WAAW,KAAK,0DAA0D,CACzJ,CAAA;YACH,CAAC;YACD,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,CAAU,CAAC,CAAA;QAC5B,CAAC;QACD,OAAO,IAAI,CAAA;IACb,CAAC;CACF;AA3KD,sBA2KC"}
|
package/dist/indexFile.d.ts
CHANGED
|
@@ -1,6 +1,21 @@
|
|
|
1
1
|
import type Chunk from './chunk.ts';
|
|
2
2
|
import type VirtualOffset from './virtualOffset.ts';
|
|
3
3
|
import type { GenericFilehandle } from 'generic-filehandle2';
|
|
4
|
+
/**
|
|
5
|
+
* The virtual offset past which a coordinate-sorted file holds nothing
|
|
6
|
+
* overlapping `[.., end)`, from the binning index alone: htslib's `max_off`
|
|
7
|
+
* (`hts_itr_query` in hts.c).
|
|
8
|
+
*
|
|
9
|
+
* Walk right from the finest bin after the one holding `end - 1`, stepping up
|
|
10
|
+
* to the parent at every first child, so each bin visited begins at or past
|
|
11
|
+
* `end` and never overlaps the query. Every record in such a bin starts at or
|
|
12
|
+
* past `end`, so the first chunk of the first bin that exists is a record past
|
|
13
|
+
* the query, and in a sorted file so is every record after it.
|
|
14
|
+
*
|
|
15
|
+
* SYNC: ~/src/gmod/bam-js/src/indexFile.ts maxOffset, and see bam-js ADR 0023
|
|
16
|
+
* for why the caller drops whole merged chunks with it rather than trimming.
|
|
17
|
+
*/
|
|
18
|
+
export declare function maxOffset(binIndex: Record<number, Chunk[]>, end: number, minShift: number, depth: number): VirtualOffset | undefined;
|
|
4
19
|
export interface Options {
|
|
5
20
|
signal?: AbortSignal;
|
|
6
21
|
/**
|
|
@@ -111,6 +126,7 @@ export default abstract class IndexFile {
|
|
|
111
126
|
* The chunks of the data file that may hold records overlapping the region.
|
|
112
127
|
* The two index formats differ only in their binning scheme and in where
|
|
113
128
|
* they keep the pruning floor, which is what the two hooks above supply.
|
|
129
|
+
* The ceiling, `maxOffset`, needs only the bins, so it serves both.
|
|
114
130
|
*
|
|
115
131
|
* @internal
|
|
116
132
|
*/
|
package/dist/indexFile.js
CHANGED
|
@@ -1,8 +1,57 @@
|
|
|
1
1
|
"use strict";
|
|
2
2
|
Object.defineProperty(exports, "__esModule", { value: true });
|
|
3
|
+
exports.maxOffset = maxOffset;
|
|
3
4
|
const bgzf_filehandle_1 = require("@gmod/bgzf-filehandle");
|
|
4
5
|
const shared_read_cache_1 = require("@gmod/shared-read-cache");
|
|
5
6
|
const util_ts_1 = require("./util.js");
|
|
7
|
+
/**
|
|
8
|
+
* The virtual offset past which a coordinate-sorted file holds nothing
|
|
9
|
+
* overlapping `[.., end)`, from the binning index alone: htslib's `max_off`
|
|
10
|
+
* (`hts_itr_query` in hts.c).
|
|
11
|
+
*
|
|
12
|
+
* Walk right from the finest bin after the one holding `end - 1`, stepping up
|
|
13
|
+
* to the parent at every first child, so each bin visited begins at or past
|
|
14
|
+
* `end` and never overlaps the query. Every record in such a bin starts at or
|
|
15
|
+
* past `end`, so the first chunk of the first bin that exists is a record past
|
|
16
|
+
* the query, and in a sorted file so is every record after it.
|
|
17
|
+
*
|
|
18
|
+
* SYNC: ~/src/gmod/bam-js/src/indexFile.ts maxOffset, and see bam-js ADR 0023
|
|
19
|
+
* for why the caller drops whole merged chunks with it rather than trimming.
