feat(api): Greenhouse, Lever, and Ashby public-board adapters wired into discovery (#261)

[Taleef7]

Closes #261 (Epic #245 — Feed Engine).

Summary of Changes

1. Board Fetchers & Normalizers (apps/api/src/lib/job-sources/boards.ts):

   • htmlToText(html): Two-pass HTML entity decoding and tag stripping via node-html-parser, correctly handling double-encoded HTML from Greenhouse (<p>).
   • Normalizers for Greenhouse, Lever, and Ashby producing canonical SourcedJob records with full JD descriptions, inferred workplace types, and parsed salaries (via parseSalaryFromText).
   • fetchBoardJobs(target): Strict token validation using /^[A-Za-z0-9._-]+$/ before reaching external URLs; dispatches to fixed HTTPS host endpoints with 15s timeout and User-Agent: JobOpsCopilot/1.0 (+board-watch).
   • fetchTargetCompanyBoards(targets): Iterates enabled target companies; per-target failures log a warning and continue without failing the overall sweep.

2. Source Normalization (apps/api/src/lib/job-sources/normalize.ts):

   • Exported clean, employmentLabel, and inferWorkplaceType for reuse across external board adapters.

3. Discovery Integration (apps/api/src/lib/discovery.ts & apps/api/src/routes/discovery.ts):

   • Added listTargetCompanies and fetchBoards to DiscoveryDeps.
   • Extracted shared job insertion pipeline (insertIfNew) that encapsulates deduplication against seen keys, deps.createJob, best-effort prerankAnalysis, and concurrent Postgres 23505 race handling.
   • Polled enabled target company boards after saved searches and pushed board postings through insertIfNew.
   • Updated DiscoveryResult.source to reflect contributing board types (e.g. adzuna+greenhouse).
   • Wired listTargetCompanies from @/data/target-company-store and fetchBoards: fetchTargetCompanyBoards into route defaults.

4. Testing & Verification:

   • Added unit test suite apps/api/src/lib/job-sources/boards.test.ts (6 tests covering entity decoding, normalizer field mappings, salary extraction, invalid token rejection, and fault tolerance).
   • Extended apps/api/src/lib/discovery.test.ts (11 tests verifying cross-deduplication between board jobs and search results, composite source reporting, and disabled target skips).
   • Total 309 API unit tests pass. Full npm run check (typecheck + ESLint + Next.js build + API build) clean with 0 errors.

[Taleef7]

@codex review

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 183cd410c9

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".