@astragenie/astramemory-local 0.7.2

package/dist/distill/stages/02-normalize.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"02-normalize.js","sourceRoot":"","sources":["../../../src/distill/stages/02-normalize.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,OAAO,EAAE,OAAO,EAAE,MAAM,SAAS,CAAC;AAElC;;;GAGG;AACH,MAAM,aAAa,GAA2B;IAC5C,aAAa,EAAE,aAAa;IAC5B,aAAa,EAAE,aAAa;IAC5B,YAAY,EAAE,aAAa;IAC3B,QAAQ,EAAE,QAAQ;IAClB,SAAS,EAAE,gBAAgB;IAC3B,gBAAgB,EAAE,gBAAgB;IAClC,gBAAgB,EAAE,gBAAgB;CACnC,CAAC;AAEF,+DAA+D;AAC/D,MAAM,eAAe,GAAG,mCAAmC,CAAC;AAE5D,0DAA0D;AAC1D,MAAM,gBAAgB,GAAG,mCAAmC,CAAC;AAE7D;;;;;GAKG;AACH,MAAM,iBAAiB,GAAG,sEAAsE,CAAC;AAEjG,MAAM,IAAI,GAAG,OAAO,EAAE,CAAC;AAEvB,MAAM,UAAU,SAAS,CAAC,IAAY;IACpC,IAAI,MAAM,GAAG,IAAI,CAAC;IAElB,2DAA2D;IAC3D,MAAM,GAAG,MAAM,CAAC,OAAO,CAAC,eAAe,EAAE,CAAC,KAAK,EAAE,EAAE;QACjD,sDAAsD;QACtD,MAAM,YAAY,GAAG,KAAK,CAAC,OAAO,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC;QAC/C,OAAO,YAAY,CAAC,YAAY,CAAC,CAAC;IACpC,CAAC,CAAC,CAAC;IAEH,0DAA0D;IAC1D,MAAM,GAAG,MAAM,CAAC,OAAO,CAAC,gBAAgB,EAAE,CAAC,KAAK,EAAE,EAAE;QAClD,OAAO,YAAY,CAAC,KAAK,CAAC,CAAC;IAC7B,CAAC,CAAC,CAAC;IAEH,8CAA8C;IAC9C,MAAM,GAAG,MAAM,CAAC,OAAO,CAAC,iBAAiB,EAAE,CAAC,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,EAAE;QAChE,OAAO,GAAG,IAAI,IAAI,IAAI,GAAG,CAAC;IAC5B,CAAC,CAAC,CAAC;IAEH,uCAAuC;IACvC,MAAM,GAAG,mBAAmB,CAAC,MAAM,CAAC,CAAC;IAErC,OAAO,MAAM,CAAC;AAChB,CAAC;AAED,SAAS,YAAY,CAAC,IAAY;IAChC,uDAAuD;IACvD,MAAM,WAAW,GAAG,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC;IAC7C,MAAM,WAAW,GAAG,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,GAAG,CAAC,CAAC;IAE7C,IAAI,WAAW,CAAC,UAAU,CAAC,WAAW,CAAC,EAAE,CAAC;QACxC,OAAO,GAAG,GAAG,WAAW,CAAC,KAAK,CAAC,WAAW,CAAC,MAAM,CAAC,CAAC;IACrD,CAAC;IACD,gFAAgF;IAChF,8CAA8C;IAC9C,OAAO,WAAW,CAAC;AACrB,CAAC;AAED,SAAS,mBAAmB,CAAC,IAAY;IACvC,0DAA0D;IAC1D,KAAK,MAAM,CAAC,KAAK,EAAE,SAAS,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,aAAa,CAAC,EAAE,CAAC;QAC/D,sCAAsC;QACtC,MAAM,OAAO,GAAG,KAAK,CAAC,OAAO,CAAC,0BAA0B,EAAE,MAAM,CAAC,CAAC;QAClE,MAAM,EAAE,GAAG,IAAI,MAAM,CAAC,MAAM,OAAO,KAAK,EAAE,IAAI,CAAC,CAAC;QAChD,IAAI,GAAG,IAAI,CAAC,OAAO,CAAC,EAAE,EAAE,SAAS,CAAC,CAAC;IACrC,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC"}

