Files
picoclaw/pkg/seahorse/schema.go
T
Liu Yuan 15a70ac45c feat(seahorse): implement short-term memory engine (LCM) (#2285)
* feat(seahorse): implement short-term memory engine of seahorse

Add pkg/seahorse/ module implementing a SQLite-backed DAG-based summary
hierarchy for context management, ported from lossless-claw's LCM design:

- types.go + short_constants.go: core types (Message, Summary, Conversation,
  ContextItem) and configuration constants (fanout, token targets, thresholds)
- migration.go: idempotent DB schema with FTS5 trigram tokenizer for CJK
- store.go: full SQLite CRUD (conversations, messages, summaries DAG,
  context_items with ordinal gap numbering, FTS5 search)
- short_engine.go: Engine lifecycle (NewEngine, Ingest, Assemble, Compact),
  session pattern filtering (ignore/stateless glob→regex compilation),
  per-session mutex via sync.Map
- short_assembler.go: budget-aware context assembly with fresh tail protection
  (32 messages), oldest-first eviction, summary XML formatting, RebuildContextItems
- short_compaction.go: leaf compaction (messages→summary) and condensed
  compaction (summaries→higher-level summary), 3-level LLM escalation,
  CompactUntilUnder for emergency overflow
- short_retrieval.go: lookupByID, FTS5/LIKE search, recursive expand with
  token cap
- context_seahorse.go: agent.ContextManager adapter, registered as "seahorse",
  provider↔seahorse message type conversion (ToolCalls, tool_result)

* fix(seahorse): correct 3 adapter bugs in context management

- TokenCount: use full message (Content+ToolCalls+Media) instead of Content-only
- Empty Content: rebuild Content from tool_result Parts when stored empty
- Duplicate summaries: summaries only in Summary field, not in History messages
- Grep: fix SearchResult.Snippet→Content for summaries
- Schema: fix FTS5 SQL uses VIRTUAL TABLE not TEMP TABLE
- TestFTS5SQLConstants: verify FTS5 SQL syntax correctness
- Test: fix flaky TestCompactLeaf

* fix(agent): ingest steering messages into seahorse SQLite

Steering messages were only persisted to session JSONL but not ingested
into seahorse SQLite, causing them to be missing from context assembly.

Added `ts.ingestMessage(turnCtx, al, pm)` call in the steering message
injection block alongside the existing JSONL persistence.

Test: TestSeahorseSteeringMessageIngested verifies steering messages
appear in seahorse SQLite DB after being processed.

* fix(seahorse): address 3 blocking bugs from code review

- Fix resequenceContextItemsTx scan error handling (store.go:850)
  Changed `return err` to `return scanErr` to properly propagate scan errors
  instead of returning nil (which silently corrupts data)

- Fix sql.NullString for INTEGER column (store.go:847)
  Changed `mid` from sql.NullString to sql.NullInt64 since message_id
  is INTEGER in schema. Removed unnecessary strconv.ParseInt call.

- Fix compactCondensed fallback deleting non-candidate items
  Added ReplaceContextItemsWithSummary method for per-item deletion
  when candidates are not contiguous in ordinal space.
  Optimized to use range deletion when candidates are consecutive.

* fix(seahorse): pass Budget to Compact for correct condensed threshold

Issue #4 from PR review: When Budget was not passed to seahorse.Compact,
it defaulted to `tokensBefore * 0.75`, making `tokensBefore > budget`
always true and causing condensed compaction to trigger unnecessarily.

Changes:
- context_seahorse.go: Forward Budget from CompactRequest to CompactInput
- loop.go: Pass Budget (ContextWindow) in all 3 Compact calls
- Add test verifying condensed is skipped when tokens < threshold
- Fix lint issues in store.go and store_test.go

* fix(seahorse): add mutex for assembler lazy initialization

Issue #5 from PR review: The check-then-create pattern for e.assembler
was a data race when multiple goroutines called Assemble() concurrently:
    if e.assembler == nil {
        e.assembler = &Assembler{...}
    }

Changes:
- Add assemblerMu sync.Mutex to Engine struct
- Add initAssemblerOnce() using double-checked locking (same pattern as initCompactionOnce)
- Add TestAssemblerLazyInitRace to verify thread-safety

* fix(seahorse): handle non-consecutive depths in selectShallowestCondensationCandidate

Issue #8 from PR review: the loop iterated depth 0, 1, 2... assuming
consecutive keys, but break when key was missing caused deeper depths
to never be checked.

Fix: collect all existing depth keys, sort, then iterate in order.

