[djleamen]

Summary

In isReasonableInput, the generic short-token gibberish guard

if (/^[a-z]{1,3}$/.test(trimmed) && ['o', 'c', 'n', 'h'].includes(trimmed)) return false;

ran before the per-type switch, so any 2–3 character lowercase SMILES not in the atom allowlist was rejected with "Invalid {sourceType}" before lookupChemicalIdentifiers ever consulted the local DB or PubChem. That silently breaks valid molecules such as CC (ethane), CCO (ethanol), CCN (ethylamine), and CCC (propane) — even ones present in the built-in commonCompounds table.

Change

Skip that heuristic for smiles and inchi, which already have dedicated format checks inside the switch. Other identifier types (IUPAC names, etc.) still get the short-gibberish guard unchanged.

Verification

Simulated isReasonableInput over representative inputs:

type	value	before	after
smiles	CC / CCO / CCN / CCC	rejected ❌	accepted ✅
smiles	c1ccccc1	accepted	accepted ✅
iupacName	xyz (gibberish)	rejected	rejected ✅
iupacName	water	accepted	accepted ✅
casNumber / pubchemCID	valid	accepted	accepted ✅

Single-atom SMILES (O, C, N) remain gated by the pre-existing length < 2 check — a separate limitation left out of scope here.

Relates to #4 (limited SMILES parser). The repo has no test framework configured, so the change is validated by logic simulation rather than a unit test.

---

Generated by Claude Code

[github-actions]

Dependency Review

✅ No vulnerabilities or license issues or OpenSSF Scorecard issues found.

Scanned Files

None

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[djleamen]

The build matrix is failing at npm ci with an ERESOLVE peer-dependency conflict ([email protected] in the root vs [email protected], which only supports eslint <=8). This happens during dependency install — before any of this PR's code is compiled — so it is not caused by this change; the same failure reproduces on main.

It's the exact conflict already tracked/fixed by #41. Once #41 lands, I'll merge main into this branch and CI should go green. This PR's diff is a single-line guard in isReasonableInput and is unrelated to the dependency graph.

---

Generated by Claude Code

[Copilot]

Pull request overview

This PR fixes pre-validation in src/utils/identifierLookup.ts so valid short SMILES strings (e.g., CC, CCO, CCN, CCC) are no longer rejected as “very short gibberish” before the per-identifier-type validation and lookup logic runs.

Changes:

• Skips the short lowercase-token “gibberish” heuristic for smiles and inchi.
• Adds inline comments explaining why short SMILES must bypass that heuristic.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[djleamen]

Heads-up on the red check here: the build (18.x) job fails at npm ci with an ERESOLVE between [email protected] and [email protected] (the 20.x/22.x jobs are fail-fast cancellations of the same run). This is a pre-existing dependency conflict on main, not something this change introduces — this PR only touches src/utils/identifierLookup.ts.

The actual fix is #41, which migrates to the ESLint 10 flat config, bumps eslint-plugin-react-hooks to v7 (its peer range covers eslint 10), and drops Node 18 from the CI matrix. Once #41 merges into main, updating this branch will clear the red check with no changes needed here.

---

Generated by Claude Code