fix(cognition): claimant-tracked speculative markers — the refcount fix was wrong (H-13)
71542e61 fixed only half of H-13 and asserted a premise that is false. This
replaces it, and records why, because the failure mode is the one this arc
exists to catch.
WHAT WAS WRONG. The refcount rested on "seeding and eviction walk the same
source list, so the counts balance." They do not. The call site's arms are
MUTUALLY EXCLUSIVE (pipeline.py §10b): a proposal is seeded when SPECULATIVE or
released when COHERENT, never both. So a release never balances a claim the
same proposal made, and a COHERENT proposal ALWAYS releases claims it never
held. Refcounting therefore repaired only the sub-case where both proposals had
been taught speculatively. The ordinary case — an unreviewed `wisdom`, then a
separate correction about `practical wisdom` that passes review on its own turn
— still stripped the marker. Measured against the landed refcount:
`sibling marked? False`.
The first attempt's tests passed because their helpers split tokens themselves,
mirroring a call pattern production no longer used — tests written to the fix
instead of to the defect. That is the same "asserted past the evidence" failure
diagnosed in two external assessments earlier today, committed here one commit
later, and catchable by reading fifteen lines at the call site.
WHAT IS CORRECT. _speculative_subjects maps token -> SET OF CLAIMING SUBJECTS.
A token is evicted exactly when the subject that actually claimed it is
reviewed; discarding an absent claimant is a no-op, so a neighbour's marker
survives. Token derivation moved out of the call site into
_speculative_index_tokens, because ownership is what the call site cannot see —
doing the split out there made every token its own claimant, which was the bug.
Tests rewritten to call the cache exactly as §10b does. The lead pin
(test_coherent_proposal_does_not_unmark_an_unrelated_subject) was confirmed to
FAIL against the refcount design before this landed, so it discriminates rather
than merely passing. 13 pins green; the 8 pre-existing lifecycle pins unchanged.
SUITE MEMBERSHIP, resolved rather than deferred. The file pinning this behavior
sat in NO curated suite, reachable only through `full` — which is why the
regression was invisible (G-7's mechanism, on the pin that would have caught
it). Registered now in BOTH TEST_SUITES["smoke"] and smoke.yml, per the #136
precedent that an unregistered pin runs nowhere. Verified the local/CI delta is
unchanged at exactly ten files (local 24, CI 14, CI-only 0), so this is
compatible with all three R-14 options and settles none of them; PR-4 still
owns the standing ten.
Cleanup as found: three unambiguously dead imports removed from pipeline.py
(hashlib, RatificationOutcome, RatifiedIntent). ruff 4 -> 1; the survivor is a
pre-existing structural E402, left alone.
Ratification note unchanged: this changes served surfaces — markers appear on
turns where they previously vanished. Direction is toward disclosure and it
restores ADR-0021 §Articulation's stated intent rather than extending it.
Co-Authored-By: Claude <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Wcw2pnMBwyvmNyQg4uPEt4This commit is contained in:
Claude
parent
71542e61ee
commit
2cfc6ad251
5 changed files with 187 additions and 120 deletions
3
.github/workflows/smoke.yml
vendored
3
.github/workflows/smoke.yml
vendored
|
|
@ -60,4 +60,5 @@ jobs:
