feat(ppc): commit ranker_gici_cluster_override selector mode#102
Open
rsasaki0109 wants to merge 1 commit into
Open
feat(ppc): commit ranker_gici_cluster_override selector mode#102rsasaki0109 wants to merge 1 commit into
rsasaki0109 wants to merge 1 commit into
Conversation
The Phase 43/71 production n/r2 selector mode `ranker_gici_cluster_override` was
referenced by run scripts but never committed to exp_ppc_ctrbpf_fgo.py (absent
from argparse choices -> invalid-choice error), so the 86.21% OFFICIAL run was
not reproducible from committed code. This implements it.
- experiments/ppc_gici_override.py: pure, numpy-only helper implementing the
documented rule -- when the base ranker picks a high-risk GICI variant
(xd_gici_{c4,oa,combo,z,hs}), re-pick within the xd_gici family toward a tight
cluster (rms_rank<=12, >=6 members within 0.5 m, <=0.8 m from the pick;
largest cluster wins, RMS tie-break). pf_bridge excluded.
- exp_ppc_ctrbpf_fgo.py: add the argparse choice and apply the override after
the base ranker pick.
- tests/test_ppc_gici_override.py: 8 synthetic cases for the rule.
- docs: mark the replay blocker resolved.
Caveat: exact score replay still needs the regenerated candidate pool
(libgnss_diag_phase10/19 + selector_ranker_predictions_v5_nlos.csv); committed
thresholds follow plan.md (12 / 6 / 0.8 / 0.5 m) where the original build was
under-documented.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
rsasaki0109
force-pushed
the
feat/ppc-ranker-gici-cluster-override
branch
2 times, most recently
from
9877c9f to
9cf5e8d
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The Phase 43/71 production n/r2 selector mode
ranker_gici_cluster_overridewas referenced by run scripts but never committed toexp_ppc_ctrbpf_fgo.py(absent from argparse choices → invalid-choice error), so the 86.21% OFFICIAL run was not reproducible from committed code. This implements it.Changes
experiments/ppc_gici_override.py(new, pure numpy) — documented rule: when the base ranker picks a high-risk GICI variant (xd_gici_{c4,oa,combo,z,hs}), re-pick within thexd_gicifamily toward a tight cluster (rms_rank≤12, ≥6 members within 0.5 m, ≤0.8 m from the pick; largest cluster wins, RMS tie-break).pf_bridgeexcluded.exp_ppc_ctrbpf_fgo.py— add the argparse choice; apply the override after the base ranker pick.tests/test_ppc_gici_override.py— 8 synthetic cases (8/8 pass).No GPL
gici-opencode is used —xd_gici_*are precomputed candidate-position labels; the clustering logic is original.Caveat: exact score replay still needs the regenerated candidate pool (
libgnss_diag_phase10/19+selector_ranker_predictions_v5_nlos.csv, not all checked in); committed thresholds follow plan.md (12 / 6 / 0.8 / 0.5 m) where the original build was under-documented.🤖 Generated with Claude Code