@martian-engineering/lossless-claw 0.4.0 → 0.5.0

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@martian-engineering/lossless-claw",
-  "version": "0.4.0",
+  "version": "0.5.0",
   "description": "Lossless Context Management plugin for OpenClaw — DAG-based conversation summarization with incremental compaction",
   "type": "module",
   "main": "index.ts",
@@ -35,6 +35,7 @@
     "@sinclair/typebox": "0.34.48"
   },
   "devDependencies": {
+    "@changesets/changelog-github": "^0.6.0",
     "@changesets/cli": "^2.30.0",
     "typescript": "^5.7.0",
     "vitest": "^3.0.0"

package/src/compaction.ts

@@ -142,6 +142,14 @@ function generateSummaryId(content: string): string {
 /** Maximum characters for the deterministic fallback truncation (512 tokens * 4 chars). */
 const FALLBACK_MAX_CHARS = 512 * 4;
 const DEFAULT_LEAF_CHUNK_TOKENS = 20_000;
+
+/**
+ * Pattern matching MEDIA:/... file path references that appear in message content
+ * when the original message contained only a media attachment (image, file, etc.)
+ * with no meaningful text.
+ */
+const MEDIA_PATH_RE = /^MEDIA:\/.+$/;
+
 const CONDENSED_MIN_INPUT_RATIO = 0.1;
 
 function dedupeOrderedIds(ids: Iterable<string>): string[] {
@@ -231,6 +239,7 @@ export class CompactionEngine {
     summarize: CompactionSummarizeFn;
     force?: boolean;
     hardTrigger?: boolean;
+    summaryModel?: string;
   }): Promise<CompactionResult> {
     return this.compactFullSweep(input);
   }
@@ -246,6 +255,7 @@ export class CompactionEngine {
     summarize: CompactionSummarizeFn;
     force?: boolean;
     previousSummaryContent?: string;
+    summaryModel?: string;
   }): Promise<CompactionResult> {
     const { conversationId, tokenBudget, summarize, force } = input;
 
@@ -281,6 +291,7 @@ export class CompactionEngine {
       leafChunk.items,
       summarize,
       previousSummaryContent,
+      input.summaryModel,
     );
     if (!leafResult) {
       return {
@@ -322,6 +333,7 @@ export class CompactionEngine {
           chunk.items,
           targetDepth,
           summarize,
+          input.summaryModel,
         );
         if (!condenseResult) {
           break;
@@ -370,6 +382,7 @@ export class CompactionEngine {
     summarize: CompactionSummarizeFn;
     force?: boolean;
     hardTrigger?: boolean;
+    summaryModel?: string;
   }): Promise<CompactionResult> {
     const { conversationId, tokenBudget, summarize, force, hardTrigger } = input;
 
@@ -416,6 +429,7 @@ export class CompactionEngine {
         leafChunk.items,
         summarize,
         previousSummaryContent,
+        input.summaryModel,
       );
       if (!leafResult) {
         break;
@@ -461,6 +475,7 @@ export class CompactionEngine {
         candidate.chunk.items,
         candidate.targetDepth,
         summarize,
+        input.summaryModel,
       );
       if (!condenseResult) {
         break;
@@ -511,6 +526,7 @@ export class CompactionEngine {
     targetTokens?: number;
     currentTokens?: number;
     summarize: CompactionSummarizeFn;
+    summaryModel?: string;
   }): Promise<{ success: boolean; rounds: number; finalTokens: number }> {
     const { conversationId, tokenBudget, summarize } = input;
     const targetTokens =
@@ -542,6 +558,7 @@ export class CompactionEngine {
         tokenBudget,
         summarize,
         force: true,
+        summaryModel: input.summaryModel,
       });
 
       if (result.tokensAfter <= targetTokens) {
@@ -998,6 +1015,17 @@ export class CompactionEngine {
       };
     }
     const inputTokens = Math.max(1, estimateTokens(sourceText));
+    const buildDeterministicFallback = (): { content: string; level: CompactionLevel } => {
+      const truncated =
+        sourceText.length > FALLBACK_MAX_CHARS
+          ? sourceText.slice(0, FALLBACK_MAX_CHARS)
+          : sourceText;
+      return {
+        content: `${truncated}
+[Truncated from ${inputTokens} tokens]`,
+        level: "fallback",
+      };
+    };
 
