The easy-to-use open source Business Intelligence and Embedded Analytics tool that lets everyone work with data :bar_chart:
TypeScript analyticsbibusiness-intelligencebusinessintelligenceclojuredashboarddatadata-analysisdata-visualizationdatabasemetabasemysqlpostgrespostgresqlreportingslacksql-editorvisualization
// Backend (Clojure/hawk) adapter: hawk's JUnit XML → NormalizedTest[].
//
// hawk always writes JUnit XML to `target/junit/`, so the backend "collector"
// is a post-run scan of that artifact (one runner, one artifact, covering both
// the backend and driver CI paths). The XML parsing itself is shared
// (`../junit.ts`); this owns the backend-specific parts — hawk's JUnit file
// layout — while the suite label is supplied by the entrypoint.
import { readFileSync } from "node:fs";
import type { NormalizedTest } from "../contract.ts";
import { findJunitFiles, parseJunit } from "../junit.ts";
import { log } from "../util.ts";
const JUNIT_DIR = process.env.JUNIT_DIR || "target/junit";
// hawk names its per-namespace JUnit files after the namespace, whatever its
// name (`metabase.app_db.general_schema_tests.xml`, not only `*_test.xml`),
// plus one fixed-name file (`mb_hawk_var_less_errors.xml`) for errors it can't
// tie to a test var (a fixture-init throw or a namespace load/compile error).
// That var-less file carries a `<testcase>` with a `name` but no `classname`,
// so it parses to a namespace-less failure — which is exactly what makes the
// granular-rerun collector fall back to a full rerun instead of a misleadingly
// narrow one. hawk clears `target/junit` at the start of a run and is its only
// writer, so every `.xml` file there is hawk's. Miss one and its failures
// vanish from JUnit.
const selectHawkJunit = (entries: string[]): string[] =>
entries.filter((entry) => entry.endsWith(".xml"));
/**
* Normalize every JUnit file hawk wrote under `dir` into `NormalizedTest[]` —
* the source-specific half of the adapter pattern (hawk XML → the shared
* normalized shape).
*/
export function normalizeBackendJunit(
ignorePassingTests: boolean = true,
dir: string = JUNIT_DIR,
): NormalizedTest[] {
const files = findJunitFiles(dir, selectHawkJunit);
const failures = files.flatMap((file) => {
try {
return parseJunit(readFileSync(file, "utf8"), ignorePassingTests);
} catch (error) {
console.error(`[ci-conductor] failed to read ${file}`, error);
return [];
}
});
log(
`scanned ${dir}: ${files.length} JUnit file(s), ${failures.length} ${ignorePassingTests ? "failing " : ""}test(s)`,
);
for (const test of failures) {
log(
` ${test.status ?? "failure"}: ${test.path || "(no namespace)"} / ${test.name}`,
);
}
return failures;
}