Address #287 review: configurable browser prediction and evaluation

.github/workflows/build-wasm.yml

@@ -2,7 +2,6 @@ name: WebAssembly and web interface
 
 on:
   pull_request:
-    branches: [master]
   push:
     branches: [master]
   workflow_dispatch:
@@ -36,7 +35,9 @@ jobs:
       - name: Cross-compile dependencies and IntaRNA
         run: bash scripts/wasm/build.sh
       - name: Check real WASM predictions against the native reference
-        run: node scripts/wasm/smoke.mjs
+        run: |
+          node scripts/wasm/smoke.mjs
+          node scripts/wasm/smoke-protocol.mjs
       - name: Stage the static site and notices
         run: bash scripts/wasm/stage.sh
       - name: Install browser test tools

.github/workflows/build.yml

@@ -4,7 +4,6 @@ on:
   push:
     branches: [master]
   pull_request:
-    branches: [master]
 
 defaults:
   run:

ChangeLog

@@ -14,6 +14,11 @@
 
 ## Interface and handling
 
+- extend the browser interface with vaRRI styling, prediction/evaluation tasks,
+  multi-FASTA, start indices, CLI parameter groups, selectable CSV columns and
+  additional output previews/downloads; label pair-relative differences mfe Δ
+  and display actual compiled versions (PR #287 review)
+
 - display sortable/filterable browser results and native CSV downloads;
   publish the tested interface alongside existing Pages documentation (#286)
 
@@ -75,6 +80,10 @@
 
 ## Technical changes and Optimizations
 
+- annotate inherited virtual methods/destructors with override, preserve compiled
+  WASM asset provenance through staging, and expand protocol/browser validation
+  for the PR #287 review
+
 - cross-compile IntaRNA and pinned dependencies to WebAssembly (#286, step 1)
 - verify browser predictions against native accessibility/energy references and
   exercise real WASM cancellation and nested Pages paths in Chromium/Firefox
@@ -232,6 +241,29 @@ energies, restricted partition sums, and trackers.
 ################################################################################  
 ################################################################################
 
+261021 Alexander Mitrofanov
+ * src/IntaRNA/*.h
+   * Mark 194 inherited virtual methods and 45 derived destructors override;
+     keep algorithms, interfaces and third-party headers unchanged.
+ * webinterface/{options,columns,protocol,app,results,intarna-worker}.js
+   + Add CLI-based parameter groups and validation, multi-FASTA prediction,
+     signed per-strand start indices and IntaRNAeval structure evaluation.
+   + Provide all CSV columns, table visibility controls, pair-relative mfe Δ,
+     local SHAPE/accessibility data and previews/downloads of native extra files.
+ * webinterface/index.html, styles.css, assets/*
+   * Follow vaRRI main-v2 colors and institutional header; keep one parameter
+     group open at a time and adapt controls/defaults to the selected task.
+ * scripts/wasm/{build,stage}.sh, write-build-info.mjs
+   + Record actual compiled engine/library versions and asset hashes in text/JSON
+     at build time; preserve/check provenance during staging and show it in UI.
+   * Reject inconsistent missing overrides in the Emscripten build.
+ * webinterface/tests/*, scripts/wasm/{smoke-protocol.mjs,test-browser.py}
+   + Exercise task validation/defaults, real WASM evaluation/prediction, multiple
+     records/coordinates, imported data, columns/reports and worker lifecycle.
+ * .github/workflows/{build,build-wasm}.yml, README.md, webinterface/docs/*
+   * Run WASM/browser checks on dependent PRs and document revised behavior,
+     provenance, native option replacements and browser limits (PR #287 review).
+
 261020 Alexander Mitrofanov
  * scripts/wasm/*, .github/workflows/build-wasm.yml, webinterface/app.js
    + Execute dependency API probes and test structured accessibility costs,

README.md

@@ -5,8 +5,8 @@
 **Efficient RNA-RNA interaction prediction incorporating accessibility and
 seeding of interaction sites**
 
-Predict short RNA pairs locally with the [browser interface](https://backofenlab.github.io/IntaRNA/webinterface/).
-See the [WebAssembly build instructions](docs/step1-wasm-compilation.md) and
+Predict short RNA pairs or evaluate supplied structures locally with the [browser interface](https://backofenlab.github.io/IntaRNA/webinterface/).
+See the [WebAssembly build instructions](webinterface/docs/step1-wasm-compilation.md) and
 [interface/deployment notes](webinterface/docs/step4-results-and-deployment.md).
 