package/dist/distill/stages/03-chunk.d.ts

@@ -0,0 +1,22 @@
+/**
+ * Stage 3 — Chunk (deterministic)
+ *
+ * Token-aware split using word-count proxy (~800 tokens ≈ ~3200 chars).
+ * Respects turn boundaries — does not split mid role: block.
+ *
+ * Turn detection: lines starting with "user:", "assistant:", "system:",
+ * "human:", "ai:" (case-insensitive) are treated as turn boundaries.
+ */
+export interface Chunk {
+    index: number;
+    text: string;
+}
+/**
+ * Split text into chunks respecting turn boundaries.
+ *
+ * Algorithm:
+ * 1. Split into "turns" — sequences of lines starting with a role: prefix.
+ * 2. Pack turns into chunks until MAX_CHUNK_CHARS would be exceeded.
+ * 3. If a single turn exceeds MAX_CHUNK_CHARS, split it by paragraph then by line.
+ */
+export declare function chunk(text: string): Chunk[];

package/dist/distill/stages/03-chunk.js

@@ -0,0 +1,138 @@
+/**
+ * Stage 3 — Chunk (deterministic)
+ *
+ * Token-aware split using word-count proxy (~800 tokens ≈ ~3200 chars).
+ * Respects turn boundaries — does not split mid role: block.
+ *
+ * Turn detection: lines starting with "user:", "assistant:", "system:",
+ * "human:", "ai:" (case-insensitive) are treated as turn boundaries.
+ */
+/** Approximate chars per token (conservative estimate) */
+const CHARS_PER_TOKEN = 4;
+/** Target tokens per chunk */
+const TARGET_TOKENS = 800;
+/** Max chars per chunk before we must split even mid-turn */
+const MAX_CHUNK_CHARS = TARGET_TOKENS * CHARS_PER_TOKEN; // 3200
+/** Turn boundary patterns */
+const TURN_BOUNDARY_RE = /^(user|assistant|system|human|ai)\s*:/i;
+/**
+ * Split text into chunks respecting turn boundaries.
+ *
+ * Algorithm:
+ * 1. Split into "turns" — sequences of lines starting with a role: prefix.
+ * 2. Pack turns into chunks until MAX_CHUNK_CHARS would be exceeded.
+ * 3. If a single turn exceeds MAX_CHUNK_CHARS, split it by paragraph then by line.
+ */
+export function chunk(text) {
+    const turns = splitIntoTurns(text);
+    return packTurns(turns);
+}
+function splitIntoTurns(text) {
+    const lines = text.split('\n');
+    const turns = [];
+    let current = [];
+    for (const line of lines) {
+        if (TURN_BOUNDARY_RE.test(line) && current.length > 0) {
+            turns.push({ text: current.join('\n') });
+            current = [line];
+        }
+        else {
+            current.push(line);
+        }
+    }
+    if (current.length > 0) {
+        const t = current.join('\n').trim();
+        if (t)
+            turns.push({ text: t });
+    }
+    return turns;
+}
+function packTurns(turns) {
+    const chunks = [];
+    let current = '';
+    let idx = 0;
+    for (const turn of turns) {
+        const turnText = turn.text.trim();
+        if (!turnText)
+            continue;
+        // If this single turn is too large, flush current and split the turn
+        if (turnText.length > MAX_CHUNK_CHARS) {
+            if (current.trim()) {
+                chunks.push({ index: idx++, text: current.trim() });
+                current = '';
+            }
+            // Split the oversized turn by paragraph
+            const subChunks = splitLargeTurn(turnText, idx);
+            for (const sc of subChunks) {
+                chunks.push({ index: idx++, text: sc.text });
+            }
+            continue;
+        }
+        // Would adding this turn exceed the limit?
+        const candidate = current ? current + '\n\n' + turnText : turnText;
+        if (candidate.length > MAX_CHUNK_CHARS && current.trim()) {
+            chunks.push({ index: idx++, text: current.trim() });
+            current = turnText;
+        }
+        else {
+            current = candidate;
+        }
+    }
+    if (current.trim()) {
+        chunks.push({ index: idx, text: current.trim() });
+    }
+    // Edge case: empty input
+    if (chunks.length === 0) {
+        return [{ index: 0, text: '' }];
+    }
+    return chunks;
+}
+/**
+ * Split a single turn that exceeds MAX_CHUNK_CHARS into sub-chunks.
+ * First tries paragraph boundaries, then line-by-line as fallback.
+ */
+function splitLargeTurn(text, startIdx) {
+    const paragraphs = text.split(/\n\n+/);
+    const chunks = [];
+    let current = '';
+    let idx = startIdx;
+    for (const para of paragraphs) {
+        if (!para.trim())
+            continue;
+        if (para.length > MAX_CHUNK_CHARS) {
+            // Split by line
+            if (current.trim()) {
+                chunks.push({ index: idx++, text: current.trim() });
+                current = '';
+            }
+            const lines = para.split('\n');
+            for (const line of lines) {
+                if (!line.trim())
+                    continue;
+                const cand = current ? current + '\n' + line : line;
+                if (cand.length > MAX_CHUNK_CHARS && current.trim()) {
+                    chunks.push({ index: idx++, text: current.trim() });
+                    current = line;
+                }
+                else {
+                    current = cand;
+                }
+            }
+        }
+        else {
+            const cand = current ? current + '\n\n' + para : para;
+            if (cand.length > MAX_CHUNK_CHARS && current.trim()) {
+                chunks.push({ index: idx++, text: current.trim() });
+                current = para;
+            }
+            else {
+                current = cand;
+            }
+        }
+    }
+    if (current.trim()) {
+        chunks.push({ index: idx, text: current.trim() });
+    }
+    return chunks;
+}
+//# sourceMappingURL=03-chunk.js.map