* fix(seahorse): wrap DeleteMessagesAfterID and appendContextItems in transactions

- DeleteMessagesAfterID: wrap all DELETE operations in a transaction for
  atomicity, remove redundant manual FTS delete (handled by trigger)
- appendContextItems: use transaction to fix read-then-write race condition
- Add GetMaxOrdinalTx and resolveItemTokenCountTx for transaction-scoped queries
- Remove unused resolveItemTokenCount function

Fixes PR review issues 6 and 7.

* fix(seahorse): derive readable content from Parts and cap CompactUntilUnder iterations

- Derive readable content from MessageParts in AddMessageWithParts so
  FTS5 indexing and summary formatting can access tool call information
- formatMessagesForSummary and truncateSummary now fall back to Parts
  when Content is empty, fixing blank summaries for Part-based messages
- Add MaxCompactIterations (20) to prevent CompactUntilUnder infinite
  loops; exceeded iterations are logged as warnings
2026-04-05 09:05:16 +08:00

186 lines
7.0 KiB
Go

package seahorse
import (
"database/sql"
"fmt"
"github.com/sipeed/picoclaw/pkg/logger"
)
// SQL statements for FTS5 tables with trigram tokenizer.
const (
sqlCreateSummariesFTS = `CREATE VIRTUAL TABLE IF NOT EXISTS summaries_fts USING fts5(
summary_id,
content,
tokenize="trigram"
)`
sqlCreateMessagesFTS = `CREATE VIRTUAL TABLE IF NOT EXISTS messages_fts USING fts5(
message_id,
content,
tokenize="trigram"
)`
sqlCheckFTS5Available = `CREATE VIRTUAL TABLE IF NOT EXISTS _fts5_check USING fts5(content)`
sqlCheckTrigramAvailable = `CREATE VIRTUAL TABLE IF NOT EXISTS _trigram_check USING fts5(content, tokenize="trigram")`
sqlDropFTS5Check = `DROP TABLE IF EXISTS _fts5_check`
sqlDropTrigramCheck = `DROP TABLE IF EXISTS _trigram_check`
)
// runSchema creates or upgrades the database schema.
// All schemas are idempotent (safe to run multiple times).
func runSchema(db *sql.DB) error {
// Check FTS5 support before creating tables
if err := checkFTS5Support(db); err != nil {
return fmt.Errorf("FTS5 check: %w", err)
}
stmts := []string{
`CREATE TABLE IF NOT EXISTS conversations (
conversation_id INTEGER PRIMARY KEY AUTOINCREMENT,
session_key TEXT NOT NULL UNIQUE,
created_at TEXT NOT NULL DEFAULT (datetime('now')),
updated_at TEXT NOT NULL DEFAULT (datetime('now'))
)`,
`CREATE TABLE IF NOT EXISTS messages (
message_id INTEGER PRIMARY KEY AUTOINCREMENT,
conversation_id INTEGER NOT NULL REFERENCES conversations(conversation_id),
role TEXT NOT NULL,
content TEXT NOT NULL DEFAULT '',
token_count INTEGER NOT NULL DEFAULT 0,
created_at TEXT NOT NULL DEFAULT (datetime('now'))
)`,
`CREATE TABLE IF NOT EXISTS message_parts (
part_id INTEGER PRIMARY KEY AUTOINCREMENT,
message_id INTEGER NOT NULL REFERENCES messages(message_id),
type TEXT NOT NULL,
text TEXT,
name TEXT,
arguments TEXT,
tool_call_id TEXT,
media_uri TEXT,
mime_type TEXT,
ordinal INTEGER NOT NULL DEFAULT 0
)`,
`CREATE TABLE IF NOT EXISTS summaries (
summary_id TEXT PRIMARY KEY,
conversation_id INTEGER NOT NULL REFERENCES conversations(conversation_id),
kind TEXT NOT NULL,
depth INTEGER NOT NULL DEFAULT 0,
content TEXT NOT NULL,
token_count INTEGER NOT NULL DEFAULT 0,
earliest_at TEXT,
latest_at TEXT,
descendant_count INTEGER NOT NULL DEFAULT 0,
descendant_token_count INTEGER NOT NULL DEFAULT 0,
source_message_token_count INTEGER NOT NULL DEFAULT 0,
model TEXT,
created_at TEXT NOT NULL DEFAULT (datetime('now'))
)`,
`CREATE TABLE IF NOT EXISTS summary_parents (
summary_id TEXT NOT NULL,
parent_summary_id TEXT NOT NULL,
PRIMARY KEY (summary_id, parent_summary_id)
)`,
`CREATE TABLE IF NOT EXISTS summary_messages (
summary_id TEXT NOT NULL,
message_id INTEGER NOT NULL,
ordinal INTEGER NOT NULL DEFAULT 0,
PRIMARY KEY (summary_id, message_id)
)`,
`CREATE TABLE IF NOT EXISTS context_items (
