Add GSC opportunity scorer and wire into analytics/onboarding flows

[AJaySi]

Motivation

• Provide a reusable opportunity-scoring module that identifies high-impact search opportunities from normalized GSC rows so UI/agents can consume a single stable schema.
• Enable window-over-window comparisons by returning a previous_period payload from the GSC service so rising/declining trends can be detected.
• Integrate scoring into both the realtime analytics handler and the onboarding data flow so both surface the same ranked opportunities.
• Skills: no AGENTS skill was needed for this change (helpers and integrations implemented directly).

Description

• Added backend/services/analytics/opportunity_scorer.py with pure functions: high_impression_low_ctr_queries, rising_queries, declining_pages, score_and_rank_opportunities, and categorize_opportunities, returning the stable opportunity schema (id, query, page_url, reason, score, current_metrics, previous_metrics).
• Integrated scorer in GSCAnalyticsHandler by normalizing current/previous query & page rows, calling categorize_opportunities(...), and returning opportunities alongside existing metrics in handler responses.
• Extended GSCService.get_search_analytics to compute a matching previous-date window and optionally fetch previous-period query and page rows (added previous_period in the analytics payload) to enable window-over-window deltas.
• Hooked opportunity extraction into onboarding integration (backend/api/content_planning/.../data_integration.py) so _get_gsc_analytics includes opportunities in its output and the onboarding flow reuses the same stable schema.

Testing

• Compiled modified modules with python -m py_compile backend/services/analytics/opportunity_scorer.py backend/services/analytics/handlers/gsc_handler.py backend/services/gsc_service.py backend/api/content_planning/services/content_strategy/onboarding/data_integration.py; compilation succeeded.
• Ran a focused smoke test invoking the scorer (from services.analytics.opportunity_scorer import categorize_opportunities) with representative query/page rows to validate categorization and ranking; the function produced a ranked list matching the stable schema.

---

Codex Task