package/dist/distill/stages/03-chunk.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"03-chunk.js","sourceRoot":"","sources":["../../../src/distill/stages/03-chunk.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,0DAA0D;AAC1D,MAAM,eAAe,GAAG,CAAC,CAAC;AAE1B,8BAA8B;AAC9B,MAAM,aAAa,GAAG,GAAG,CAAC;AAE1B,6DAA6D;AAC7D,MAAM,eAAe,GAAG,aAAa,GAAG,eAAe,CAAC,CAAC,OAAO;AAEhE,6BAA6B;AAC7B,MAAM,gBAAgB,GAAG,wCAAwC,CAAC;AAOlE;;;;;;;GAOG;AACH,MAAM,UAAU,KAAK,CAAC,IAAY;IAChC,MAAM,KAAK,GAAG,cAAc,CAAC,IAAI,CAAC,CAAC;IACnC,OAAO,SAAS,CAAC,KAAK,CAAC,CAAC;AAC1B,CAAC;AAMD,SAAS,cAAc,CAAC,IAAY;IAClC,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;IAC/B,MAAM,KAAK,GAAW,EAAE,CAAC;IACzB,IAAI,OAAO,GAAa,EAAE,CAAC;IAE3B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,IAAI,gBAAgB,CAAC,IAAI,CAAC,IAAI,CAAC,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;YACtD,KAAK,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;YACzC,OAAO,GAAG,CAAC,IAAI,CAAC,CAAC;QACnB,CAAC;aAAM,CAAC;YACN,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACrB,CAAC;IACH,CAAC;IAED,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACvB,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,CAAC;QACpC,IAAI,CAAC;YAAE,KAAK,CAAC,IAAI,CAAC,EAAE,IAAI,EAAE,CAAC,EAAE,CAAC,CAAC;IACjC,CAAC;IAED,OAAO,KAAK,CAAC;AACf,CAAC;AAED,SAAS,SAAS,CAAC,KAAa;IAC9B,MAAM,MAAM,GAAY,EAAE,CAAC;IAC3B,IAAI,OAAO,GAAG,EAAE,CAAC;IACjB,IAAI,GAAG,GAAG,CAAC,CAAC;IAEZ,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,QAAQ,GAAG,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,CAAC;QAClC,IAAI,CAAC,QAAQ;YAAE,SAAS;QAExB,qEAAqE;QACrE,IAAI,QAAQ,CAAC,MAAM,GAAG,eAAe,EAAE,CAAC;YACtC,IAAI,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC;gBACnB,MAAM,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,EAAE,IAAI,EAAE,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;gBACpD,OAAO,GAAG,EAAE,CAAC;YACf,CAAC;YACD,wCAAwC;YACxC,MAAM,SAAS,GAAG,cAAc,CAAC,QAAQ,EAAE,GAAG,CAAC,CAAC;YAChD,KAAK,MAAM,EAAE,IAAI,SAAS,EAAE,CAAC;gBAC3B,MAAM,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,EAAE,IAAI,EAAE,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC;YAC/C,CAAC;YACD,SAAS;QACX,CAAC;QAED,2CAA2C;QAC3C,MAAM,SAAS,GAAG,OAAO,CAAC,CAAC,CAAC,OAAO,GAAG,MAAM,GAAG,QAAQ,CAAC,CAAC,CAAC,QAAQ,CAAC;QACnE,IAAI,SAAS,CAAC,MAAM,GAAG,eAAe,IAAI,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC;YACzD,MAAM,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,EAAE,IAAI,EAAE,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