LCORE-1572: Integrate conversation compaction into the query flow - #1796

Merged

tisnik merged 23 commits into

lightspeed-core:mainfrom

max-svistunov:lcore-1572-integrate-compaction-query-flow

Jun 9, 2026

Commits on Jun 8, 2026

LCORE-1572: add conversation compaction and wire it into /v1/query
max-svistunov
committed
532f23a
LCORE-1572: unit tests for conversation compaction core and /v1/query
max-svistunov
committed
bb59773
LCORE-1572: apply conversation compaction in the A2A endpoint
max-svistunov
committed
4b89235
LCORE-1572: apply conversation compaction in the streaming_query endpoint
max-svistunov
committed
f026f22
LCORE-1572: tests for the streaming compaction gate
max-svistunov
committed
edc0ba5
LCORE-1572: apply conversation compaction in the /v1/responses endpoint
max-svistunov
committed
9384e8e
LCORE-1572: tests for /v1/responses compacted-turn storage
max-svistunov
committed
694b2f0
LCORE-1572: update spec doc to the as-built compaction design
max-svistunov
committed
707ce1f
LCORE-1572: fix needs_compaction_path docstring (pydocstyle D400)
max-svistunov
committed
09020b9
LCORE-1572: build compacted input as typed messages (silence Pydantic warning)
max-svistunov
committed
5ba83b2
LCORE-1572: raise instead of assert on the drained compaction result
max-svistunov
committed
8b4f4ed
LCORE-1572: wire persisted recursive fold (R3) via the summary cache
max-svistunov
committed
ffd452a
LCORE-1572: address CodeRabbit review — list-form input tokens + clarity rename
max-svistunov
committed
ba8b29c
LCORE-1572: persist compacted streaming turns with structure (CodeRabbit #4)
max-svistunov
committed
f50f721
LCORE-1572: do not initialize the conversation cache when compaction is disabled
max-svistunov
committed
63cf5ac
LCORE-1572: address CodeRabbit round 2 (compacted-mode persistence edges)
max-svistunov
committed
8461c51
LCORE-1572: document the disable-after-compaction limitation in the spec doc (CodeRabbit E)
max-svistunov
committed
5448923
LCORE-1572: document as-built divergences in spec doc (cache source-of-truth, persisted fold)
max-svistunov
committed
461753d
LCORE-1572: fix line-too-long (C0301) in interrupted-turn test docstring
max-svistunov
committed
a7178a1
LCORE-1572: harden disabled-cache regression test to fail on eager cache access (CodeRabbit)
max-svistunov
committed
a87e90d
LCORE-1572: ref-count per-conversation lock + extract apply_compaction helpers (review)
max-svistunov
committed
9976097
LCORE-1572: tighten typed-item handling in compaction helpers (review)
max-svistunov
committed
82b491f
LCORE-1572: soften R12 doc on silent /v1/responses compaction (review)
max-svistunov
committed
15b4f30

Navigation Menu

Uh oh!

LCORE-1572: Integrate conversation compaction into the query flow - #1796

LCORE-1572: Integrate conversation compaction into the query flow#1796
tisnik merged 23 commits into
lightspeed-core:mainfrom
max-svistunov:lcore-1572-integrate-compaction-query-flow

Commits on Jun 8, 2026

LCORE-1572: add conversation compaction and wire it into /v1/query

LCORE-1572: unit tests for conversation compaction core and /v1/query

LCORE-1572: apply conversation compaction in the A2A endpoint

LCORE-1572: apply conversation compaction in the streaming_query endpoint

LCORE-1572: tests for the streaming compaction gate

LCORE-1572: apply conversation compaction in the /v1/responses endpoint

LCORE-1572: tests for /v1/responses compacted-turn storage

LCORE-1572: update spec doc to the as-built compaction design

LCORE-1572: fix needs_compaction_path docstring (pydocstyle D400)

LCORE-1572: build compacted input as typed messages (silence Pydantic warning)

LCORE-1572: raise instead of assert on the drained compaction result

LCORE-1572: wire persisted recursive fold (R3) via the summary cache

LCORE-1572: address CodeRabbit review — list-form input tokens + clarity rename

LCORE-1572: persist compacted streaming turns with structure (CodeRabbit #4)

LCORE-1572: do not initialize the conversation cache when compaction is disabled

LCORE-1572: address CodeRabbit round 2 (compacted-mode persistence edges)

LCORE-1572: document the disable-after-compaction limitation in the spec doc (CodeRabbit E)

LCORE-1572: document as-built divergences in spec doc (cache source-of-truth, persisted fold)

LCORE-1572: fix line-too-long (C0301) in interrupted-turn test docstring

LCORE-1572: harden disabled-cache regression test to fail on eager cache access (CodeRabbit)

LCORE-1572: ref-count per-conversation lock + extract apply_compaction helpers (review)

LCORE-1572: tighten typed-item handling in compaction helpers (review)

LCORE-1572: soften R12 doc on silent /v1/responses compaction (review)