|
|
20
|
+
*/
|
|
21
|
+
function maxOffset(binIndex, end, minShift, depth) {
|
|
22
|
+
if (end > 2 ** (minShift + depth * 3)) {
|
|
23
|
+
return undefined;
|
|
24
|
+
}
|
|
25
|
+
const binCount = (8 ** (depth + 1) - 1) / 7;
|
|
26
|
+
let bin = (8 ** depth - 1) / 7 + Math.floor((end - 1) / 2 ** minShift) + 1;
|
|
27
|
+
if (bin >= binCount) {
|
|
28
|
+
bin = 0;
|
|
29
|
+
}
|
|
30
|
+
for (;;) {
|
|
31
|
+
while (bin % 8 === 1) {
|
|
32
|
+
bin = (bin - 1) / 8;
|
|
33
|
+
}
|
|
34
|
+
if (bin === 0) {
|
|
35
|
+
return undefined;
|
|
36
|
+
}
|
|
37
|
+
const chunks = binIndex[bin];
|
|
38
|
+
if (chunks?.length) {
|
|
39
|
+
let lowest = chunks[0].minv;
|
|
40
|
+
for (let i = 1; i < chunks.length; i++) {
|
|
41
|
+
const minv = chunks[i].minv;
|
|
42
|
+
if (minv.compareTo(lowest) < 0) {
|
|
43
|
+
lowest = minv;
|
|
44
|
+
}
|
|
45
|
+
}
|
|
46
|
+
return lowest;
|
|
47
|
+
}
|
|
48
|
+
bin++;
|
|
49
|
+
}
|
|
50
|
+
}
|
|
51
|
+
// htslib files every GAF record under ref 0 and ignores a query's name
|
|
52
|
+
function refIdFor(indexData, refName) {
|
|
53
|
+
return indexData.format === 'GAF' ? 0 : indexData.refNameToId[refName];
|
|
54
|
+
}
|
|
6
55
|
class IndexFile {
|
|
7
56
|
filehandle;
|
|
8
57
|
/**
|
|
@@ -31,7 +80,7 @@ class IndexFile {
|
|
|
31
80
|
/** @internal */
|
|
32
81
|
async lineCount(refName, opts = {}) {
|
|
33
82
|
const indexData = await this.parse(opts);
|
|
34
|
-
const refId = indexData
|
|
83
|
+
const refId = refIdFor(indexData, refName);
|
|
35
84
|
if (refId === undefined) {
|
|
36
85
|
return -1;
|
|
37
86
|
}
|
|
@@ -46,12 +95,13 @@ class IndexFile {
|
|
|
46
95
|
* The chunks of the data file that may hold records overlapping the region.
|
|
47
96
|
* The two index formats differ only in their binning scheme and in where
|
|
48
97
|
* they keep the pruning floor, which is what the two hooks above supply.
|
|
98
|
+
* The ceiling, `maxOffset`, needs only the bins, so it serves both.
|
|
49
99
|
*
|
|
50
100
|
* @internal
|
|
51
101
|
*/
|
|
52
102
|
async blocksForRange(refName, min, max, opts = {}) {
|
|
53
103
|
const indexData = await this.parse(opts);
|
|
54
|
-
const refId = indexData
|
|
104
|
+
const refId = refIdFor(indexData, refName);
|
|
55
105
|
if (refId === undefined) {
|
|
56
106
|
return [];
|
|
57
107
|
}
|
|
@@ -59,6 +109,12 @@ class IndexFile {
|
|
|
59
109
|
if (!ba) {
|
|
60
110
|
return [];
|
|
61
111
|
}
|
|
112
|
+
// htslib indexes a GAF read as [minNode, maxNode), one short of the
|
|
113
|
+
// [minNode, maxNode] that getLines matches, so look one node lower, as
|
|
114
|
+
// `tabix` does for a region's 1-based start
|
|
115
|
+
if (indexData.format === 'GAF') {
|
|
116
|
+
min = Math.max(min - 1, 0);
|
|
117
|
+
}
|
|
62
118
|
// Find chunks in overlapping bins. Leaf bins are not pruned.
|
|
63
119
|
const chunks = [];
|
|
64
120
|
for (const [start, end] of this.reg2bins(min, max, indexData)) {
|
|
@@ -71,7 +127,16 @@ class IndexFile {
|
|
|
71
127
|
}
|
|
72
128
|
}
|
|
73
129
|
}
|
|
74
|
-
|
|
130
|
+
const merged = (0, util_ts_1.optimizeChunks)(chunks, this.lowestOffset(ba, min, indexData));
|
|
131
|
+
const past = maxOffset(ba.binIndex, max, indexData.minShift, indexData.depth);
|
|
132
|
+
if (past) {
|
|
133
|
+
let n = merged.length;
|
|
134
|
+
while (n > 0 && merged[n - 1].minv.compareTo(past) >= 0) {
|
|
135
|
+
n--;
|
|
136
|
+
}
|
|
137
|
+
merged.length = n;
|
|
138
|
+
}
|
|
139
|
+
return merged;
|
|
75
140
|
}
|
|
76
141
|
// SYNC: ~/src/gmod/bam-js/src/indexFile.ts parse — same shape and the same
|
|
77
142
|
// reasoning below.