;YACpD,OAAO,GAAG,QAAQ,CAAC;QACrB,CAAC;aAAM,CAAC;YACN,OAAO,GAAG,SAAS,CAAC;QACtB,CAAC;IACH,CAAC;IAED,IAAI,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC;QACnB,MAAM,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,IAAI,EAAE,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;IACpD,CAAC;IAED,yBAAyB;IACzB,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACxB,OAAO,CAAC,EAAE,KAAK,EAAE,CAAC,EAAE,IAAI,EAAE,EAAE,EAAE,CAAC,CAAC;IAClC,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC;AAED;;;GAGG;AACH,SAAS,cAAc,CAAC,IAAY,EAAE,QAAgB;IACpD,MAAM,UAAU,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;IACvC,MAAM,MAAM,GAAY,EAAE,CAAC;IAC3B,IAAI,OAAO,GAAG,EAAE,CAAC;IACjB,IAAI,GAAG,GAAG,QAAQ,CAAC;IAEnB,KAAK,MAAM,IAAI,IAAI,UAAU,EAAE,CAAC;QAC9B,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE;YAAE,SAAS;QAE3B,IAAI,IAAI,CAAC,MAAM,GAAG,eAAe,EAAE,CAAC;YAClC,gBAAgB;YAChB,IAAI,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC;gBACnB,MAAM,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,EAAE,IAAI,EAAE,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;gBACpD,OAAO,GAAG,EAAE,CAAC;YACf,CAAC;YACD,MAAM,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC;YAC/B,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;gBACzB,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE;oBAAE,SAAS;gBAC3B,MAAM,IAAI,GAAG,OAAO,CAAC,CAAC,CAAC,OAAO,GAAG,IAAI,GAAG,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC;gBACpD,IAAI,IAAI,CAAC,MAAM,GAAG,eAAe,IAAI,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC;oBACpD,MAAM,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,EAAE,IAAI,EAAE,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;oBACpD,OAAO,GAAG,IAAI,CAAC;gBACjB,CAAC;qBAAM,CAAC;oBACN,OAAO,GAAG,IAAI,CAAC;gBACjB,CAAC;YACH,CAAC;QACH,CAAC;aAAM,CAAC;YACN,MAAM,IAAI,GAAG,OAAO,CAAC,CAAC,CAAC,OAAO,GAAG,MAAM,GAAG,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC;YACtD,IAAI,IAAI,CAAC,MAAM,GAAG,eAAe,IAAI,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC;gBACpD,MAAM,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,EAAE,IAAI,EAAE,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;gBACpD,OAAO,GAAG,IAAI,CAAC;YACjB,CAAC;iBAAM,CAAC;gBACN,OAAO,GAAG,IAAI,CAAC;YACjB,CAAC;QACH,CAAC;IACH,CAAC;IAED,IAAI,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC;QACnB,MAAM,CAAC,IAAI,CAAC,EAAE,KAAK,EAAE,GAAG,EAAE,IAAI,EAAE,OAAO,CAAC,IAAI,EAAE,EAAE,CAAC,CAAC;IACpD,CAAC;IAED,OAAO,MAAM,CAAC;AAChB,CAAC"}

