feat(retrieval): a tuned number carries the space it was measured in (#4104)

Milestone 416 step 6. A retrieval floor is a cosine similarity, which only
means something inside one embedding model's geometry over documents cut one
particular way. Change either and every floor on the install keeps applying
while describing nothing — and nothing anywhere says so, because the scores
simply come out different and the bar goes on cutting.

`CHUNKER_VERSION` already solved this for documents: stamped per row, so the
backfill re-embeds precisely what is stale. The same idea, applied to the
numbers:

- `calibration_stamp()` — embedding model + document shape, one definition.
  TWO fields, never a fused string (rule 149): a mismatch has to say WHICH half
  moved, because they call for different responses.
- `retrieval_tuning_events` gains `embedding_model` / `shape_version`
  (migration 0104), stamped on every write. Nullable and NOT backfilled —
  "unstamped" is the honest answer for a row written before this existed, and
  it reports as `stale: null`, never as fine.
- `current_settings` reports calibration per dial: tuned rows from their event,
  untouched dials from the registry default's own stamp.
- `retrieval_surfaces` and the Settings panel show the mismatch. The panel
  renders ONLY when something is stale, so seeing it at all is the signal.
- `migrate_floor` / `migrate_retrieval_floor` answers "a path for thresholds to
  be inherited by the next model so that they don't have to recalibrate a lot":
  the raw cosine cannot cross models, but the PERCENTILE it represented can.
  Measure what fraction of a surface's logged calls the old floor admitted,
  re-score those queries under the current model, take the value admitting the
  same fraction. Dry run by default; applying writes an ordinary tuning event
  with the arithmetic in its reason.

Nothing auto-retunes. A stale stamp says a number is no longer a measurement;
it does not say what the number should be, and #4102 measured the one case
where the statistic and the correct action pointed opposite ways.

The load-bearing test is an ABSENCE: no chat-model identifier may appear
anywhere in the calibration path. Claude produces none of these scores, so a
Claude upgrade must trigger nothing — a false alarm here teaches the operator
to ignore the real one on the day bge-small becomes bge-base.

Backup v17 carries both columns, unfilled on the way out and on the way back:
a round trip must not turn "we don't know" into a stated fact.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01821k5B3Ysecp9fNYs92Kuy

src/scribe/models/retrieval_tuning.py

@@ -75,6 +75,23 @@ class RetrievalTuningEvent(Base):
     # becomes a formality; the service refuses a blank one.
     reason: Mapped[str] = mapped_column(Text, nullable=False)
 
+    # WHAT SPACE THIS NUMBER WAS MEASURED IN (#4104). A floor is a cosine
+    # distance in one embedding model's geometry, over documents cut one
+    # particular way — change either and the number describes something that
+    # no longer exists.
+    #
+    # On the EVENT rather than beside the setting, because this table already
+    # holds one row per change and `current_settings` already reads the latest
+    # per dial. A parallel stamp row next to the value would be a second place
+    # to keep in sync, and the two disagreeing is worse than neither.
+    #
+    # Both NULLABLE for the rows written before this step: those were measured
+    # under something, but claiming to know which would be inventing a fact.
+    # Null here means "unstamped", which is a different and honest answer from
+    # naming a model that may be wrong.
+    embedding_model: Mapped[str | None] = mapped_column(Text, nullable=True)
+    shape_version: Mapped[int | None] = mapped_column(Integer, nullable=True)
+
     __table_args__ = (
         # The only read this table has: one surface's history, newest first.
         Index(
@@ -94,4 +111,8 @@ class RetrievalTuningEvent(Base):
             "new_value": self.new_value,
             "actor": self.actor,
             "reason": self.reason,
+            # Null for every row written before #4104, and rendered as
+            # "unstamped" rather than guessed at — see the column comments.
+            "embedding_model": self.embedding_model,
+            "shape_version": self.shape_version,
         }