| 1 | """Durable, provider-free canonical sector discovery.""" |
| 2 | from __future__ import annotations |
| 3 | |
| 4 | from collections import defaultdict |
| 5 | from collections.abc import Iterable |
| 6 | |
| 7 | from app.sector_leaderboard import normalize_sector |
| 8 | |
| 9 | |
| 10 | def canonical_sector_counts( |
| 11 | instruments: Iterable[dict], |
| 12 | sector_by_instrument: dict[str, str | None], |
| 13 | ) -> list[dict[str, object]]: |
| 14 | """Count unique durable instruments by the backend's canonical sector label.""" |
| 15 | identities_by_sector: dict[str, set[str]] = defaultdict(set) |
| 16 | labels: dict[str, str] = {} |
| 17 | |
| 18 | for instrument in instruments: |
| 19 | identity = str(instrument.get("globalInstrumentId") or "").strip() |
| 20 | raw_sector = sector_by_instrument.get(identity) |
| 21 | if not identity or not raw_sector or not str(raw_sector).strip(): |
| 22 | continue |
| 23 | normalized, label = normalize_sector(str(raw_sector)) |
| 24 | identities_by_sector[normalized].add(identity) |
| 25 | labels.setdefault(normalized, label) |
| 26 | |
| 27 | return [ |
| 28 | {"name": labels[key], "instrumentCount": len(identities_by_sector[key])} |
| 29 | for key in sorted(identities_by_sector, key=lambda value: (labels[value].casefold(), labels[value])) |
| 30 | ] |