|
|||
tests/test_architectural_invariants.py \
|
||||
tests/test_audio_*.py \
|
||||
tests/test_pack_measurements_phase2.py \
|
||||
tests/test_register_substantive_consumption.py
|
||||
tests/test_register_substantive_consumption.py \
|
||||
tests/test_speculative_subject_lifecycle.py
|
||||
|
|
|
|||
|
|
@ -79,6 +79,19 @@ TEST_SUITES: dict[str, tuple[str, ...]] = {
|
|||
# creates it, per the #136 finding that an unregistered pin runs
|
||||
# nowhere. ~6s, mostly two real pipeline turns per denial pair.
|
||||
"tests/test_negation_survives_articulation.py",
|
||||
# The speculative marker must survive a neighbour's review. Teaching
|
||||
# indexes a subject under its tokens, so one COHERENT correction used
|
||||
# to release a token an unrelated, still-unreviewed proposal was
|
||||
# relying on — serving unreviewed material with no "(speculative, not
|
||||
# yet reviewed)" prefix. That is a truth-on-the-served-surface defect,
|
||||
# so it belongs on the pre-push gate rather than under `full`. The file
|
||||
# predates the fix but ran in NO curated suite, which is why the
|
||||
# regression was invisible; it is registered here **and in smoke.yml**
|
||||
# in the same PR as the fix, per the #136 finding that an unregistered
|
||||
# pin runs nowhere. Registering it in both keeps the local/CI delta at
|
||||
# the same ten files (H-12), so it settles nothing R-14 has to rule.
|
||||
# ~9s, cache-level with two served-surface assertions.
|
||||
"tests/test_speculative_subject_lifecycle.py",
|
||||
# ADR governance. An ADR that governs a default-ON flag records a
|
||||
# decision already in force, so leaving it "Proposed" makes the
|
||||
# governance record assert something false about the running system —
|
||||
|
|
|
|||
|
|
@ -15,7 +15,6 @@ Constraint: ChatRuntime.chat() and ChatResponse contract are unchanged.
|
|||
|
||||
from __future__ import annotations
|
||||
|
||||
import hashlib
|
||||
import json
|
||||
from collections import OrderedDict
|
||||
|
||||
|
|
@ -35,11 +34,7 @@ from core.physics.wave_manifold import multivector_content_digest
|
|||
from core.reasoning.adapters import evidence_from_entailment_trace
|
||||
from generate.intent import classify_compound_intent
|
||||
from generate.intent_bridge import _is_useful_surface
|
||||
from generate.intent_ratifier import (
|
||||
RatificationOutcome,
|
||||
RatifiedIntent,
|
||||
ratify_intent,
|
||||
)
|
||||
from generate.intent_ratifier import ratify_intent
|
||||
from generate.graph_planner import (
|
||||
GraphNode,
|
||||
PropositionGraph,
|
||||
|
|
@ -159,19 +154,27 @@ class CognitiveTurnPipeline:
|
|||
# Iteration order matches insertion / refresh order; lookups
|
||||
# remain O(1).
|
||||
#
|
||||
# H-13 (external-assessment triage, 2026-07-28) — the value is a
|
||||
# REFERENCE COUNT, not a placeholder. Seeding splits a subject
|
||||
# into tokens, so two independent proposals can claim the same
|
||||
# token ("wisdom" from both ``wisdom`` and ``practical wisdom``).
|
||||
# While this was a flat set, promoting either one popped the
|
||||
# shared token and the *other* proposal — still unreviewed — lost
|
||||
# its marker, serving speculative material as though reviewed.
|
||||
# A token now survives while any unpromoted proposal still claims
|
||||
# it. Seeding and eviction walk the same source list, so the
|
||||
# counts balance; where they disagree the count stays positive
|
||||
# and the subject keeps its marker, which is the honest direction
|
||||
# to fail.
|
||||
self._speculative_subjects: OrderedDict[str, int] = OrderedDict()
