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>

docs-site/docs/features/settings.md

@@ -119,6 +119,7 @@ Readiness requires:
 - Penalize missing salary data
 - Set penalty amount
 - Optional auto-skip threshold for low-score jobs
+- Block jobs from companies that match configured keyword tokens
 
 ### Danger Zone
 

orchestrator/src/client/pages/SettingsPage.test.tsx

@@ -261,4 +261,38 @@ describe("SettingsPage", () => {
       }),
     );
   });
+
+  it("saves blocked company keywords from scoring settings", async () => {
+    vi.mocked(api.getSettings).mockResolvedValue(baseSettings);
+    vi.mocked(api.updateSettings).mockResolvedValue({
+      ...baseSettings,
+      blockedCompanyKeywords: {
+        value: ["staffing"],
+        default: [],
+        override: ["staffing"],
+      },
+    });
+
+    renderPage();
+
+    const scoringTrigger = await screen.findByRole("button", {
+      name: /scoring settings/i,
+    });
+    fireEvent.click(scoringTrigger);
+
+    const input = screen.getByPlaceholderText('e.g. "recruitment", "staffing"');
+    fireEvent.change(input, { target: { value: "staffing" } });
+    fireEvent.keyDown(input, { key: "Enter" });
+
+    const saveButton = screen.getByRole("button", { name: /^save$/i });
+    await waitFor(() => expect(saveButton).toBeEnabled());
+    fireEvent.click(saveButton);
+
+    await waitFor(() => expect(api.updateSettings).toHaveBeenCalled());
+    expect(api.updateSettings).toHaveBeenCalledWith(
+      expect.objectContaining({
+        blockedCompanyKeywords: ["staffing"],
+      }),
+    );
+  });
 });

orchestrator/src/client/pages/SettingsPage.tsx

@@ -18,6 +18,7 @@ import {
   resumeProjectsEqual,
 } from "@client/pages/settings/utils";
 import { zodResolver } from "@hookform/resolvers/zod";
+import { normalizeStringArray } from "@shared/normalize-string-array.js";
 import {
   type UpdateSettingsInput,
   updateSettingsSchema,
@@ -72,6 +73,7 @@ const DEFAULT_FORM_VALUES: UpdateSettingsInput = {
   penalizeMissingSalary: null,
   missingSalaryPenalty: null,
   autoSkipScoreThreshold: null,
+  blockedCompanyKeywords: [],
 };
 
 type LlmProviderValue = LlmProviderId | null;
@@ -114,6 +116,7 @@ const NULL_SETTINGS_PAYLOAD: UpdateSettingsInput = {
   penalizeMissingSalary: null,
   missingSalaryPenalty: null,
   autoSkipScoreThreshold: null,
+  blockedCompanyKeywords: null,
 };
 
 const mapSettingsToForm = (data: AppSettings): UpdateSettingsInput => ({
@@ -149,6 +152,7 @@ const mapSettingsToForm = (data: AppSettings): UpdateSettingsInput => ({
   penalizeMissingSalary: data.penalizeMissingSalary.override,
   missingSalaryPenalty: data.missingSalaryPenalty.override,
   autoSkipScoreThreshold: data.autoSkipScoreThreshold.override,
+  blockedCompanyKeywords: data.blockedCompanyKeywords.override ?? [],
 });
 
 const normalizeString = (value: string | null | undefined) => {
@@ -162,6 +166,11 @@ const normalizePrivateInput = (value: string | null | undefined) => {
   return trimmed || undefined;
 };
 
+const stringArraysEqual = (left: string[], right: string[]): boolean => {
+  if (left.length !== right.length) return false;
+  return left.every((value, index) => value === right[index]);
+};
+
 const nullIfSame = <T,>(value: T | null | undefined, defaultValue: T) =>
   value === defaultValue ? null : (value ?? null);
 
@@ -291,6 +300,10 @@ const getDerivedSettings = (settings: AppSettings | null) => {
         effective: settings?.autoSkipScoreThreshold?.value ?? null,
         default: settings?.autoSkipScoreThreshold?.default ?? null,
       },
+      blockedCompanyKeywords: {
+        effective: settings?.blockedCompanyKeywords?.value ?? [],
+        default: settings?.blockedCompanyKeywords?.default ?? [],
+      },
     },
   };
 };
@@ -627,6 +640,15 @@ export const SettingsPage: React.FC = () => {
           data.missingSalaryPenalty,
           scoring.missingSalaryPenalty.default,
         ),
+        blockedCompanyKeywords: (() => {
+          const normalized = normalizeStringArray(data.blockedCompanyKeywords);
+          const normalizedDefault = normalizeStringArray(
+            scoring.blockedCompanyKeywords.default,
+          );
+          return stringArraysEqual(normalized, normalizedDefault)
+            ? null
+            : normalized;
+        })(),
         ...envPayload,
       };
 

