feat: add scheduled crawls (closes #14)

[barunTayenjam]

Closes #14.

Projects can now have a crawl schedule (hourly, daily or weekly), selected when the project is created or edited. While the server is running, a scheduler ticks every minute and starts a new crawl for every scheduled project whose next run time has passed.

Changes

• migration 0076: projects.schedule_interval + projects.next_run
• models/repository: schedule fields, FindScheduledProjects, UpdateProjectNextRun
• SchedulerService (stdlib time.Ticker only, injected clock for tests); next_run always advances (even on error) so a failing project never blocks the schedule; BasicAuth projects are skipped as their credentials are only provided interactively and are not stored
• SaveProject sets next_run; UpdateProject recomputes it only when the interval changes
• [scheduler] enabled config (SEONAUT_SCHEDULER_ENABLED), default on
• project add/edit forms: schedule select + validation error
• translations: schedule strings in en, es, nl, uk, fa
• docs: README + INSTALL.md (scheduled crawls section)

Design decisions

• preset intervals (hourly/daily/weekly), not cron expressions: no new dependency, matches the "given time intervals" ask in Scheduling crawling job #14
• no catch-up storm on restart: a project whose time passed while down runs at most once on the next tick, then resumes its interval
• notification/report delivery is deliberately out of scope here; tracked in Notification feature. #15

Verification

• go test -race -parallel 1 -short ./internal/... pass, go vet clean, go build -race clean, gofmt clean on touched files
• new tests: TestRunOnce* (due/not-due/bogus-interval, error advances, BasicAuth skip), TestSchedule (interval validation + next_run), config scheduler assertion
• end-to-end against MySQL 8.4 + local test site: manual crawl produces report/CSV; forcing next_run past due triggers a scheduled crawl within one tick and advances next_run +1h; invalid interval rejected with form error; edit form preselects interval; restart causes no extra crawls

[barunTayenjam]

CI note (for maintainers): both PR workflows are gated on first-time contributor approval (action_required). I verified the equivalent steps locally on this branch:

• go vet ./internal/... — clean
• go test -race -parallel 1 -short ./internal/... — all packages pass (incl. new TestRunOnce*, TestSchedule)
• backend build replicated (CGO_ENABLED=0 go build ./cmd/server/main.go) — OK
• frontend step replicated in node:18-alpine3.18 (same esbuild command as Dockerfile) — OK
• manual e2e (MySQL 8.4 + local test site): scheduled crawl fires within one tick when next_run is past due, next_run advances, BasicAuth projects skipped, restart causes no catch-up crawls.

Happy to address any review feedback or rebase as needed.

[barunTayenjam]

Closing for now while I finish local verification. Will reopen once tested. Sorry for the noise.