|
||||
# H-13 (external-assessment triage, 2026-07-28) — the value is the
|
||||
# set of SUBJECTS CLAIMING this token, not a placeholder and not a
|
||||
# count. Seeding indexes a subject under itself *and* under each
|
||||
# of its ≥4-char tokens, so two independent subjects can land on
|
||||
# one token ("wisdom" from both ``wisdom`` and ``practical
|
||||
# wisdom``). While this was a flat set, teaching ``practical
|
||||
# wisdom`` coherently popped the shared token and an unrelated,
|
||||
# still-unreviewed ``wisdom`` proposal lost its marker — serving
|
||||
# speculative material as though reviewed.
|
||||
#
|
||||
# A reference count does not fix it, and the reason is the branch
|
||||
# structure at the call site: a proposal is EITHER seeded (when
|
||||
# SPECULATIVE) OR released (when COHERENT), never both, so a
|
||||
# release never balances a claim the same proposal made. A
|
||||
# COHERENT proposal always releases claims it never held.
|
||||
# Claimants make that a no-op on other subjects' tokens: a token
|
||||
# dies only when the subject that actually claimed it is
|
||||
# reviewed. Token derivation lives in these methods rather than
|
||||
# at the call site, because ownership is exactly what the call
|
||||
# site cannot see.
|
||||
self._speculative_subjects: OrderedDict[str, set[str]] = OrderedDict()
|
||||
|
||||
# ------------------------------------------------------------------
|
||||
# Public API
|
||||
|
|
@ -737,19 +740,19 @@ class CognitiveTurnPipeline:
|
|||
if proposal.epistemic_status is EpistemicStatus.SPECULATIVE:
|
||||
for src in sources:
|
||||
self._remember_speculative_subject(src)
|
||||
for tok in _SUBJECT_SPLIT_RE.split(src.lower()):
|
||||
if len(tok) >= 4 and tok not in _SUBJECT_STOPWORDS:
|
||||
self._remember_speculative_subject(tok)
|
||||
elif proposal.epistemic_status is EpistemicStatus.COHERENT:
|
||||
# Finding 5 (audit 2026-05-20) — once teaching review
|
||||
# promotes a proposal to COHERENT, the subject is no
|
||||
# longer speculative; evict its tokens so the marker
|
||||
# stops appearing on subsequent probes about it.
|
||||
# returns COHERENT, the subject is no longer speculative;
|
||||
# release its claims so the marker stops appearing on
|
||||
# subsequent probes about it.
|
||||
#
|
||||
# H-13 — the split-and-add loop that used to live here has
|
||||
# moved inside the cache. Doing it out here made every
|
||||
# token its own claimant, so releasing "practical wisdom"
|
||||
# also released the token "wisdom" that an unrelated,
|
||||
# still-unreviewed proposal was relying on.
|
||||
for src in sources:
|
||||
self._forget_speculative_subject(src)
|
||||
for tok in _SUBJECT_SPLIT_RE.split(src.lower()):
|
||||
if len(tok) >= 4 and tok not in _SUBJECT_STOPWORDS:
|
||||
self._forget_speculative_subject(tok)
|
||||
|
||||
# Advance turn counter and remember surface for next correction binding.
|
||||
# The truth-path surface (not the served, register-decorated bytes)
|
||||
|
|
@ -1039,53 +1042,76 @@ class CognitiveTurnPipeline:
|
|||
prompt_versor = field_state.F
|
||||
return ratify_intent(intent, prompt_versor, vocab=vocab)
|
||||
|
||||
@staticmethod
|
||||
def _speculative_index_tokens(subject: str) -> tuple[str, tuple[str, ...]]:
|
||||
"""Return ``(owner, tokens)`` — how one subject is indexed.
|
||||
|
||||
The owner is the normalized subject itself; the tokens are the
|
||||
owner plus each ≥4-char non-stopword fragment, so a prefixed
|
||||
parse like ``"correction: wisdom"`` still matches a probe about
|
||||
``"wisdom"``. Both the claim and the release derive their
|
||||
tokens here, so the two can never disagree about what a subject
|
||||
covers.
|
||||
"""
|
||||
owner = subject.lower().strip()
|
||||
if not owner:
|
||||
return "", ()
|
||||
tokens = [owner]
|
||||
for tok in _SUBJECT_SPLIT_RE.split(owner):
|
||||
if len(tok) >= 4 and tok not in _SUBJECT_STOPWORDS and tok != owner:
|
||||
tokens.append(tok)
|
||||
return owner, tuple(tokens)
|
||||
|
||||
def _remember_speculative_subject(self, subject: str) -> None:
|
||||
"""Claim a speculative subject token (and refresh its LRU position).
|
||||
"""Record ``subject`` as claiming its index tokens (refreshing LRU).
|
||||
|
||||
Finding 5 (audit 2026-05-20). Caps the cache at
|
||||
``_MAX_SPECULATIVE_SUBJECTS`` via insertion-order eviction.
|
||||
Empty / whitespace-only inputs are dropped silently so callers
|
||||
can pass raw fragments without guarding.
|
||||
|
||||
H-13 — increments the token's reference count rather than
|
||||
re-seating a placeholder, so a token claimed by two unreviewed
|
||||
proposals needs two promotions to clear.
|
||||
H-13 — the subject is recorded as the *claimant* of each token it
|
||||
indexes under, so a later release can tell its own tokens from a
|
||||
neighbour's.
|
||||
"""
|
||||
subject = subject.lower().strip()
|
||||
if not subject:
|
||||
owner, tokens = self._speculative_index_tokens(subject)
|
||||
if not owner:
|
||||
return
|
||||
count = self._speculative_subjects.get(subject, 0)
|
||||
self._speculative_subjects[subject] = count + 1
|
||||
self._speculative_subjects.move_to_end(subject)
|
||||
for token in tokens:
|
||||
claimants = self._speculative_subjects.get(token)
|
||||
if claimants is None:
|
||||
self._speculative_subjects[token] = {owner}
|
||||
else:
|
||||
claimants.add(owner)
|
||||
self._speculative_subjects.move_to_end(token)
|
||||
while len(self._speculative_subjects) > _MAX_SPECULATIVE_SUBJECTS:
|
||||
self._speculative_subjects.popitem(last=False)
|
||||
|
||||
def _forget_speculative_subject(self, subject: str) -> None:
|
||||
"""Release one claim on a subject; evict it once no claim remains.