orchestrator/src/client/pages/orchestrator/TokenizedInput.tsx

@@ -16,6 +16,7 @@ interface TokenizedInputProps {
   helperText: string;
   removeLabelPrefix: string;
   collapsedTextLimit?: number;
+  disabled?: boolean;
 }
 
 function mergeUnique(values: string[], nextValues: string[]): string[] {
@@ -41,6 +42,7 @@ export const TokenizedInput: React.FC<TokenizedInputProps> = ({
   helperText,
   removeLabelPrefix,
   collapsedTextLimit = 3,
+  disabled = false,
 }) => {
   const [isFocused, setIsFocused] = useState(false);
   const tokensRef = useRef<HTMLDivElement | null>(null);
@@ -118,6 +120,7 @@ export const TokenizedInput: React.FC<TokenizedInputProps> = ({
           }
         }}
         placeholder={placeholder}
+        disabled={disabled}
       />
       <p className="text-xs text-muted-foreground">{helperText}</p>
       {values.length > 0 ? (
@@ -152,6 +155,7 @@ export const TokenizedInput: React.FC<TokenizedInputProps> = ({
                     variant="outline"
                     className="h-auto rounded-full px-2 py-1 text-xs text-muted-foreground"
                     aria-label={`${removeLabelPrefix} ${value}`}
+                    disabled={disabled}
                     onPointerDown={(event) => event.preventDefault()}
                     onClick={() =>
                       onValuesChange(

orchestrator/src/client/pages/settings/components/ScoringSettingsSection.tsx

@@ -1,7 +1,9 @@
+import { TokenizedInput } from "@client/pages/orchestrator/TokenizedInput";
 import { SettingsInput } from "@client/pages/settings/components/SettingsInput";
 import type { ScoringValues } from "@client/pages/settings/types";
 import type { UpdateSettingsInput } from "@shared/settings-schema.js";
 import type React from "react";
+import { useState } from "react";
 import { Controller, useFormContext } from "react-hook-form";
 import {
   AccordionContent,
@@ -17,6 +19,13 @@ type ScoringSettingsSectionProps = {
   isSaving: boolean;
 };
 
+function parseTokenizedKeywordInput(input: string): string[] {
+  return input
+    .split(/[\n,]/g)
+    .map((value) => value.trim())
+    .filter(Boolean);
+}
+
 export const ScoringSettingsSection: React.FC<ScoringSettingsSectionProps> = ({
   values,
   isLoading,
@@ -26,8 +35,11 @@ export const ScoringSettingsSection: React.FC<ScoringSettingsSectionProps> = ({
     penalizeMissingSalary,
     missingSalaryPenalty,
     autoSkipScoreThreshold,
+    blockedCompanyKeywords,
   } = values;
-  const { control, watch } = useFormContext<UpdateSettingsInput>();
+  const { control, watch, setValue } = useFormContext<UpdateSettingsInput>();
+  const [blockedCompanyKeywordDraft, setBlockedCompanyKeywordDraft] =
+    useState("");
 
   // Watch the current form value to conditionally show/hide penalty input
   const currentPenalizeEnabled =
@@ -35,6 +47,8 @@ export const ScoringSettingsSection: React.FC<ScoringSettingsSectionProps> = ({
 
   // Watch auto-skip threshold to show current value
   const currentAutoSkipThreshold = watch("autoSkipScoreThreshold");
+  const blockedCompanyKeywordValues =
+    watch("blockedCompanyKeywords") ?? blockedCompanyKeywords.default;
 
   return (
     <AccordionItem value="scoring" className="border rounded-lg px-4">
@@ -154,6 +168,41 @@ export const ScoringSettingsSection: React.FC<ScoringSettingsSectionProps> = ({
 