package/dist/distill/stages/04-compact.d.ts

@@ -0,0 +1,28 @@
+/**
+ * Stage 4 — Compact (LLM call per chunk)
+ *
+ * Removes redundancy, repeated thinking, false starts from each chunk.
+ * Preserves decisions, file paths, commands, errors, rationale.
+ *
+ * Budget: checks cap before each call, records after.
+ * Throws BudgetExceeded if over cap (caller moves job to paused).
+ */
+import type { LLMProvider } from '../../contracts/index.js';
+import type { Chunk } from './03-chunk.js';
+import { BudgetTracker, BudgetExceeded } from '../../budget/tracker.js';
+export { BudgetExceeded };
+export interface CompactResult {
+    index: number;
+    text: string;
+    usageUsd: number;
+}
+/**
+ * Compact a single chunk via LLM.
+ * Returns the compressed text and actual USD cost.
+ */
+export declare function compactChunk(chk: Chunk, provider: LLMProvider, budget: BudgetTracker, capUsd: number): Promise<CompactResult>;
+/**
+ * Compact all chunks in sequence.
+ * Stops at first BudgetExceeded — caller catches and pauses job.
+ */
+export declare function compact(chunks: Chunk[], provider: LLMProvider, budget: BudgetTracker, capUsd: number): Promise<CompactResult[]>;

package/dist/distill/stages/04-compact.js

@@ -0,0 +1,69 @@
+/**
+ * Stage 4 — Compact (LLM call per chunk)
+ *
+ * Removes redundancy, repeated thinking, false starts from each chunk.
+ * Preserves decisions, file paths, commands, errors, rationale.
+ *
+ * Budget: checks cap before each call, records after.
+ * Throws BudgetExceeded if over cap (caller moves job to paused).
+ */
+import { BudgetTracker, BudgetExceeded } from '../../budget/tracker.js';
+export { BudgetExceeded };
+const COMPACT_SYSTEM = `You are a transcript compressor for AI coding agent sessions.
+
+Compress the provided transcript chunk. Remove:
+- Repeated or near-identical lines
+- Thinking-out-loud that reaches no conclusion
+- False starts and self-corrections that don't change the outcome
+- Generic acknowledgment phrases ("Sure!", "Of course!", "I'll help you with that")
+
+PRESERVE:
+- All decisions and chosen approaches
+- All file paths and commands
+- All error messages and their resolutions
+- All rationale and architectural reasoning
+- All TODO items and pending work
+
+Output plain text only. No markdown formatting. No commentary.`;
+/**
+ * Compact a single chunk via LLM.
+ * Returns the compressed text and actual USD cost.
+ */
+export async function compactChunk(chk, provider, budget, capUsd) {
+    if (!chk.text.trim()) {
+        return { index: chk.index, text: '', usageUsd: 0 };
+    }
+    const messages = [
+        { role: 'system', content: COMPACT_SYSTEM },
+        { role: 'user', content: chk.text },
+    ];
+    // Estimate cost: prompt = system + user text chars
+    const promptChars = COMPACT_SYSTEM.length + chk.text.length;
+    // For Ollama providers, usd=0 always — use a negligible estimate to pass budget check
+    // For Azure, pricing is handled after the call via usage.usd
+    const estimateUsd = BudgetTracker.estimateUsd(promptChars, 0.000002, 0.000002);
+    budget.assertCanSpend(estimateUsd, capUsd);
+    const result = await provider.chat(messages, {
+        temperature: 0.1,
+        maxTokens: 2000,
+    });
+    // Record actual cost (0 for Ollama, real for Azure)
+    budget.record(result.usage.usd);
+    return {
+        index: chk.index,
+        text: result.text.trim(),
+        usageUsd: result.usage.usd,
+    };
+}
+/**
+ * Compact all chunks in sequence.
+ * Stops at first BudgetExceeded — caller catches and pauses job.
+ */
+export async function compact(chunks, provider, budget, capUsd) {
+    const results = [];
+    for (const chk of chunks) {
+        results.push(await compactChunk(chk, provider, budget, capUsd));
+    }
+    return results;
+}
+//# sourceMappingURL=04-compact.js.map