|
||||
"""Release ``subject``'s claims; evict each token no one still claims.
|
||||
|
||||
Called when a SPECULATIVE proposal is promoted to COHERENT via
|
||||
the teaching review loop, so reviewed material stops being
|
||||
marked speculative on later probes. No-op if the subject is
|
||||
not present.
|
||||
Called when teaching review returns a COHERENT proposal, so
|
||||
material that has now been reviewed stops being marked
|
||||
speculative on later probes. No-op for subjects that hold no
|
||||
claim.
|
||||
|
||||
H-13 — decrements rather than pops. Token seeding is
|
||||
subject-split, so two proposals can claim one token; popping
|
||||
outright let a promotion strip the marker from a *different*
|
||||
proposal that was still unreviewed. An unmatched release (the
|
||||
promoted proposal's token set differing from what it seeded)
|
||||
floors at removal and never goes negative, so it cannot borrow
|
||||
against a later proposal's claim.
|
||||
H-13 — releases only what *this* subject claimed. A COHERENT
|
||||
proposal never seeded anything (the call site's branches are
|
||||
mutually exclusive), so it holds no claim on a neighbour's
|
||||
token: discarding an absent claimant leaves that token live, and
|
||||
the still-unreviewed subject keeps its marker. A token is
|
||||
evicted exactly when its last claimant is reviewed.
|
||||
"""
|
||||
subject = subject.lower().strip()
|
||||
if not subject:
|
||||
owner, tokens = self._speculative_index_tokens(subject)
|
||||
if not owner:
|
||||
return
|
||||
count = self._speculative_subjects.get(subject)
|
||||
if count is None:
|
||||
return
|
||||
if count <= 1:
|
||||
self._speculative_subjects.pop(subject, None)
|
||||
else:
|
||||
self._speculative_subjects[subject] = count - 1
|
||||
for token in tokens:
|
||||
claimants = self._speculative_subjects.get(token)
|
||||
if claimants is None:
|
||||
continue
|
||||
claimants.discard(owner)
|
||||
if not claimants:
|
||||
self._speculative_subjects.pop(token, None)
|
||||
|
||||
def _should_mark_speculative(self, text: str, surface: str) -> bool:
|
||||
"""Decide whether ``surface`` should carry the SPECULATIVE marker.
|
||||
|
|
|
|||
|
|
@ -105,8 +105,17 @@ Ranked by leverage (cognitive/structural load removed ÷ effort), per the AGENTS
|
|||
**Why it hinders:** the marker is the honesty surface for unreviewed material, and the failure direction is served-unhedged — the direction this architecture treats as the expensive one. *(Provenance: surfaced by an external assessment that described this seam with the direction inverted — it claimed under-eviction; the code over-evicts. The verification that caught the inversion is what confirmed the real bug.)*
|
||||
**Better home:** a reference-counted cache (token → count of live proposals; decrement on promotion, evict at zero), with the LRU cap retained as a size valve. **Failing test first:** promote one proposal, assert the sibling's marker survives — red before green.
|
||||
**Authority:** mechanical PR + red-then-green test; flag at review whether ADR cover is wanted, since served-marker behavior changes (in the honest direction).
|
||||
**Status (2026-07-28): FIXED.** `_speculative_subjects` is now `OrderedDict[str, int]` — a reference count; `_forget_speculative_subject` decrements and evicts at zero, flooring at removal so an unmatched release cannot borrow against a later proposal's claim. Four new pins in `tests/test_speculative_subject_lifecycle.py`, **all four observed red before green**; the eight pre-existing lifecycle pins pass unchanged. **Ratification note:** this changes served surfaces — markers now appear on turns where they previously vanished. The direction is toward disclosure, and it restores ADR-0021 §Articulation's stated intent rather than extending it, so it is offered as a defect fix; say if you want ADR cover before it ships.