           <Separator />
 
+          <div className="space-y-3">
+            <label
+              htmlFor="blocked-company-keywords"
+              className="text-sm font-medium leading-none"
+            >
+              Blocked Company Keywords
+            </label>
+            <TokenizedInput
+              id="blocked-company-keywords"
+              values={blockedCompanyKeywordValues}
+              draft={blockedCompanyKeywordDraft}
+              parseInput={parseTokenizedKeywordInput}
+              onDraftChange={setBlockedCompanyKeywordDraft}
+              onValuesChange={(value) =>
+                setValue("blockedCompanyKeywords", value, { shouldDirty: true })
+              }
+              placeholder='e.g. "recruitment", "staffing"'
+              helperText="Jobs whose company name contains one of these keywords will be dropped during discovery."
+              removeLabelPrefix="Remove blocked keyword"
+              disabled={isLoading || isSaving}
+            />
+            <div className="break-words font-mono text-xs text-muted-foreground">
+              Effective:{" "}
+              {blockedCompanyKeywordValues.length > 0
+                ? blockedCompanyKeywordValues.join(", ")
+                : "None"}{" "}
+              | Default:{" "}
+              {blockedCompanyKeywords.default.length > 0
+                ? blockedCompanyKeywords.default.join(", ")
+                : "None"}
+            </div>
+          </div>
+
+          <Separator />
+
           {/* Effective/Default values display */}
           <div className="grid gap-2 text-sm sm:grid-cols-2">
             <div>

orchestrator/src/client/pages/settings/types.ts

@@ -48,4 +48,5 @@ export type ScoringValues = {
   penalizeMissingSalary: EffectiveDefault<boolean>;
   missingSalaryPenalty: EffectiveDefault<number>;
   autoSkipScoreThreshold: EffectiveDefault<number | null>;
+  blockedCompanyKeywords: EffectiveDefault<string[]>;
 };

orchestrator/src/index.css

@@ -76,12 +76,12 @@
 :root {
   --radius: 0.5rem;
   --background: oklch(0.9818 0.0054 95.0986);
-  --foreground: oklch(0.3438 0.0269 95.7226);
+  --foreground: oklch(28.815% 0.01599 96.892);
   --card: oklch(0.9818 0.0054 95.0986);
   --card-foreground: oklch(0.1908 0.002 106.5859);
   --popover: oklch(1 0 0);
   --popover-foreground: oklch(0.2671 0.0196 98.939);
-  --primary: oklch(0.6171 0.1375 39.0427);
+  --primary: oklch(0.6916 0.1719 40.93);
   --primary-foreground: oklch(1 0 0);
   --secondary: oklch(0.9245 0.0138 92.9892);
   --secondary-foreground: oklch(0.4334 0.0177 98.6048);
@@ -143,7 +143,7 @@
   --card-foreground: oklch(0.9818 0.0054 95.0986);
   --popover: oklch(0.3085 0.0035 106.6039);
   --popover-foreground: oklch(0.9211 0.004 106.4781);
-  --primary: oklch(0.6724 0.1308 38.7559);
+  --primary: oklch(0.6916 0.1719 40.93);
   --primary-foreground: oklch(1 0 0);
   --secondary: oklch(0.9818 0.0054 95.0986);
   --secondary-foreground: oklch(0.3085 0.0035 106.6039);
@@ -155,7 +155,7 @@
   --border: oklch(0.3618 0.0101 106.8928);
   --input: oklch(0.4336 0.0113 100.2195);
   --ring: oklch(0.6724 0.1308 38.7559);
-  --chart-1: oklch(0.5583 0.1276 42.9956);
+  --chart-1: oklch(0.6916 0.1719 40.93);
   --chart-2: oklch(0.6898 0.1581 290.4107);
   --chart-3: oklch(0.213 0.0078 95.4245);
   --chart-4: oklch(0.3074 0.0516 289.323);

orchestrator/src/server/pipeline/steps/discover-jobs.test.ts

@@ -646,6 +646,44 @@ describe("discoverJobsStep", () => {
     expect(vi.mocked(ukVisa.runUkVisaJobs)).not.toHaveBeenCalled();
   });
 
+  it("drops discovered jobs when employer matches blocked company keywords", async () => {
+    const settingsRepo = await import("../../repositories/settings");
+    const jobSpy = await import("../../services/jobspy");
+
+    vi.mocked(settingsRepo.getAllSettings).mockResolvedValue({
+      searchTerms: JSON.stringify(["engineer"]),
+      blockedCompanyKeywords: JSON.stringify(["recruit", "staffing"]),
+    } as any);
+
+    vi.mocked(jobSpy.runJobSpy).mockResolvedValue({
+      success: true,
+      jobs: [
+        {
+          source: "linkedin",
+          title: "Engineer",
+          employer: "Acme Staffing",
+          jobUrl: "https://example.com/job-1",
+        },
+        {
+          source: "linkedin",
+          title: "Engineer II",
+          employer: "Contoso",
+          jobUrl: "https://example.com/job-2",
+        },
+      ],
+    } as any);
+
+    const result = await discoverJobsStep({
+      mergedConfig: {
+        ...config,
+        sources: ["linkedin"],
+      },
+    });
+
+    expect(result.discoveredJobs).toHaveLength(1);
+    expect(result.discoveredJobs[0]?.employer).toBe("Contoso");
+  });
+
   it("tracks source completion counters across source transitions", async () => {
     const settingsRepo = await import("../../repositories/settings");
     const jobSpy = await import("../../services/jobspy");

