refactor(evidence): extract TypeScript lifecycle (#32)

This commit is contained in:
2026-08-24 02:50:56 +02:00
parent 840848df94
commit f375515dc0
8 changed files with 432 additions and 135 deletions
@@ -0,0 +1,34 @@
import { readdirSync, readFileSync } from "node:fs";
import { basename, join } from "node:path";
import { fileURLToPath } from "node:url";
import ts from "typescript";
import { expect, test } from "vitest";
const evidenceRoot = fileURLToPath(new URL("../../../src/workspaces/evidence/", import.meta.url));
const coreInfrastructure = new Set([
"preprocessing-service",
"qdrant-collection",
"registry",
]);
test("Evidence modules do not import core-owned registry or Qdrant lifecycle", () => {
const violations: Array<{ file: string; dependency: string }> = [];
for (const file of readdirSync(evidenceRoot).filter((name) => name.endsWith(".ts"))) {
const source = ts.createSourceFile(
file,
readFileSync(join(evidenceRoot, file), "utf8"),
ts.ScriptTarget.Latest,
true,
ts.ScriptKind.TS,
);
for (const statement of source.statements) {
if (!ts.isImportDeclaration(statement) || !ts.isStringLiteral(statement.moduleSpecifier)) {
continue;
}
const dependency = basename(statement.moduleSpecifier.text).replace(/\.js$/, "");
if (coreInfrastructure.has(dependency)) violations.push({ file, dependency });
}
}
expect(violations).toEqual([]);
});
@@ -0,0 +1,134 @@
import { execFile } from "node:child_process";
import { mkdtempSync, mkdirSync, readFileSync, rmSync, symlinkSync, writeFileSync, readdirSync } from "node:fs";
import { tmpdir } from "node:os";
import { join } from "node:path";
import { promisify } from "node:util";
import { afterEach, expect, test } from "vitest";
import { GitWorkspaceRepository } from "../../../src/workspaces/git-repository.js";
import { materializeEvidenceTree } from "../../../src/workspaces/evidence/materialization.js";
import type { WorkspaceRegistryConfig } from "../../../src/workspaces/types.js";
const runFile = promisify(execFile);
const temporaryRoots: string[] = [];
afterEach(() => {
temporaryRoots.splice(0).forEach((root) => rmSync(root, { recursive: true, force: true }));
});
async function git(cwd: string, args: string[]): Promise<string> {
const { stdout } = await runFile("git", args, { cwd });
return stdout.trim();
}
function config(root: string, remoteUrl: string): WorkspaceRegistryConfig {
return {
root,
remoteUrl,
branch: "main",
gitAuthorName: "Evidence Materializer Test",
gitAuthorEmail: "evidence-materializer@example.invalid",
installationId: "test",
secretRoots: [],
maxImportBytes: 1024,
maxImportEntries: 1,
};
}
async function fixture(): Promise<{ root: string; remote: string; commit: string }> {
const root = mkdtempSync(join(tmpdir(), "thoth-evidence-materializer-"));
temporaryRoots.push(root);
const remote = join(root, "remote.git");
const source = join(root, "source");
await git(root, ["init", "--bare", "--initial-branch=main", remote]);
mkdirSync(source);
await git(source, ["init", "--initial-branch=main"]);
await git(source, ["config", "user.name", "Evidence Materializer Test"]);
await git(source, ["config", "user.email", "evidence-materializer@example.invalid"]);
writeFileSync(join(source, "thoth-workspaces.yaml"), "schema_version: 1\nworkspaces: [{id: research, name: Research}]\n");
mkdirSync(join(source, "research", "evidence", "nested"), { recursive: true });