package/dist/distill/stages/04-compact.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"04-compact.js","sourceRoot":"","sources":["../../../src/distill/stages/04-compact.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAIH,OAAO,EAAE,aAAa,EAAE,cAAc,EAAE,MAAM,yBAAyB,CAAC;AAExE,OAAO,EAAE,cAAc,EAAE,CAAC;AAE1B,MAAM,cAAc,GAAG;;;;;;;;;;;;;;;+DAewC,CAAC;AAQhE;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,YAAY,CAChC,GAAU,EACV,QAAqB,EACrB,MAAqB,EACrB,MAAc;IAEd,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,IAAI,EAAE,EAAE,CAAC;QACrB,OAAO,EAAE,KAAK,EAAE,GAAG,CAAC,KAAK,EAAE,IAAI,EAAE,EAAE,EAAE,QAAQ,EAAE,CAAC,EAAE,CAAC;IACrD,CAAC;IAED,MAAM,QAAQ,GAAc;QAC1B,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,cAAc,EAAE;QAC3C,EAAE,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,GAAG,CAAC,IAAI,EAAE;KACpC,CAAC;IAEF,mDAAmD;IACnD,MAAM,WAAW,GAAG,cAAc,CAAC,MAAM,GAAG,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC;IAC5D,sFAAsF;IACtF,6DAA6D;IAC7D,MAAM,WAAW,GAAG,aAAa,CAAC,WAAW,CAAC,WAAW,EAAE,QAAQ,EAAE,QAAQ,CAAC,CAAC;IAE/E,MAAM,CAAC,cAAc,CAAC,WAAW,EAAE,MAAM,CAAC,CAAC;IAE3C,MAAM,MAAM,GAAG,MAAM,QAAQ,CAAC,IAAI,CAAC,QAAQ,EAAE;QAC3C,WAAW,EAAE,GAAG;QAChB,SAAS,EAAE,IAAI;KAChB,CAAC,CAAC;IAEH,oDAAoD;IACpD,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC;IAEhC,OAAO;QACL,KAAK,EAAE,GAAG,CAAC,KAAK;QAChB,IAAI,EAAE,MAAM,CAAC,IAAI,CAAC,IAAI,EAAE;QACxB,QAAQ,EAAE,MAAM,CAAC,KAAK,CAAC,GAAG;KAC3B,CAAC;AACJ,CAAC;AAED;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,OAAO,CAC3B,MAAe,EACf,QAAqB,EACrB,MAAqB,EACrB,MAAc;IAEd,MAAM,OAAO,GAAoB,EAAE,CAAC;IACpC,KAAK,MAAM,GAAG,IAAI,MAAM,EAAE,CAAC;QACzB,OAAO,CAAC,IAAI,CAAC,MAAM,YAAY,CAAC,GAAG,EAAE,QAAQ,EAAE,MAAM,EAAE,MAAM,CAAC,CAAC,CAAC;IAClE,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC"}

package/dist/distill/stages/05-extract.d.ts

@@ -0,0 +1,35 @@
+/**
+ * Stage 5 — Extract (LLM JSON-mode per chunk)
+ *
+ * Emits typed atoms {type, text, importance, confidence, evidence}.
+ * Zod-validates the JSON response. Retries once with a stricter prompt
+ * on parse failure. Throws BudgetExceeded if over cap.
+ */
+import type { LLMProvider } from '../../contracts/index.js';
+import { type Atom } from '../prompts/extract.js';
+import { BudgetTracker, BudgetExceeded } from '../../budget/tracker.js';
+export { BudgetExceeded };
+export type { Atom };
+export interface ExtractResult {
+    chunkIndex: number;
+    atoms: Atom[];
+    usageUsd: number;
+    retried: boolean;
+}
+/**
+ * Extract atoms from a single compacted chunk.
+ *
+ * Parse / Zod validation failure → throws DeterministicError immediately.
+ * The worker catches this, poisons the job after a single attempt, and does not
+ * issue a second LLM call. This prevents burning budget on a malformed transcript
+ * that will never produce valid JSON regardless of how many times we retry.
+ */
+export declare function extractChunk(chunkIndex: number, text: string, provider: LLMProvider, budget: BudgetTracker, capUsd: number): Promise<ExtractResult>;
+/**
+ * Extract atoms from all compacted chunks.
+ * Stops at first BudgetExceeded.
+ */
+export declare function extract(chunks: Array<{
+    index: number;
+    text: string;
+}>, provider: LLMProvider, budget: BudgetTracker, capUsd: number): Promise<ExtractResult[]>;