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

@@ -5,6 +5,7 @@ import {
   isSourceAllowedForCountry,
   normalizeCountryKey,
 } from "@shared/location-support.js";
+import { normalizeStringArray } from "@shared/normalize-string-array.js";
 import { parseSearchCitiesSetting } from "@shared/search-cities.js";
 import type { CreateJobInput, PipelineConfig } from "@shared/types";
 import * as jobsRepo from "../../repositories/jobs";
@@ -31,6 +32,31 @@ type DiscoverySourceTask = {
   run: () => Promise<DiscoveryTaskResult>;
 };
 
+function parseBlockedCompanyKeywords(raw: string | undefined): string[] {
+  if (!raw) return [];
+  try {
+    const parsed = JSON.parse(raw);
+    if (!Array.isArray(parsed)) return [];
+    return normalizeStringArray(
+      parsed.filter((value): value is string => typeof value === "string"),
+    );
+  } catch {
+    return [];
+  }
+}
+
+function isBlockedEmployer(
+  employer: string | null | undefined,
+  blockedKeywordsLowerCase: string[],
+): boolean {
+  if (!employer) return false;
+  if (blockedKeywordsLowerCase.length === 0) return false;
+  const normalizedEmployer = employer.toLowerCase();
+  return blockedKeywordsLowerCase.some((keyword) =>
+    normalizedEmployer.includes(keyword),
+  );
+}
+
 export async function discoverJobsStep(args: {
   mergedConfig: PipelineConfig;
   shouldCancel?: () => boolean;
@@ -525,11 +551,41 @@ export async function discoverJobsStep(args: {
     sourceErrors.push(...sourceResult.sourceErrors);
   }
 
-  if (args.shouldCancel?.()) {
-    return { discoveredJobs, sourceErrors };
+  const blockedCompanyKeywords = parseBlockedCompanyKeywords(
+    settings.blockedCompanyKeywords,
+  );
+  const blockedKeywordsLowerCase = blockedCompanyKeywords.map((value) =>
+    value.toLowerCase(),
+  );
+  const filteredDiscoveredJobs = discoveredJobs.filter(
+    (job) => !isBlockedEmployer(job.employer, blockedKeywordsLowerCase),
+  );
+  const droppedCount = discoveredJobs.length - filteredDiscoveredJobs.length;
+
+  if (droppedCount > 0) {
+    const blockedCompanyKeywordsPreview = blockedCompanyKeywords.slice(0, 10);
+    const blockedCompanyKeywordsTruncated =
+      blockedCompanyKeywordsPreview.length < blockedCompanyKeywords.length;
+
+    logger.info("Dropped discovered jobs matching blocked company keywords", {
+      step: "discover-jobs",
+      droppedCount,
+      blockedKeywordCount: blockedCompanyKeywords.length,
+      blockedCompanyKeywordsPreview,
+      blockedCompanyKeywordsTruncated,
+    });
+
+    logger.debug("Full blocked company keywords used for filtering", {
+      step: "discover-jobs",
+      blockedCompanyKeywords,
+    });
   }
 
-  if (discoveredJobs.length === 0 && sourceErrors.length > 0) {
+  if (args.shouldCancel?.()) {
+    return { discoveredJobs: filteredDiscoveredJobs, sourceErrors };
+  }
+
+  if (filteredDiscoveredJobs.length === 0 && sourceErrors.length > 0) {
     throw new Error(`All sources failed: ${sourceErrors.join("; ")}`);
   }
 
@@ -537,7 +593,7 @@ export async function discoverJobsStep(args: {
     logger.warn("Some discovery sources failed", { sourceErrors });
   }
 
-  progressHelpers.crawlingComplete(discoveredJobs.length);
+  progressHelpers.crawlingComplete(filteredDiscoveredJobs.length);
 
-  return { discoveredJobs, sourceErrors };
+  return { discoveredJobs: filteredDiscoveredJobs, sourceErrors };
 }

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>;