Compaction: a button, and automatically when the window fills

A long conversation eventually just stops working. Compaction summarises the
earlier turns and sends the summary in their place.

The messages are kept. They stay in the transcript behind a collapsed
divider and simply stop being part of the request, which is what makes the
button safe to press and automatic compaction safe to have at all: a summary
that came out badly is a bad turn, not a lost conversation.

Stored on the Chat, not as a synthetic Message. A synthetic row needs a
role -- `system` breaks the one-system-message rule the moment build_messages
emits it beside the harness, and user/assistant makes it a turn people can
edit, regenerate from and copy, indistinguishable from a real one in all
four places a bubble is rendered. Worse, "editing rewinds, it does not
branch" would silently delete it and leave no marker that compaction had
happened at all.

The summary goes out as a user turn and an assistant turn, not one. A
leading assistant breaks templates requiring the first non-system message to
be user; a lone leading user produces user, user whenever the kept history
starts on a user turn -- which it always does, because the cutoff lands on a
finished reply.

compacted_through_id is a plain id rather than a foreign key: migrations.py
compiles only the column type, so a REFERENCES clause would exist on a fresh
database and not on an upgraded one, and a constraint half the fleet has is
worse than none. cutoff_message validates it on every read instead, and a
rewind past the boundary clears it.

Compacting again summarises only the delta, with the previous summary
supplied to be subsumed. Re-summarising the whole chat each time grows
quadratically and eventually exceeds the window it is protecting.

Automatically at the top of _run, not in post_message: that route's contract
is to return immediately and leave the slow part to a resumable connection,
and it also means build_request is called once, after compaction, with no
second assembly path. The trigger is the last reply's recorded usage plus an
estimate of the new turn -- retrospective because true prompt_tokens are only
knowable after a response, plus the delta because otherwise fifty thousand
characters pasted into the composer overflow a window that read 90% last
turn. It never fires when the context length is unknown. It does fire on
estimated counts, which is safe here precisely because nothing is lost.

_maybe_compact never raises: a failure logs and sends the uncompacted
request. A `status` event says "Summarising earlier messages…" in the
meantime, because a silent multi-second pause before the first token is what
a hang looks like.

The wording is three fragments under Admin - Prompts. Clearing task.compact
turns compaction off entirely.

Also adds compaction.moment(): SQLite does not store the offset, so a row
loaded from disk is naive while one in the session's identity map keeps its
tzinfo, and comparing the two raises. Every comparison here is between
exactly those.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

src/lembas/web/templates/admin/general.html

@@ -41,6 +41,29 @@
     </div>
   </section>
 
+  <section class="card">
+    <h2 class="card__title">Compaction</h2>
+    <p class="card__lede">
+      A long conversation eventually fills the model's context. When it gets
+      close, the earlier turns are summarised and the summary is sent in their
+      place. The messages themselves are kept and stay readable in the
+      transcript — they simply stop being sent.
+    </p>
+    <div class="field">
+      <label class="field__label" for="compact-threshold">Compact at</label>
+      <input class="input" id="compact-threshold" name="compact_threshold" type="number"
+             min="0" max="99" value="{{ values.compact_threshold }}">
+      <p class="field__hint">
+        Percent of the model's context length. <code>0</code> turns automatic
+        compaction off; the button in each chat still works. Nothing happens for
+        a model whose context length is unset under
+        <a href="/admin/models">Models</a> — that is "unknown", not "small", and
+        this will not act on a number nobody supplied. The wording of the
+        summary is under <a href="/admin/prompts">Prompts</a>.
+      </p>
+    </div>
+  </section>
+
   <section class="card">
     <h2 class="card__title">
       Registration