Feat/company blacklist tokenized input (#219)

* initial commit

* docs mention!

* Update orchestrator/src/server/pipeline/steps/discover-jobs.ts

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* normalizeStringArray

* poppier orange

* comments

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

shared/src/normalize-string-array.test.ts

@@ -0,0 +1,22 @@
+import { describe, expect, it } from "vitest";
+import { normalizeStringArray } from "./normalize-string-array";
+
+describe("normalizeStringArray", () => {
+  it("returns empty array for nullish/empty input", () => {
+    expect(normalizeStringArray(undefined)).toEqual([]);
+    expect(normalizeStringArray(null)).toEqual([]);
+    expect(normalizeStringArray([])).toEqual([]);
+  });
+
+  it("trims values and removes empty entries", () => {
+    expect(normalizeStringArray(["  staffing  ", " ", "\n"])).toEqual([
+      "staffing",
+    ]);
+  });
+
+  it("deduplicates values case-insensitively while preserving first casing", () => {
+    expect(
+      normalizeStringArray(["Recruit", "staffing", "recruit", "STAFFING"]),
+    ).toEqual(["Recruit", "staffing"]);
+  });
+});

shared/src/normalize-string-array.ts

@@ -0,0 +1,21 @@
+export function normalizeStringArray(
+  values: readonly string[] | null | undefined,
+): string[] {
+  if (!values || values.length === 0) return [];
+
+  const seen = new Set<string>();
+  const normalized: string[] = [];
+
+  for (const value of values) {
+    const trimmed = value.trim();
+    if (!trimmed) continue;
+
+    const key = trimmed.toLowerCase();
+    if (seen.has(key)) continue;
+
+    seen.add(key);
+    normalized.push(trimmed);
+  }
+
+  return normalized;
+}

shared/src/settings-registry.ts

@@ -177,6 +177,13 @@ export const settingsRegistry = {
     parse: parseJsonArrayOrNull,
     serialize: serializeNullableJsonArray,
   },
+  blockedCompanyKeywords: {
+    kind: "typed" as const,
+    schema: z.array(z.string().trim().min(1).max(200)).max(200),
+    default: (): string[] => [],
+    parse: parseJsonArrayOrNull,
+    serialize: serializeNullableJsonArray,
+  },
   searchCities: {
     kind: "typed" as const,
     schema: z.string().trim().max(100),

shared/src/testing/factories.ts

@@ -156,6 +156,11 @@ export const createAppSettings = (
     default: ["Software Engineer"],
     override: null,
   },
+  blockedCompanyKeywords: {
+    value: [],
+    default: [],
+    override: null,
+  },
   searchCities: {
     value: "United Kingdom",
     default: "United Kingdom",

shared/src/types/settings.ts

@@ -123,6 +123,7 @@ export interface AppSettings {
   adzunaMaxJobsPerTerm: Resolved<number>;
   gradcrackerMaxJobsPerTerm: Resolved<number>;
   searchTerms: Resolved<string[]>;
+  blockedCompanyKeywords: Resolved<string[]>;
   searchCities: Resolved<string>;
   jobspyResultsWanted: Resolved<number>;
   jobspyCountryIndeed: Resolved<string>;