package/dist/distill/stages/05-extract.js

@@ -0,0 +1,101 @@
+/**
+ * Stage 5 — Extract (LLM JSON-mode per chunk)
+ *
+ * Emits typed atoms {type, text, importance, confidence, evidence}.
+ * Zod-validates the JSON response. Retries once with a stricter prompt
+ * on parse failure. Throws BudgetExceeded if over cap.
+ */
+import { z } from 'zod';
+import { AtomSchema, EXTRACTION_SYSTEM_PROMPT, } from '../prompts/extract.js';
+import { BudgetTracker, BudgetExceeded } from '../../budget/tracker.js';
+import { DeterministicError } from '../../pipeline/errors.js';
+/**
+ * Lenient wrapper: accepts any objects in the atoms array so that individual
+ * bad atoms (short text, unknown type, etc.) don't reject the whole response.
+ * Valid atoms are filtered with AtomSchema after JSON parsing.
+ */
+const LenientExtractionSchema = z.object({ atoms: z.array(z.unknown()) });
+export { BudgetExceeded };
+/**
+ * Extract atoms from a single compacted chunk.
+ *
+ * Parse / Zod validation failure → throws DeterministicError immediately.
+ * The worker catches this, poisons the job after a single attempt, and does not
+ * issue a second LLM call. This prevents burning budget on a malformed transcript
+ * that will never produce valid JSON regardless of how many times we retry.
+ */
+export async function extractChunk(chunkIndex, text, provider, budget, capUsd) {
+    if (!text.trim()) {
+        return { chunkIndex, atoms: [], usageUsd: 0, retried: false };
+    }
+    const promptChars = EXTRACTION_SYSTEM_PROMPT.length + text.length;
+    const estimateUsd = BudgetTracker.estimateUsd(promptChars, 0.000002, 0.000002);
+    budget.assertCanSpend(estimateUsd, capUsd);
+    const result = await callExtract(text, EXTRACTION_SYSTEM_PROMPT, provider);
+    budget.record(result.usageUsd);
+    const parsed = tryParseAtoms(result.text);
+    if (parsed !== null) {
+        return { chunkIndex, atoms: parsed, usageUsd: result.usageUsd, retried: false };
+    }
+    // Parse / schema failure — deterministic. Let the worker decide retry policy.
+    throw new DeterministicError(`[extract] chunk ${chunkIndex}: Zod/JSON parse failed — raw: ${result.text.slice(0, 200)}`);
+}
+async function callExtract(text, systemPrompt, provider) {
+    const messages = [
+        { role: 'system', content: systemPrompt },
+        { role: 'user', content: `Extract knowledge atoms from this transcript chunk:\n\n${text}` },
+    ];
+    const result = await provider.chat(messages, {
+        temperature: 0.1,
+        json: true,
+        maxTokens: 1500,
+    });
+    return { text: result.text, usageUsd: result.usage.usd };
+}
+/**
+ * Try to parse the LLM response into an Atom[].
+ * Strips markdown fences if present.
+ *
+ * Returns an Atom[] (possibly empty) when the JSON is structurally valid and
+ * contains an `atoms` array — even if individual atoms are malformed (they
+ * are silently dropped via AtomSchema filtering).
+ *
+ * Returns null only when the raw text is not parseable JSON or is missing the
+ * top-level `atoms` key — this is the signal for a DeterministicError.
+ */
+function tryParseAtoms(raw) {
+    try {
+        let text = raw.trim();
+        // Strip markdown code fences
+        text = text.replace(/^```(?:json)?\n?/i, '').replace(/\n?```\s*$/i, '').trim();
+        // Find the outermost JSON object
+        const start = text.indexOf('{');
+        const end = text.lastIndexOf('}');
+        if (start === -1 || end === -1)
+            return null;
+        const jsonStr = text.slice(start, end + 1);
+        const parsed = JSON.parse(jsonStr);
+        // Use the lenient schema — only requires `atoms` to be an array
+        const wrapper = LenientExtractionSchema.safeParse(parsed);
+        if (!wrapper.success)
+            return null;
+        // Filter atoms individually: drop those that fail strict AtomSchema
+        // (short text, unknown type, out-of-range numbers, etc.)
+        return wrapper.data.atoms.filter((a) => AtomSchema.safeParse(a).success);
+    }
+    catch {
+        return null;
+    }
+}
+/**
+ * Extract atoms from all compacted chunks.
+ * Stops at first BudgetExceeded.
+ */
+export async function extract(chunks, provider, budget, capUsd) {
+    const results = [];
+    for (const chk of chunks) {
+        results.push(await extractChunk(chk.index, chk.text, provider, budget, capUsd));
+    }
+    return results;
+}
+//# sourceMappingURL=05-extract.js.map