writeFileSync(join(source, "research", "workspace.yaml"), "workspace:\n schema_version: 3\n id: research\n");
writeFileSync(join(source, "research", "evidence", "guide.md"), "# guide\n");
writeFileSync(join(source, "research", "evidence", "nested", "deep.md"), "# deep\n");
await git(source, ["add", "-A"]);
await git(source, ["commit", "-m", "initial"]);
await git(source, ["remote", "add", "origin", remote]);
await git(source, ["push", "origin", "main"]);
const commit = await git(source, ["rev-parse", "HEAD"]);
return { root, remote, commit };
}
async function repo(fixture: { root: string; remote: string }): Promise<GitWorkspaceRepository> {
const repository = new GitWorkspaceRepository(config(join(fixture.root, "registry"), fixture.remote));
await repository.bootstrap();
return repository;
}
test("materializes the tree, hashes every file, and writes a bounded manifest", async () => {
const fixtureValue = await fixture();
const repository = await repo(fixtureValue);
const target = mkdtempSync(join(tmpdir(), "thoth-evidence-target-"));
temporaryRoots.push(target);
const result = await materializeEvidenceTree({
repository,
revision: fixtureValue.commit,
id: "research",
targetDirectory: target,
});
expect(readFileSync(join(result.root, "guide.md"), "utf8")).toBe("# guide\n");
expect(readFileSync(join(result.root, "nested", "deep.md"), "utf8")).toBe("# deep\n");
expect(result.manifest).toMatchObject({
schemaVersion: 1,
workspace: "research",
commit: fixtureValue.commit,
entryCount: 2,
});
expect(Object.keys(result.manifest.files).sort()).toEqual(["guide.md", "nested/deep.md"]);
expect(result.manifest.files["guide.md"]!.digest).toMatch(/^sha256:[0-9a-f]{64}$/);
expect(result.manifestDigest).toMatch(/^[0-9a-f]{64}$/);
expect(readFileSync(result.manifestPath, "utf8")).toContain('"entryCount":2');
});
test("refuses symlink-containing trees and bound violations without publishing", async () => {
const root = mkdtempSync(join(tmpdir(), "thoth-evidence-unsafe-"));
temporaryRoots.push(root);
const remote = join(root, "remote.git");
const source = join(root, "source");
await git(root, ["init", "--bare", "--initial-branch=main", remote]);
mkdirSync(source);
await git(source, ["init", "--initial-branch=main"]);
await git(source, ["config", "user.name", "E"]);
await git(source, ["config", "user.email", "e@e.invalid"]);
writeFileSync(join(source, "thoth-workspaces.yaml"), "schema_version: 1\nworkspaces: [{id: research, name: Research}]\n");
mkdirSync(join(source, "research", "evidence"), { recursive: true });
writeFileSync(join(source, "research", "workspace.yaml"), "workspace:\n schema_version: 3\n id: research\n");
writeFileSync(join(source, "research", "outside.md"), "# outside\n");
symlinkSync("../outside.md", join(source, "research", "evidence", "link.md"));
await git(source, ["add", "-A"]);
await git(source, ["commit", "-m", "symlink"]);
await git(source, ["remote", "add", "origin", remote]);
await git(source, ["push", "origin", "main"]);
const commit = await git(source, ["rev-parse", "HEAD"]);
const repository = await repo({ root, remote });
const target = mkdtempSync(join(tmpdir(), "thoth-evidence-unsafe-target-"));
temporaryRoots.push(target);
await expect(materializeEvidenceTree({ repository, revision: commit, id: "research", targetDirectory: target }))
.rejects.toThrow();
expect(readdirSync(target)).toEqual([]);
// A valid tree but a per-file bound of 1 byte must also refuse.