 

scripts/wasm/build.sh

@@ -89,7 +89,7 @@ export VRNA_LIBS="-L$prefix/lib -lRNA -lm"
 (cd "$repo" && bash autotools-init.sh)
 mkdir -p "$build/intarna"
 (cd "$build/intarna"
- CPPFLAGS="-I$build/intarna/src -I$repo/src $CPPFLAGS" emconfigure "$repo/configure" --host=wasm32-unknown-emscripten \
+ CXXFLAGS="$CXXFLAGS -Werror=inconsistent-missing-override" CPPFLAGS="-I$build/intarna/src -I$repo/src $CPPFLAGS" emconfigure "$repo/configure" --host=wasm32-unknown-emscripten \
    --disable-multithreading --disable-log-coloring --with-mdspan=kokkos \
    --with-vrna="$prefix" --with-boost="$prefix" \
    --with-boost-libdir="$prefix/lib" --with-zlib="$prefix"
@@ -99,4 +99,7 @@ mkdir -p "$build/intarna"
  cp src/bin/IntaRNA.wasm "$build/site/wasm/intarna.wasm")
 # ES modules also work in Node smoke tests without changing package scope.
 printf '{"type":"module"}\n' > "$build/site/wasm/package.json"
+node "$repo/scripts/wasm/write-build-info.mjs" "$repo" "$build/site" \
+  "$(emcc -dumpversion)" "$(git -C "$repo" rev-parse HEAD)" \
+  "$(git -C "$repo" status --porcelain --untracked-files=no)"
 echo "WASM assets: $build/site/wasm"

scripts/wasm/smoke-protocol.mjs

@@ -0,0 +1,47 @@
+// Integration coverage of browser argument translation against the real CLI.
+import assert from 'node:assert/strict';
+import { resolve } from 'node:path';
+import { pathToFileURL } from 'node:url';
+import { DEFAULTS, EVALUATION_EXAMPLE, prepareRequest, makeArguments } from '../../webinterface/protocol.js';
+import { CSV_CATALOG } from '../../webinterface/columns.js';
+import { parseResults } from '../../webinterface/results.js';
+const asset = resolve(process.argv[2] || 'build-wasm/site/wasm/intarna.js');
+const { default: createIntaRNA } = await import(pathToFileURL(asset));
+async function run(input, format = 'C') {
+  const request = prepareRequest({ ...DEFAULTS, ...input });
+  const diagnostics = [];
+  const module = await createIntaRNA({ noInitialRun: true, locateFile: () => new URL('./intarna.wasm', pathToFileURL(asset)).href, print: (line) => diagnostics.push(line), printErr: (line) => diagnostics.push(line) });
+  module.FS.writeFile('/query.fa', request.query.fasta);
+  module.FS.writeFile('/target.fa', request.target.fasta);
+  module.FS.mkdir('/output');
+  for (const [key, value] of Object.entries(request.parameters)) if (/^(q|t)(Shape|AccFile)$/.test(key) && value) module.FS.writeFile(`/${key}.txt`, value + '\n');
+  const args = makeArguments(request, format);
+  assert.equal(module.callMain(args), 0, `${args.join(' ')}\n${diagnostics.join('\n')}`);
+  const text = module.FS.readFile(format === 'C' ? '/results.csv' : '/report.txt', { encoding: 'utf8' });
+  return { rows: format === 'C' ? parseResults(text) : [], text, fs: module.FS };
+}
+const defaults = await run({ parameters: { outNumber: 10 } });
+assert.equal(defaults.rows.length, 10);
+assert.equal(defaults.rows[0].E, -21.4);
+const shifted = await run({ qIdxPos0: -3, tIdxPos0: -5 });
+assert.deepEqual(['start1', 'end1', 'start2', 'end2'].map((key) => shifted.rows[0][key]), [-5, 5, -3, 7]);
+const multi = await run({ query: '>q1\nGGGAGGAGGG\n>q2\nGGGAGGAGGG', target: '>t1\nCCCUCCUCCC\n>t2\nCCCUCCUCCC', outputs: ['qMinE', 'tMinE'] });
+assert.equal(multi.rows.length, 4);
+assert.ok(multi.rows.every((row) => row.deltaE === 0));
+assert.equal(multi.fs.readdir('/output').filter((name) => name.endsWith('.csv')).length, 8);
+const pairwise = await run({ query: '>q1\nGGGAGGAGGG\n>q2\nGGGAGGAGGG', target: '>t1\nCCCUCCUCCC\n>t2\nCCCUCCUCCC', parameters: { outPairwise: true } });
+assert.equal(pairwise.rows.length, 2);
+const evaluation = await run({ task: 'evaluation', rri: EVALUATION_EXAMPLE + ':1|||||||||&2|||||||||', outputs: ['qAcc', 'tPu'], csvColumns: CSV_CATALOG.map((c) => c.key) });
+assert.equal(evaluation.rows.length, 2);
+assert.equal(evaluation.rows[0].E, -21.4);
+assert.equal(evaluation.rows[0].seedE, 'NAN');
+assert.ok((await run({ task: 'evaluation', rri: EVALUATION_EXAMPLE }, 'D')).text.includes('-21.4'));
+assert.ok((await run({ task: 'evaluation', rri: EVALUATION_EXAMPLE }, 'E')).text.length > 0);
+const probabilities = await run({ parameters: { model: 'P' }, csvColumns: CSV_CATALOG.map((c) => c.key), outputs: ['bpProb', 'bpsvg', 'qSpotProb', 'tSpotProb', 'spotProb', 'pMinE'] });
+assert.equal(probabilities.rows.length, 1);
+assert.ok(probabilities.fs.readFile('/output/bpsvg.svg', { encoding: 'utf8' }).includes('<svg'));
+const accessibility = await run({ parameters: { outNumber: 0 }, outputs: ['qPu'] });
+const imported = await run({ parameters: { qAcc: 'P', qAccFile: accessibility.fs.readFile('/output/qPu.csv', { encoding: 'utf8' }) } });
+assert.equal(imported.rows[0].E, (await run({ parameters: { qIntLenMax: 9 } })).rows[0].E);
+for (const parameters of [{ energyVRNA: 'Turner99', temperature: 20 }, { energyVRNA: 'Andronescu07' }, { model: 'B' }, { mode: 'K' }, { noSeed: true, mode: 'M' }, { seedMaxUP: 1 }, { qAccConstr: '..........', qShape: '1 0.2\n2 0.3' }, { outOverlap: 'N', outNumber: 3 }, { outCsvSort: 'E', outNumber: 3 }]) assert.ok((await run({ parameters })).rows.length >= 1);
+console.log('Browser protocol: native-reference prediction/evaluation, signed coordinates, multi-FASTA, all columns, imported accessibility, energy/seed/helix/SHAPE options and output files passed');

