@jbrowse/bandage-core 0.1.20 → 4.0.25

package/dist/cli/figure.js

@@ -1,38 +1,27 @@
 #!/usr/bin/env node
 import {
-  cutWindowGFA,
+  HPRC_GBZ,
+  cutGbzRegion,
+  featuresOnBackbone,
   figureSvg,
   forceLayout,
-  haplotypeWanted,
+  genesFromBed,
+  genesFromGff3Lines,
+  genesFromText,
+  graphBackbone,
   layoutModeByValue,
   loadBandage,
   loadGraph,
-  referencePathQuery,
-  referenceSamplesOf,
-  resolveReferenceSample
-} from "../chunks/chunk-3XDR3ZBS.js";
+  openGbz,
+  parseRegion
+} from "../chunks/chunk-B5ZHLU4N.js";
 
 // src/cli/figure.ts
 import { open, readFile, writeFile } from "node:fs/promises";
 import path from "node:path";
 import { parseArgs } from "node:util";
 import { gunzipSync } from "node:zlib";
-import { GBZBase } from "@gmod/gbz-base";
-var HPRC = {
-  db: "https://s3-us-west-2.amazonaws.com/human-pangenomics/pangenomes/freeze/release2/minigraph-cactus/v2.1/hprc-v2.1-mc-grch38/hprc-v2.1-mc-grch38.gbz.db",
-  index: "https://jbrowse.org/demos/hprc/hprc-v2.1-mc-grch38.haplotype-index.anchored.db"
-};
-var WALK_LIMIT = 1e5;
-function parseRegion(text) {
-  const m = /^\s*([^:\s]+):([\d,]+)-([\d,]+)\s*$/.exec(text);
-  if (!m) {
-    throw new Error(
-      `"${text}" is not a region like chr6:160,614,798-160,647,758`
-    );
-  }
-  const n = (s) => Number(s.replaceAll(",", ""));
-  return { refName: m[1], start: n(m[2]), end: n(m[3]) };
-}
+import { TabixIndexedFile } from "@gmod/tabix";
 var isUrl = (s) => /^https?:\/\//.test(s);
 function byteSource(location) {
   if (isUrl(location)) {
@@ -64,36 +53,40 @@ function byteSource(location) {
     }
   };
 }
-async function gfaText(location) {
+async function text(location) {
   const bytes = isUrl(location) ? new Uint8Array(await (await fetch(location)).arrayBuffer()) : await readFile(location);
   const gzipped = bytes[0] === 31 && bytes[1] === 139;
   return new TextDecoder().decode(gzipped ? gunzipSync(bytes) : bytes);
 }
-async function cutGbz(gbz) {
-  const { db, index } = gbz.db === "hprc" ? HPRC : gbz;
-  const base = await GBZBase.open(
-    byteSource(db),
-    index ? { haplotypeIndex: byteSource(index) } : {}
+async function readGenes(genes, graph, resolve) {
+  const backbone = graphBackbone(graph);
+  if (!backbone) {
+    return [];
+  }
+  const file = resolve(genes.file);
+  const format = genes.format ?? (/\.bed(\.gz)?$/i.test(genes.file) ? "bed" : "gff3");
+  if (!genes.index) {
+    return featuresOnBackbone(genesFromText(await text(file)), backbone);
+  }
+  const index = resolve(genes.index);
+  const csi = index.endsWith(".csi");
+  const tabix = new TabixIndexedFile(
+    isUrl(file) ? { url: file, ...csi ? { csiUrl: index } : { tbiUrl: index } } : { path: file, ...csi ? { csiPath: index } : { tbiPath: index } }
   );
-  const samples = await referenceSamplesOf(base);
-  const sample = resolveReferenceSample({
-    configured: "",
-    anchorPrefix: gbz.referenceSample ?? samples[0] ?? "",
-    referenceSamples: samples
-  });
-  const region = parseRegion(gbz.region);
-  const query = await referencePathQuery(base, sample, region.refName);
-  if (!query) {
-    throw new Error(`${sample} has no indexed path named ${region.refName}`);
+  const indexed = new Set(await tabix.getReferenceSequenceNames());
+  const lines = [];
+  for (const contig of backbone.contigs) {
+    const refName = [contig.contig, contig.refName].find((n) => indexed.has(n));
+    if (refName !== void 0) {
+      await tabix.getLines(refName, contig.start, contig.end, {
+        lineCallback: (line) => lines.push(line)
+      });
+    }
   }
-  const wanted = gbz.haplotypes?.length ? gbz.haplotypes : void 0;
-  const text = await cutWindowGFA(base, query, region.start, region.end, {
-    context: 1e3,
-    snarls: "contained",
-    limit: WALK_LIMIT,
-    ...wanted ? { keep: (name) => haplotypeWanted(name, wanted) } : {}
-  });
-  return { text, region, name: `${sample} ${gbz.region}` };
+  return featuresOnBackbone(
+    format === "gff3" ? genesFromGff3Lines(lines) : genesFromBed(lines.join("\n")),
+    backbone
+  );
 }
 var engine = async (request) => {
   const bandage = await loadBandage();
@@ -101,15 +94,26 @@ var engine = async (request) => {
   const result = bandage.computeLayout(request.graph, request.options);
   return { result, duration: performance.now() - start };
 };
-async function renderSpec(spec, base = ".") {
-  const source = spec.gbz ? await cutGbz(spec.gbz) : spec.gfa ? {
-    text: await gfaText(
-      isUrl(spec.gfa) ? spec.gfa : path.resolve(base, spec.gfa)
-    ),
-    region: spec.region ? parseRegion(spec.region) : void 0,
-    name: path.basename(spec.gfa)
-  } : void 0;
-  if (!source) {
+async function renderSpec(spec, base) {
+  const resolve = (location) => isUrl(location) ? location : path.resolve(base, location);
+  let source;
+  if (spec.gbz) {
+    const { db, index } = spec.gbz.db === "hprc" ? HPRC_GBZ : spec.gbz;
+    const cut = await cutGbzRegion(
+      await openGbz(
+        byteSource(resolve(db)),
+        index ? byteSource(resolve(index)) : void 0
+      ),
+      spec.gbz
+    );
+    source = { ...cut, name: `${cut.sample} ${spec.gbz.region}` };
+  } else if (spec.gfa) {
+    source = {
+      text: await text(resolve(spec.gfa)),
+      name: path.basename(spec.gfa),
+      region: spec.region ? parseRegion(spec.region) : void 0
+    };
+  } else {
     throw new Error("a spec names its graph as `gfa` or `gbz`");
   }
   const graph = loadGraph(source.text, source.name, {
@@ -128,7 +132,8 @@ async function renderSpec(spec, base = ".") {
     ...spec,
     walks: spec.walks?.map((w) => typeof w === "string" ? { walk: w } : w),
     region: source.region,
-    metadata: JSON.stringify(spec)
+    genes: spec.genes ? await readGenes(spec.genes, graph, resolve) : void 0,
+    spec
   });
 }
 async function main() {