Python tool for aircraft Doppler audio geolocation using historical ADS-B flight data (Source-available)
Python acoustic-geolocationads-baircraft-trackingairplaneaudio-forensicsdoppler-effectdoppler-shiftgeolocationosintsignal-processing
Summarize drift-limited candidate groups
Cinna Davis3399cd497d6e
diff --git a/docs/PUBLIC_RECORDING_EXPERIMENT_2026-11-03.md b/docs/PUBLIC_RECORDING_EXPERIMENT_2026-11-03.mdindex e96b826..029716c 100644--- a/docs/PUBLIC_RECORDING_EXPERIMENT_2026-11-03.md+++ b/docs/PUBLIC_RECORDING_EXPERIMENT_2026-11-03.md@@ -67,6 +67,9 @@ coordinates back into a search.- **Rochester:** six proposal groups were discovered, but the leading locationhypothesis matches only group 6; five groups are unmatched. Its two reportedridge IDs are harmonics within that one group, not two independent flyovers.+ That group's shared fit also reaches the source-drift bound (-3.142 Hz with a+ 3.142 Hz limit), a warning hidden behind the higher-priority+ `search_incomplete` result.The finite-search assessment is incomplete, with 515 unknown clock trials and87 unclassified score candidates. One matched flyover cannot locate therecorder, so the `search_incomplete` status is appropriate.@@ -74,12 +77,15 @@ coordinates back into a search.groups 3 and 4 are unmatched. Groups 1 and 2 both map to the same archivedaircraft segment, so the three matched groups do not establish threeindependent flyovers. The source-frequency drift fit reaches its allowed bound- for both of those groups, which is why the result is- `source_drift_unresolved` even though the finite grid assessment completed.+ for both of those groups (-2.869/2.869 Hz and -2.783/2.783 Hz), which is why+ the result is `source_drift_unresolved` even though the finite grid assessment+ completed.- **East Texas:** six proposal groups were discovered, but the leadinghypothesis matches only groups 4 and 5; four are unmatched. The saved searchhas 999,703 unknown clock trials and 31,333 unclassified score candidates, so- the assessment is incomplete. Its `search_incomplete` status is appropriate.+ the assessment is incomplete. Group 5's source-drift fit also reaches its+ bound (-6.907 Hz with a 6.907 Hz limit); the `search_incomplete` status masks+ that additional warning.### Audit of the saved retention ledgersdiff --git a/src/dopplergeist/confidence.py b/src/dopplergeist/confidence.pyindex 356c337..7dc141f 100644--- a/src/dopplergeist/confidence.py+++ b/src/dopplergeist/confidence.py@@ -164,6 +164,7 @@ def _candidate_evidence_summary(events, candidate, evidence):matched_events = set(candidate["event_ids"])group_flights = {}flights_groups = {}+ drift_limited_groups = set()fit_summaries = []for event_id, flight_id in zip(candidate["event_ids"], candidate["flight_ids"]):event = lookup[event_id]@@ -172,8 +173,12 @@ def _candidate_evidence_summary(events, candidate, evidence):flights_groups.setdefault(flight_id, set()).add(group_id)fit = evidence.get(event_id)if isinstance(fit, dict):+ if fit.get("drift_at_bound") is True:+ drift_limited_groups.add(group_id)fit_summaries.append({"event_id": event_id,+ "harmonic_group_id": group_id,+ "flight_id": flight_id,**{name: fit[name] for name in("rmse_hz", "outlier_fraction", "support_fraction", "drift_at_bound")if name in fit},@@ -198,6 +203,8 @@ def _candidate_evidence_summary(events, candidate, evidence):{"group_id": group_id, "flight_ids": sorted(flight_ids)}for group_id, flight_ids in sorted(group_flights.items()) if len(flight_ids) > 1],+ "drift_limited_group_ids": sorted(drift_limited_groups),+ "drift_limited_group_count": len(drift_limited_groups),"event_fits": fit_summaries,}