const fixtureValue = await fixture();
const repository2 = await repo(fixtureValue);
const target2 = mkdtempSync(join(tmpdir(), "thoth-evidence-bound-target-"));
temporaryRoots.push(target2);
await expect(materializeEvidenceTree({
repository: repository2,
revision: fixtureValue.commit,
id: "research",
targetDirectory: target2,
limits: { maxFileBytes: 1 },
})).rejects.toThrow();
expect(readdirSync(target2)).toEqual([]);
});
@@ -0,0 +1,161 @@
import { expect, test, vi } from "vitest";
import {
continueEvidencePreprocessing,
preprocessEvidence,
type EvidenceJobState,
type EvidencePreprocessingDependencies,
} from "../../../src/workspaces/evidence/preprocessing.js";
import type { WorkspaceDescriptor } from "../../../src/workspaces/schema.js";
type EvidenceConfig = NonNullable<WorkspaceDescriptor["evidence"]>;
const filesystemEvidence = {
source: { type: "filesystem", uri: "research/evidence" },
} as EvidenceConfig;
const privateHttpEvidence = {
source: {
type: "http",
uris: ["http://127.0.0.1/private.md"],
authentication: "none",
connect_timeout_ms: 1000,
read_timeout_ms: 2000,
max_bytes: 100,
max_redirects: 0,
allow_private_hosts: true,
max_cache_bytes: 100,
},
} as EvidenceConfig;
function job(overrides: Partial<EvidenceJobState> = {}): EvidenceJobState {
return {
runId: "a".repeat(32),
childRuns: {},
completedStages: [],
...overrides,
};
}
function dependencies(payload: Record<string, unknown> = {}): EvidencePreprocessingDependencies & {
runStage: ReturnType<typeof vi.fn>;
persistJob: ReturnType<typeof vi.fn>;
semanticPreflight: ReturnType<typeof vi.fn>;
} {
return {
runStage: vi.fn(async () => payload),
persistJob: vi.fn(),
semanticPreflight: vi.fn(async () => ({ ok: true as const })),
requireRunId(value) {
if (typeof value !== "string" || !/^[0-9a-f]{32}$/.test(value)) {
throw new Error("child run id is invalid");
}
return value;
},
numberRecord(value) {
if (!value || typeof value !== "object" || Array.isArray(value)) return undefined;
return Object.fromEntries(
Object.entries(value as Record<string, unknown>).map(([key, nested]) => [key, Number(nested)]),
);
},
};
}
test("owns the standalone Evidence stage argv and mutation order", async () => {
const state = job({ childRuns: { evidence: "b".repeat(32) } });
const deps = dependencies({ run_id: "c".repeat(32), counts: { added: 2 } });
const result = await preprocessEvidence(
{ evidence: filesystemEvidence, job: state, dryRun: false },
deps,
);
expect(deps.semanticPreflight).toHaveBeenCalledOnce();
expect(deps.runStage).toHaveBeenCalledWith([
"preprocess", "evidence", "--resume", "b".repeat(32), "--json", "-c", "/dev/fd/3",
]);
expect(deps.persistJob).toHaveBeenCalledOnce();
expect(state).toMatchObject({
childRuns: { evidence: "c".repeat(32) },
completedStages: ["evidence"],
});
expect(result).toEqual({
status: "succeeded",
code: "ok",
runId: "a".repeat(32),
childRuns: { evidence: "c".repeat(32) },
completedStages: ["evidence"],
counts: { added: 2 },
});
});
test("owns Evidence egress refusal before shared semantic infrastructure", async () => {
const deps = dependencies();
const result = await preprocessEvidence(
{
evidence: privateHttpEvidence,
job: job(),
httpPrivateHostAllowlist: ["metadata.internal"],
},
deps,
);
expect(result).toEqual({ status: "failed", code: "egress_policy_refused" });
expect(deps.semanticPreflight).not.toHaveBeenCalled();
expect(deps.runStage).not.toHaveBeenCalled();
expect(deps.persistJob).not.toHaveBeenCalled();
});
test("projects aggregate no-Evidence and completed-stage outcomes without rerunning", async () => {
const deps = dependencies();
const noEvidence = await continueEvidencePreprocessing(
{
evidence: undefined,
job: job({ completedStages: ["dwh", "schema_index"] }),
priorCounts: { added: 2 },
},
deps,
);
const completed = await continueEvidencePreprocessing(
{
evidence: filesystemEvidence,
job: job({ completedStages: ["dwh", "schema_index", "evidence"] }),
},
deps,
);
expect(noEvidence).toMatchObject({
status: "succeeded",
code: "ok",
warnings: ["workspace has no Evidence source"],
counts: { added: 2 },
});
expect(completed).toMatchObject({
status: "unchanged",
code: "ok",
completedStages: ["dwh", "schema_index", "evidence"],
});
expect(deps.runStage).not.toHaveBeenCalled();
expect(deps.persistJob).not.toHaveBeenCalled();
});
test("preserves the narrow standalone projection for an already completed Evidence stage", async () => {
const deps = dependencies();
const result = await preprocessEvidence(
{
evidence: filesystemEvidence,
job: job({ childRuns: { evidence: "b".repeat(32) }, completedStages: ["evidence"] }),
},
deps,
);
expect(result).toEqual({
status: "unchanged",
code: "ok",
runId: "a".repeat(32),
completedStages: ["evidence"],
});
expect(deps.runStage).not.toHaveBeenCalled();
expect(deps.persistJob).not.toHaveBeenCalled();
});