Findings normalization

Sverka normalizes output from any analysis tool into a single Finding[] model. Source: packages/verification/src/findings/.

Finding modelCopy Anchor Link

Every finding has:

normalizeSarif(sarif, context)Copy Anchor Link

Convert SARIF 2.1.0 output to Finding[].

import { normalizeSarif } from "@sverka/sdk";

const findings = normalizeSarif(sarifLog, {
  root: "/path/to/project",
  checkIdPrefix: "eslint",
  defaultConfidence: 0.5,
});

The normalizer:

• Resolves rule IDs (via ruleId or ruleIndex).
• Maps SARIF severity levels to Sverka's Severity enum.
• Resolves file paths relative to the project root.
• Computes fingerprints for each finding.

computeFingerprint(input)Copy Anchor Link

Deterministic SHA-256 fingerprint. Insensitive to message wording and severity changes; sensitive to file, rule, and line range.

import { computeFingerprint } from "@sverka/sdk";

const fp = computeFingerprint({
  checkId: "eslint",
  rule: "no-unused-vars",
  file: "src/index.ts",
  startLine: 10,
  endLine: 10,
});

Returns a lowercase hex string.

BaselinesCopy Anchor Link

Baselines track known findings so only new ones are reported.

createBaseline(findings)Copy Anchor Link

Create a baseline from a set of findings.

import { createBaseline } from "@sverka/sdk";

const baseline = createBaseline(findings);

updateBaseline(findings, baseline)Copy Anchor Link

Update an existing baseline with new findings.

loadBaseline(path)Copy Anchor Link

Load a baseline from a JSON file.

saveBaseline(baseline, path)Copy Anchor Link

Save a baseline to a JSON file.

filterOnlyNew(findings, baseline)Copy Anchor Link

Filter findings to only those not in the baseline.

import { filterOnlyNew } from "@sverka/sdk";

const newFindings = filterOnlyNew(allFindings, baseline);

A finding is "new" if its fingerprint is not in the baseline's fingerprint set and not suppressed by a suppression entry.