|
||||
**Found while fixing it (G-7's mechanism, live):** `tests/test_speculative_subject_lifecycle.py` — the file that pins this exact behavior — is **in no curated suite tuple**, reachable only through `full`. The pin that would have caught this defect on a pre-merge gate does not run on one. Assignment left to PR-4, which is designed to place every orphan; adding it to `TEST_SUITES["smoke"]` alone would widen N-3's ten-file delta and prejudge R-14's direction, so it was deliberately not done here.
|
||||
**Status (2026-07-28): FIXED — on the second attempt, and the first attempt is the more useful record.**
|
||||
|
||||
The value of `_speculative_subjects` is now the **set of subjects claiming each token** (`OrderedDict[str, set[str]]`). A token is evicted exactly when the subject that actually claimed it is reviewed. Token derivation moved from the call site into `_speculative_index_tokens`, because ownership is precisely what the call site cannot see.
|
||||
|
||||
**The first fix was a reference count, and it did not work.** It rested on "seeding and eviction walk the same source list, so counts balance." They do not: the call site's arms are **mutually exclusive** — a proposal is seeded when SPECULATIVE *or* released when COHERENT, never both — so a release never balances a claim the same proposal made, and a COHERENT proposal always releases claims it never held. Refcounting therefore only repaired the sub-case where *both* proposals had been taught speculatively; the common case (an unreviewed `wisdom`, then a separate correction about `practical wisdom` that passes review on its own turn) still stripped the marker. Demonstrated after the refcount landed: `sibling marked? False`.
|
||||
|
||||
The first attempt's tests passed because their helpers split tokens themselves, testing a call pattern production no longer used — **a test written to the fix instead of to the defect**. The replacement helpers call the cache exactly as `pipeline.py` §10b does, and the lead pin (`test_coherent_proposal_does_not_unmark_an_unrelated_subject`) was confirmed to fail against the refcount design before the claimant design landed. 13 pins pass, 8 of them the pre-existing lifecycle pins, unchanged.
|
||||
|
||||
**Ratification note:** this changes served surfaces — markers now appear on turns where they previously vanished. The direction is toward disclosure and it restores ADR-0021 §Articulation's stated intent rather than extending it, so it is offered as a defect fix; say if you want ADR cover before it goes further.
|
||||
|
||||
**Suite membership resolved, not deferred (G-7's mechanism, live).** `tests/test_speculative_subject_lifecycle.py` pinned this exact behavior while sitting **in no curated suite** — reachable only through `full` — which is why the regression was invisible. It is now registered in **both** `TEST_SUITES["smoke"]` and `smoke.yml`, per the #136 precedent that an unregistered pin runs nowhere. Registering it in both leaves the local/CI delta at **exactly ten files** (verified: local 24, CI 14, CI-only 0), so it is compatible with all three R-14 options and settles none of them. PR-4 still owns the standing ten.
|
||||
|
||||
## H-14 · The cognitive spine's turn method is 759 lines, with back-stamp repairs as the symptom
|
||||
**Verdict:** `strained` — structure debt on the highest-traffic path; H-4's sibling · **Layers:** M3/M4
|
||||
|
|
|
|||
|
|
@ -107,63 +107,74 @@ def test_should_mark_speculative_still_iterates_correctly(
|
|||
# unpromoted proposal still claims it.
|
||||
|
||||
|
||||
def _seed_proposal(pipeline: CognitiveTurnPipeline, subject: str) -> None:
|
||||
"""Mirror the seeding the pipeline performs for one SPECULATIVE proposal."""
|
||||
# The pipeline calls the cache once per source string and lets the cache do its
|
||||
# own token indexing (H-13). These helpers mirror that exactly — a test that
|
||||
# split tokens itself would be testing a call pattern production no longer uses,
|
||||
# which is how the first attempt at this fix passed while the bug survived.
|
||||
|
||||
|
||||
def _teach_speculative(pipeline: CognitiveTurnPipeline, subject: str) -> None:
|
||||
"""One SPECULATIVE proposal lands (pipeline.py 10b, SPECULATIVE arm)."""
|
||||
pipeline._remember_speculative_subject(subject)
|
||||
for tok in subject.lower().split():
|
||||
if len(tok) >= 4:
|
||||
pipeline._remember_speculative_subject(tok)
|
||||
|
||||
|
||||
def _promote_proposal(pipeline: CognitiveTurnPipeline, subject: str) -> None:
|
||||
"""Mirror the eviction the pipeline performs when that proposal goes COHERENT."""