     const runSummarizer = async (aggressiveMode: boolean): Promise<string | null> => {
       const output = await params.summarize(sourceText, aggressiveMode, params.options);
@@ -1007,7 +1035,9 @@ export class CompactionEngine {
 
     const initialSummary = await runSummarizer(false);
     if (initialSummary === null) {
-      return null;
+      // Empty provider output should still compact deterministically so auth
+      // failures or empty responses do not stall compaction entirely.
+      return buildDeterministicFallback();
     }
     let summaryText = initialSummary;
     let level: CompactionLevel = "normal";
@@ -1015,25 +1045,60 @@ export class CompactionEngine {
     if (estimateTokens(summaryText) >= inputTokens) {
       const aggressiveSummary = await runSummarizer(true);
       if (aggressiveSummary === null) {
-        return null;
+        return buildDeterministicFallback();
       }
       summaryText = aggressiveSummary;
       level = "aggressive";
 
       if (estimateTokens(summaryText) >= inputTokens) {
-        const truncated =
-          sourceText.length > FALLBACK_MAX_CHARS
-            ? sourceText.slice(0, FALLBACK_MAX_CHARS)
-            : sourceText;
-        summaryText = `${truncated}
-[Truncated from ${inputTokens} tokens]`;
-        level = "fallback";
+        return buildDeterministicFallback();
       }
     }
 
     return { content: summaryText, level };
   }
 
+  // ── Private: Media Annotation ────────────────────────────────────────────
+
+  /**
+   * Annotate a message's content with media context when it has file/media
+   * attachments. This gives the summarizer enough context to produce a
+   * meaningful summary instead of trying to compress raw file paths.
+   *
+   * - Media-only messages (just a file path, no text): content is replaced
+   *   with "[Media attachment]" or "[Image attachment]" etc.
+   * - Media-mostly messages (any real text + attachment): content is annotated
+   *   with " [with media attachment]" suffix.
+   * - Text-only messages: returned unchanged.
+   */
+  private async annotateMediaContent(
+    messageId: number,
+    content: string,
+  ): Promise<string> {
+    const parts = await this.conversationStore.getMessageParts(messageId);
+    const hasMediaParts = parts.some(
+      (p) => p.partType === "file" || p.partType === "snapshot",
+    );
+    if (!hasMediaParts) {
+      return content;
+    }
+
+    // Strip MEDIA:/... paths to see how much actual text remains
+    const textWithoutPaths = content
+      .split("\n")
+      .filter((line) => !MEDIA_PATH_RE.test(line.trim()))
+      .join("\n")
+      .trim();
+
+    if (textWithoutPaths.length === 0) {
+      // Media-only: replace with descriptive annotation
+      return "[Media attachment]";
+    }
+
+    // Media-mostly: keep the text, add annotation
+    return `${textWithoutPaths} [with media attachment]`;
+  }
+
   // ── Private: Leaf Pass ───────────────────────────────────────────────────
 
   /**
@@ -1044,6 +1109,7 @@ export class CompactionEngine {
     messageItems: ContextItemRecord[],
     summarize: CompactionSummarizeFn,
     previousSummaryContent?: string,
+    summaryModel?: string,
   ): Promise<{ summaryId: string; level: CompactionLevel; content: string } | null> {
     // Fetch full message content for each context item
     const messageContents: { messageId: number; content: string; createdAt: Date; tokenCount: number }[] =
@@ -1054,9 +1120,13 @@ export class CompactionEngine {
       }
       const msg = await this.conversationStore.getMessageById(item.messageId);
       if (msg) {
+        const annotatedContent = await this.annotateMediaContent(
+          msg.messageId,
+          msg.content,
+        );
         messageContents.push({
           messageId: msg.messageId,
-          content: msg.content,
+          content: annotatedContent,
           createdAt: msg.createdAt,
           tokenCount: this.resolveMessageTokenCount(msg),
         });
@@ -1110,6 +1180,7 @@ export class CompactionEngine {
         (sum, message) => sum + Math.max(0, Math.floor(message.tokenCount)),
         0,
       ),
+      model: summaryModel,
     });
 