|
package/dist/indexFile.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"indexFile.js","sourceRoot":"","sources":["../src/indexFile.ts"],"names":[],"mappings":";;
|
|
1
|
+
{"version":3,"file":"indexFile.js","sourceRoot":"","sources":["../src/indexFile.ts"],"names":[],"mappings":";;AAuBA,8BAkCC;AAzDD,2DAA6C;AAC7C,+DAAyD;AAEzD,uCAA0C;AAM1C;;;;;;;;;;;;;GAaG;AACH,SAAgB,SAAS,CACvB,QAAiC,EACjC,GAAW,EACX,QAAgB,EAChB,KAAa;IAEb,IAAI,GAAG,GAAG,CAAC,IAAI,CAAC,QAAQ,GAAG,KAAK,GAAG,CAAC,CAAC,EAAE,CAAC;QACtC,OAAO,SAAS,CAAA;IAClB,CAAC;IACD,MAAM,QAAQ,GAAG,CAAC,CAAC,IAAI,CAAC,KAAK,GAAG,CAAC,CAAC,GAAG,CAAC,CAAC,GAAG,CAAC,CAAA;IAC3C,IAAI,GAAG,GAAG,CAAC,CAAC,IAAI,KAAK,GAAG,CAAC,CAAC,GAAG,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,GAAG,GAAG,CAAC,CAAC,GAAG,CAAC,IAAI,QAAQ,CAAC,GAAG,CAAC,CAAA;IAC1E,IAAI,GAAG,IAAI,QAAQ,EAAE,CAAC;QACpB,GAAG,GAAG,CAAC,CAAA;IACT,CAAC;IACD,SAAS,CAAC;QACR,OAAO,GAAG,GAAG,CAAC,KAAK,CAAC,EAAE,CAAC;YACrB,GAAG,GAAG,CAAC,GAAG,GAAG,CAAC,CAAC,GAAG,CAAC,CAAA;QACrB,CAAC;QACD,IAAI,GAAG,KAAK,CAAC,EAAE,CAAC;YACd,OAAO,SAAS,CAAA;QAClB,CAAC;QACD,MAAM,MAAM,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAA;QAC5B,IAAI,MAAM,EAAE,MAAM,EAAE,CAAC;YACnB,IAAI,MAAM,GAAG,MAAM,CAAC,CAAC,CAAE,CAAC,IAAI,CAAA;YAC5B,KAAK,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,EAAE,EAAE,CAAC;gBACvC,MAAM,IAAI,GAAG,MAAM,CAAC,CAAC,CAAE,CAAC,IAAI,CAAA;gBAC5B,IAAI,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC;oBAC/B,MAAM,GAAG,IAAI,CAAA;gBACf,CAAC;YACH,CAAC;YACD,OAAO,MAAM,CAAA;QACf,CAAC;QACD,GAAG,EAAE,CAAA;IACP,CAAC;AACH,CAAC;AA6CD,uEAAuE;AACvE,SAAS,QAAQ,CAAC,SAAoB,EAAE,OAAe;IACrD,OAAO,SAAS,CAAC,MAAM,KAAK,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,WAAW,CAAC,OAAO,CAAC,CAAA;AACxE,CAAC;AAED,MAA8B,SAAS;IAC9B,UAAU,CAAmB;IACpC;;;OAGG;IACK,UAAU,GAAG,IAAI,mCAAe,CAAoB,EAAE,CAAC,CAAA;IAE/D,YAAY,EAAE,UAAU,EAAqC;QAC3D,IAAI,CAAC,UAAU,GAAG,UAAU,CAAA;IAC9B,CAAC;IAyBD;;;OAGG;IACO,KAAK,CAAC,cAAc,CAAC,IAAa;QAC1C,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,UAAU,CAAC,QAAQ,CAAC;YACzC,MAAM,EAAE,IAAI,CAAC,MAAM;YACnB,UAAU,EAAE,IAAI,CAAC,UAAU;SAC5B,CAAC,CAAA;QACF,MAAM,KAAK,GAAG,MAAM,IAAA,uBAAK,EAAC,GAAG,CAAC,CAAA;QAC9B,OAAO;YACL,KAAK;YACL,QAAQ,EAAE,IAAI,QAAQ,CAAC,KAAK,CAAC,MAAM,EAAE,KAAK,CAAC,UAAU,EAAE,KAAK,CAAC,UAAU,CAAC;SACzE,CAAA;IACH,CAAC;IAED,gBAAgB;IACT,KAAK,CAAC,SAAS,CAAC,OAAe,EAAE,OAAgB,EAAE;QACxD,MAAM,SAAS,GAAG,MAAM,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;QACxC,MAAM,KAAK,GAAG,QAAQ,CAAC,SAAS,EAAE,OAAO,CAAC,CAAA;QAC1C,