|
||||
def _teach_coherent(pipeline: CognitiveTurnPipeline, subject: str) -> None:
|
||||
"""One COHERENT proposal lands (pipeline.py 10b, COHERENT arm).
|
||||
|
||||
Note this arm is reached by proposals that were **never seeded** — review
|
||||
returns COHERENT on the same turn the correction is captured. So a release
|
||||
routinely names tokens this subject never claimed.
|
||||
"""
|
||||
pipeline._forget_speculative_subject(subject)
|
||||
for tok in subject.lower().split():
|
||||
if len(tok) >= 4:
|
||||
pipeline._forget_speculative_subject(tok)
|
||||
|
||||
|
||||
def test_promoting_one_proposal_keeps_a_sibling_token_live(
|
||||
def test_coherent_proposal_does_not_unmark_an_unrelated_subject(
|
||||
pipeline: CognitiveTurnPipeline,
|
||||
) -> None:
|
||||
"""Promoting 'practical wisdom' must not un-mark an unreviewed 'wisdom'."""
|
||||
_seed_proposal(pipeline, "wisdom")
|
||||
_seed_proposal(pipeline, "practical wisdom")
|
||||
"""The load-bearing case: the COHERENT proposal never seeded anything.
|
||||
|
||||
_promote_proposal(pipeline, "practical wisdom")
|
||||
An unreviewed ``wisdom`` is outstanding. A *different* correction about
|
||||
``practical wisdom`` passes review on its own turn — it was never
|
||||
SPECULATIVE, so it holds no claim on the token ``wisdom``. Releasing it
|
||||
must leave the unreviewed subject marked.
|
||||
"""
|
||||
_teach_speculative(pipeline, "wisdom")
|
||||
_teach_coherent(pipeline, "practical wisdom")
|
||||
|
||||
assert "wisdom" in pipeline._speculative_subjects, (
|
||||
"the independent 'wisdom' proposal is still SPECULATIVE — promoting a "
|
||||
"different proposal that merely shares the token must not evict it"
|
||||
"a COHERENT proposal released a token it never claimed, stripping the "
|
||||
"marker from an unrelated proposal that is still unreviewed"
|
||||
)
|
||||
# The promoted proposal's own exclusive tokens are gone.
|
||||
assert "practical wisdom" not in pipeline._speculative_subjects
|
||||
assert "practical" not in pipeline._speculative_subjects
|
||||
|
||||
|
||||
def test_sibling_survives_at_the_served_surface(
|
||||
pipeline: CognitiveTurnPipeline,
|
||||
) -> None:
|
||||
"""The consequence that matters: the marker still fires for the sibling."""
|
||||
_seed_proposal(pipeline, "wisdom")
|
||||
_seed_proposal(pipeline, "practical wisdom")
|
||||
_promote_proposal(pipeline, "practical wisdom")
|
||||
|
||||
assert pipeline._should_mark_speculative(
|
||||
text="Tell me about wisdom",
|
||||
surface="Wisdom is the application of knowledge.",
|
||||
), "unreviewed material must keep its speculative marker"
|
||||
|
||||
|
||||
def test_token_clears_once_every_claimant_is_promoted(
|
||||
def test_promoting_one_proposal_keeps_a_sibling_token_live(
|
||||
pipeline: CognitiveTurnPipeline,
|
||||
) -> None:
|
||||
"""Reference counting must still reach zero — no marker that never clears."""
|
||||
_seed_proposal(pipeline, "wisdom")
|
||||
_seed_proposal(pipeline, "practical wisdom")
|
||||
"""Both taught speculatively; reviewing one must not clear the other."""
|
||||
_teach_speculative(pipeline, "wisdom")
|
||||
_teach_speculative(pipeline, "practical wisdom")
|
||||
|
||||
_promote_proposal(pipeline, "practical wisdom")
|
||||
_promote_proposal(pipeline, "wisdom")
|
||||
_teach_coherent(pipeline, "practical wisdom")
|
||||
|
||||
assert "wisdom" in pipeline._speculative_subjects
|
||||
# The reviewed subject's own exclusive tokens are gone.
|
||||
assert "practical wisdom" not in pipeline._speculative_subjects
|
||||
assert "practical" not in pipeline._speculative_subjects
|
||||
|
||||
|
||||
def test_token_clears_once_its_own_claimant_is_reviewed(
|
||||
pipeline: CognitiveTurnPipeline,
|
||||
) -> None:
|
||||
"""Eviction must still happen — no marker that can never clear."""