     // Link to source messages
@@ -1141,6 +1212,7 @@ export class CompactionEngine {
     summaryItems: ContextItemRecord[],
     targetDepth: number,
     summarize: CompactionSummarizeFn,
+    summaryModel?: string,
   ): Promise<PassResult | null> {
     // Fetch full summary records
     const summaryRecords: SummaryRecord[] = [];
@@ -1242,6 +1314,7 @@ export class CompactionEngine {
             : 0;
         return count + sourceTokens;
       }, 0),
+      model: summaryModel,
     });
 
     // Link to parent summaries

package/src/db/connection.ts

@@ -3,6 +3,7 @@ import { dirname, resolve } from "node:path";
 import { DatabaseSync } from "node:sqlite";
 
 type ConnectionKey = string;
+const SQLITE_BUSY_TIMEOUT_MS = 5_000;
 
 const connectionsByPath = new Map<ConnectionKey, Set<DatabaseSync>>();
 const connectionIndex = new Map<DatabaseSync, ConnectionKey>();
@@ -29,6 +30,7 @@ function ensureDbDirectory(dbPath: string): void {
 
 function configureConnection(db: DatabaseSync): DatabaseSync {
   db.exec("PRAGMA journal_mode = WAL");
+  db.exec(`PRAGMA busy_timeout = ${SQLITE_BUSY_TIMEOUT_MS}`);
   db.exec("PRAGMA foreign_keys = ON");
   return db;
 }

package/src/db/migration.ts

@@ -80,6 +80,14 @@ function isoStringOrNull(value: Date | null): string | null {
   return value ? value.toISOString() : null;
 }
 
+function ensureSummaryModelColumn(db: DatabaseSync): void {
+  const summaryColumns = db.prepare(`PRAGMA table_info(summaries)`).all() as SummaryColumnInfo[];
+  const hasModel = summaryColumns.some((col) => col.name === "model");
+  if (!hasModel) {
+    db.exec(`ALTER TABLE summaries ADD COLUMN model TEXT NOT NULL DEFAULT 'unknown'`);
+  }
+}
+
 function backfillSummaryDepths(db: DatabaseSync): void {
   // Leaves are always depth 0, even if legacy rows had malformed values.
   db.exec(`UPDATE summaries SET depth = 0 WHERE kind = 'leaf'`);
@@ -474,6 +482,16 @@ export function runLcmMigrations(
       created_at TEXT NOT NULL DEFAULT (datetime('now'))
     );
 
+    CREATE TABLE IF NOT EXISTS conversation_bootstrap_state (
+      conversation_id INTEGER PRIMARY KEY REFERENCES conversations(conversation_id) ON DELETE CASCADE,
+      session_file_path TEXT NOT NULL,
+      last_seen_size INTEGER NOT NULL,
+      last_seen_mtime_ms INTEGER NOT NULL,
+      last_processed_offset INTEGER NOT NULL,
+      last_processed_entry_hash TEXT,
+      updated_at TEXT NOT NULL DEFAULT (datetime('now'))
+    );
+
     -- Indexes
     CREATE INDEX IF NOT EXISTS messages_conv_seq_idx ON messages (conversation_id, seq);
     CREATE INDEX IF NOT EXISTS summaries_conv_created_idx ON summaries (conversation_id, created_at);
@@ -481,6 +499,8 @@ export function runLcmMigrations(
     CREATE INDEX IF NOT EXISTS message_parts_type_idx ON message_parts (part_type);
     CREATE INDEX IF NOT EXISTS context_items_conv_idx ON context_items (conversation_id, ordinal);
     CREATE INDEX IF NOT EXISTS large_files_conv_idx ON large_files (conversation_id, created_at);
+    CREATE INDEX IF NOT EXISTS bootstrap_state_path_idx
+      ON conversation_bootstrap_state (session_file_path, updated_at);
   `);
 
   // Forward-compatible conversations migration for existing DBs.
@@ -500,6 +520,7 @@ export function runLcmMigrations(
   db.exec(`CREATE UNIQUE INDEX IF NOT EXISTS conversations_session_key_idx ON conversations (session_key)`);
   ensureSummaryDepthColumn(db);
   ensureSummaryMetadataColumns(db);
+  ensureSummaryModelColumn(db);
   backfillSummaryDepths(db);
   backfillSummaryMetadata(db);