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YACxB,OAAO,CAAC,CAAC,CAAA;QACX,CAAC;QACD,OAAO,SAAS,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,KAAK,EAAE,SAAS,IAAI,CAAC,CAAC,CAAA;IACzD,CAAC;IAED,gBAAgB;IACT,KAAK,CAAC,WAAW,CAAC,OAAgB,EAAE;QACzC,MAAM,EAAE,OAAO,EAAE,QAAQ,EAAE,GAAG,IAAI,EAAE,GAAG,MAAM,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;QAC7D,OAAO,IAAI,CAAA;IACb,CAAC;IAED;;;;;;;OAOG;IACI,KAAK,CAAC,cAAc,CACzB,OAAe,EACf,GAAW,EACX,GAAW,EACX,OAAgB,EAAE;QAElB,MAAM,SAAS,GAAG,MAAM,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;QACxC,MAAM,KAAK,GAAG,QAAQ,CAAC,SAAS,EAAE,OAAO,CAAC,CAAA;QAC1C,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YACxB,OAAO,EAAE,CAAA;QACX,CAAC;QACD,MAAM,EAAE,GAAG,SAAS,CAAC,OAAO,CAAC,KAAK,CAAC,CAAA;QACnC,IAAI,CAAC,EAAE,EAAE,CAAC;YACR,OAAO,EAAE,CAAA;QACX,CAAC;QACD,oEAAoE;QACpE,uEAAuE;QACvE,4CAA4C;QAC5C,IAAI,SAAS,CAAC,MAAM,KAAK,KAAK,EAAE,CAAC;YAC/B,GAAG,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,GAAG,CAAC,EAAE,CAAC,CAAC,CAAA;QAC5B,CAAC;QAED,6DAA6D;QAC7D,MAAM,MAAM,GAAY,EAAE,CAAA;QAC1B,KAAK,MAAM,CAAC,KAAK,EAAE,GAAG,CAAC,IAAI,IAAI,CAAC,QAAQ,CAAC,GAAG,EAAE,GAAG,EAAE,SAAS,CAAC,EAAE,CAAC;YAC9D,KAAK,IAAI,GAAG,GAAG,KAAK,EAAE,GAAG,IAAI,GAAG,EAAE,GAAG,EAAE,EAAE,CAAC;gBACxC,MAAM,SAAS,GAAG,EAAE,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAA;gBAClC,IAAI,SAAS,EAAE,CAAC;oBACd,KAAK,MAAM,CAAC,IAAI,SAAS,EAAE,CAAC;wBAC1B,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAA;oBAChB,CAAC;gBACH,CAAC;YACH,CAAC;QACH,CAAC;QAED,MAAM,MAAM,GAAG,IAAA,wBAAc,EAAC,MAAM,EAAE,IAAI,CAAC,YAAY,CAAC,EAAE,EAAE,GAAG,EAAE,SAAS,CAAC,CAAC,CAAA;QAC5E,MAAM,IAAI,GAAG,SAAS,CACpB,EAAE,CAAC,QAAQ,EACX,GAAG,EACH,SAAS,CAAC,QAAQ,EAClB,SAAS,CAAC,KAAK,CAChB,CAAA;QACD,IAAI,IAAI,EAAE,CAAC;YACT,IAAI,CAAC,GAAG,MAAM,CAAC,MAAM,CAAA;YACrB,OAAO,CAAC,GAAG,CAAC,IAAI,MAAM,CAAC,CAAC,GAAG,CAAC,CAAE,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC;gBACzD,CAAC,EAAE,CAAA;YACL,CAAC;YACD,MAAM,CAAC,MAAM,GAAG,CAAC,CAAA;QACnB,CAAC;QACD,OAAO,MAAM,CAAA;IACf,CAAC;IAED,2EAA2E;IAC3E,mBAAmB;IACnB;;;;;;;;;;;;;;;;;;;;;;;;;OAyBG;IACH,KAAK,CAAC,OAAgB,EAAE;QACtB,OAAO,IAAI,CAAC,UAAU,CAAC,GAAG,CAAC,OAAO,EAAE,IAAI,CAAC,MAAM,EAAE,MAAM,CAAC,EAAE,CACxD,IAAI,CAAC,MAAM,CAAC,EAAE,GAAG,IAAI,EAAE,MAAM,EAAE,CAAC,CACjC,CAAA;IACH,CAAC;IAED,gBAAgB;IAChB,KAAK,CAAC,SAAS,CAAC,KAAa,EAAE,OAAgB,EAAE;QAC/C,MAAM,GAAG,GAAG,MAAM,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAA;QAClC,OAAO,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,QAAQ,CAAA;IACvC,CAAC;CACF;AAtKD,4BAsKC"}
|