|
||||
_teach_speculative(pipeline, "wisdom")
|
||||
_teach_speculative(pipeline, "practical wisdom")
|
||||
|
||||
_teach_coherent(pipeline, "practical wisdom")
|
||||
_teach_coherent(pipeline, "wisdom")
|
||||
|
||||
assert "wisdom" not in pipeline._speculative_subjects
|
||||
assert not pipeline._should_mark_speculative(
|
||||
|
|
@ -172,37 +183,44 @@ def test_token_clears_once_every_claimant_is_promoted(
|
|||
)
|
||||
|
||||
|
||||
def test_repeated_speculative_teaching_needs_matching_promotions(
|
||||
def test_claim_is_idempotent_per_subject(
|
||||
pipeline: CognitiveTurnPipeline,
|
||||
) -> None:
|
||||
"""Two unreviewed proposals on one subject take two promotions to clear."""
|
||||
_seed_proposal(pipeline, "wisdom")
|
||||
_seed_proposal(pipeline, "wisdom")
|
||||
"""Re-teaching one subject speculatively does not require two reviews.
|
||||
|
||||
_promote_proposal(pipeline, "wisdom")
|
||||
assert "wisdom" in pipeline._speculative_subjects, (
|
||||
"one proposal reviewed, one still outstanding — still speculative"
|
||||
)
|
||||
The claimant is the subject, so the same subject claiming twice is one
|
||||
claim — a single review clears it. (Distinct subjects sharing a token is
|
||||
the case that needs both reviewed, pinned above.)
|
||||
"""
|
||||
_teach_speculative(pipeline, "wisdom")
|
||||
_teach_speculative(pipeline, "wisdom")
|
||||
|
||||
_promote_proposal(pipeline, "wisdom")
|
||||
_teach_coherent(pipeline, "wisdom")
|
||||
assert "wisdom" not in pipeline._speculative_subjects
|
||||
|
||||
|
||||
def test_unmatched_forget_never_underflows(
|
||||
def test_release_of_an_unheld_subject_is_a_noop(
|
||||
pipeline: CognitiveTurnPipeline,
|
||||
) -> None:
|
||||
"""A promotion whose seeding never happened must not push a count negative.
|
||||
"""Releasing a subject nobody claimed must not disturb live claims."""
|
||||
_teach_speculative(pipeline, "wisdom")
|
||||
|
||||
Seeding and eviction derive their token sets from the proposal object at
|
||||
two different moments, so they can disagree. When they do, the safe
|
||||
direction is to keep marking (a stale marker is honest-conservative; a
|
||||
missing one is not) — never a negative count that would let a later
|
||||
promotion strip a live sibling.
|
||||
"""
|
||||
pipeline._remember_speculative_subject("wisdom")
|
||||
pipeline._forget_speculative_subject("wisdom")
|
||||
pipeline._forget_speculative_subject("wisdom") # unmatched
|
||||
pipeline._remember_speculative_subject("wisdom")
|
||||
_teach_coherent(pipeline, "entirely unrelated topic")
|
||||
_teach_coherent(pipeline, "practical") # shares no claim on "wisdom"
|
||||
|
||||
assert "wisdom" in pipeline._speculative_subjects
|
||||
assert pipeline._speculative_subjects["wisdom"] == 1
|
||||
assert pipeline._speculative_subjects["wisdom"] == {"wisdom"}
|
||||
|
||||
|
||||
def test_index_tokens_are_shared_by_claim_and_release(
|
||||
pipeline: CognitiveTurnPipeline,
|
||||
) -> None:
|
||||
"""Both sides derive tokens from one helper, so they cannot disagree."""
|
||||
owner, tokens = pipeline._speculative_index_tokens(" Correction: Practical WISDOM ")
|
||||
assert owner == "correction: practical wisdom"
|
||||
# "correction" is a subject stopword — indexing it would match every
|
||||
# correction turn, so it is filtered out of the token set.
|
||||
assert set(tokens) == {"correction: practical wisdom", "practical", "wisdom"}
|
||||
# The owner is always indexed, even when it is shorter than the token floor.
|
||||
owner_short, tokens_short = pipeline._speculative_index_tokens("ash")
|
||||
assert owner_short == "ash"
|
||||
assert tokens_short == ("ash",)
|
||||
|
|
|
|||
Loading…
Reference in a new issue