scripts/wasm/stage.sh

@@ -7,7 +7,8 @@ site="$build/site"
 test -s "$site/wasm/intarna.js"
 test -s "$site/wasm/intarna.wasm"
 mkdir -p "$site/licenses"
-for file in index.html styles.css app.js mark.svg protocol.js runner.js results.js; do
+cp -R "$repo/webinterface/assets" "$site/"
+for file in index.html styles.css app.js mark.svg protocol.js runner.js results.js options.js columns.js; do
   cp "$repo/webinterface/$file" "$site/$file"
 done
 cp "$repo/webinterface/intarna-worker.js" "$site/intarna-worker.js"
@@ -42,18 +43,16 @@ cat > "$site/licenses/index.html" <<'HTML'
 </ul><p><a href="../build-info.json">Build provenance</a> · <a href="../">Back to IntaRNA</a></p>
 </html>
 HTML
-python3 - "$repo" "$site" <<'PY'
-import json
-import subprocess
-import sys
+# Provenance belongs to the compiled assets, never the later staging revision.
+python3 - "$site" <<'PYINFO'
+import hashlib, json, sys
 from pathlib import Path
-repo, site = sys.argv[1:]
-metadata = {
-    "source": "https://github.com/BackofenLab/IntaRNA",
-    "commit": subprocess.check_output(["git", "-C", repo, "rev-parse", "HEAD"], text=True).strip(),
-    "emscripten": "4.0.23", "boost": "1.87.0", "viennarna": "2.7.0", "zlib": "1.3.1",
-    "threading": "single worker per prediction",
-}
-Path(site, "build-info.json").write_text(json.dumps(metadata, indent=2) + "\n")
-PY
+site = Path(sys.argv[1])
+metadata = json.loads((site / "build-info.json").read_text())
+if not (site / "build-info.txt").is_file():
+    raise SystemExit("Missing compiled version text")
+for name, digest in metadata["sha256"].items():
+    if hashlib.sha256((site / "wasm" / name).read_bytes()).hexdigest() != digest:
+        raise SystemExit(f"Mismatched engine: {name}")
+PYINFO
 echo "Static site: $site"