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>
This commit is contained in:
Jaroslav Beneš
2026-08-01 01:02:02 +02:00
parent aa0bbe524a
commit 3b1632069c
15 changed files with 1009 additions and 9 deletions
+42
View File
@@ -167,6 +167,45 @@
.reasoning__summary::-webkit-details-marker { display: none; }
.reasoning__summary:hover { color: var(--ink); background: var(--surface-hover); }
/* --- Compacted turns -------------------------------------------------------
Summarised messages, kept and readable but out of the way. Collapsed by
default: the point of compacting was that they stopped mattering.
*/
.compacted {
border: 1px dashed var(--border-strong);
border-radius: var(--radius);
background: color-mix(in srgb, var(--surface) 50%, transparent);
}
.compacted__summary {
display: flex;
align-items: center;
gap: var(--sp-2);
padding: var(--sp-2) var(--sp-3);
cursor: pointer;
color: var(--ink-muted);
font-size: var(--text-sm);
list-style: none;
user-select: none;
}
.compacted__summary::-webkit-details-marker { display: none; }
.compacted__summary:hover { color: var(--ink); }
.compacted[open] .reasoning__chevron { transform: rotate(180deg); }
.compacted__note {
margin: 0;
padding: 0 var(--sp-3);
font-size: var(--text-xs);
color: var(--ink-faint);
line-height: var(--leading-normal);
}
.compacted__body {
display: flex;
flex-direction: column;
gap: var(--sp-6);
padding: var(--sp-4) var(--sp-3);
opacity: 0.75;
}
/* --- Suggestions -----------------------------------------------------------
Starting points on the empty screen. Cards rather than a list, because they
are things to press.
@@ -333,6 +372,9 @@
}
/* --- Stop, notes and editing ---------------------------------------------- */
.msg__status { font-size: var(--text-xs); color: var(--ink-faint); font-style: italic; }
.msg__status:empty { display: none; }
.msg__waiting {
display: flex;
align-items: center;