conversation_id INTEGER NOT NULL,
ordinal INTEGER NOT NULL,
item_type TEXT NOT NULL,
summary_id TEXT,
message_id INTEGER,
token_count INTEGER NOT NULL DEFAULT 0,
created_at TEXT NOT NULL DEFAULT (datetime('now')),
PRIMARY KEY (conversation_id, ordinal)
)`,
// FTS5 virtual table with trigram tokenizer for CJK support
sqlCreateSummariesFTS,
// FTS5 virtual table for message search with trigram tokenizer
sqlCreateMessagesFTS,
// Indexes for common query patterns
`CREATE INDEX IF NOT EXISTS idx_messages_conversation ON messages(conversation_id)`,
`CREATE INDEX IF NOT EXISTS idx_messages_created ON messages(conversation_id, created_at)`,
`CREATE INDEX IF NOT EXISTS idx_summaries_conversation ON summaries(conversation_id)`,
`CREATE INDEX IF NOT EXISTS idx_summaries_kind_depth ON summaries(conversation_id, kind, depth)`,
`CREATE INDEX IF NOT EXISTS idx_summary_parents_parent ON summary_parents(parent_summary_id)`,
`CREATE INDEX IF NOT EXISTS idx_summary_messages_message ON summary_messages(message_id)`,
`CREATE INDEX IF NOT EXISTS idx_context_items_conv ON context_items(conversation_id, ordinal)`,
// FTS5 triggers to keep summaries_fts in sync with summaries table
`CREATE TRIGGER IF NOT EXISTS summaries_ai AFTER INSERT ON summaries BEGIN
INSERT INTO summaries_fts (summary_id, content) VALUES (new.summary_id, new.content);
END`,
`CREATE TRIGGER IF NOT EXISTS summaries_ad AFTER DELETE ON summaries BEGIN
INSERT INTO summaries_fts (summaries_fts, summary_id, content) VALUES ('delete', old.summary_id, old.content);
END`,
`CREATE TRIGGER IF NOT EXISTS summaries_au AFTER UPDATE ON summaries BEGIN
INSERT INTO summaries_fts (summaries_fts, summary_id, content) VALUES ('delete', old.summary_id, old.content);
INSERT INTO summaries_fts (summary_id, content) VALUES (new.summary_id, new.content);
END`,
// FTS5 triggers to keep messages_fts in sync with messages table
`CREATE TRIGGER IF NOT EXISTS messages_ai AFTER INSERT ON messages BEGIN
INSERT INTO messages_fts (message_id, content) VALUES (new.message_id, new.content);
END`,
`CREATE TRIGGER IF NOT EXISTS messages_ad AFTER DELETE ON messages BEGIN
DELETE FROM messages_fts WHERE message_id = old.message_id;
END`,
`CREATE TRIGGER IF NOT EXISTS messages_au AFTER UPDATE ON messages BEGIN
DELETE FROM messages_fts WHERE message_id = old.message_id;
INSERT INTO messages_fts (message_id, content) VALUES (new.message_id, new.content);
END`,
}
for _, s := range stmts {
if _, err := db.Exec(s); err != nil {
return err
}
}
return nil
}
// checkFTS5Support verifies that SQLite has FTS5 with trigram tokenizer enabled.
// This is required for full-text search with CJK (Chinese, Japanese, Korean) support.
func checkFTS5Support(db *sql.DB) error {
// Check if FTS5 is compiled in
var fts5Enabled int
err := db.QueryRow(`SELECT sqlite_compileoption_used('ENABLE_FTS5')`).Scan(&fts5Enabled)
if err != nil {
// sqlite_compileoption_used might not exist in older SQLite
// Try a different approach: create a test FTS5 table
_, testErr := db.Exec(sqlCheckFTS5Available)
if testErr != nil {
return fmt.Errorf("SQLite FTS5 not available: %w (required for full-text search)", testErr)
}
db.Exec(sqlDropFTS5Check)
} else if fts5Enabled == 0 {
return fmt.Errorf("SQLite was compiled without FTS5 support (required for full-text search)")
}
// Check if trigram tokenizer is available by trying to create a test table
// Not all SQLite builds include the trigram tokenizer
_, err = db.Exec(sqlCheckTrigramAvailable)
if err != nil {
logger.WarnCF("seahorse", "SQLite trigram tokenizer not available, CJK search may be limited",
map[string]any{"error": err.Error()})
// Trigram is not strictly required, just better for CJK
// Don't return error, just log warning
} else {
db.Exec(sqlDropTrigramCheck)
}
return nil
}