package/dist/distill/stages/05-extract.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"05-extract.js","sourceRoot":"","sources":["../../../src/distill/stages/05-extract.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,OAAO,EAAE,CAAC,EAAE,MAAM,KAAK,CAAC;AAExB,OAAO,EACL,UAAU,EACV,wBAAwB,GAEzB,MAAM,uBAAuB,CAAC;AAC/B,OAAO,EAAE,aAAa,EAAE,cAAc,EAAE,MAAM,yBAAyB,CAAC;AACxE,OAAO,EAAE,kBAAkB,EAAE,MAAM,0BAA0B,CAAC;AAE9D;;;;GAIG;AACH,MAAM,uBAAuB,GAAG,CAAC,CAAC,MAAM,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC;AAE1E,OAAO,EAAE,cAAc,EAAE,CAAC;AAU1B;;;;;;;GAOG;AACH,MAAM,CAAC,KAAK,UAAU,YAAY,CAChC,UAAkB,EAClB,IAAY,EACZ,QAAqB,EACrB,MAAqB,EACrB,MAAc;IAEd,IAAI,CAAC,IAAI,CAAC,IAAI,EAAE,EAAE,CAAC;QACjB,OAAO,EAAE,UAAU,EAAE,KAAK,EAAE,EAAE,EAAE,QAAQ,EAAE,CAAC,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC;IAChE,CAAC;IAED,MAAM,WAAW,GAAG,wBAAwB,CAAC,MAAM,GAAG,IAAI,CAAC,MAAM,CAAC;IAClE,MAAM,WAAW,GAAG,aAAa,CAAC,WAAW,CAAC,WAAW,EAAE,QAAQ,EAAE,QAAQ,CAAC,CAAC;IAE/E,MAAM,CAAC,cAAc,CAAC,WAAW,EAAE,MAAM,CAAC,CAAC;IAE3C,MAAM,MAAM,GAAG,MAAM,WAAW,CAAC,IAAI,EAAE,wBAAwB,EAAE,QAAQ,CAAC,CAAC;IAC3E,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;IAE/B,MAAM,MAAM,GAAG,aAAa,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;IAC1C,IAAI,MAAM,KAAK,IAAI,EAAE,CAAC;QACpB,OAAO,EAAE,UAAU,EAAE,KAAK,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,CAAC,QAAQ,EAAE,OAAO,EAAE,KAAK,EAAE,CAAC;IAClF,CAAC;IAED,8EAA8E;IAC9E,MAAM,IAAI,kBAAkB,CAC1B,mBAAmB,UAAU,kCAAkC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,CAC3F,CAAC;AACJ,CAAC;AAED,KAAK,UAAU,WAAW,CACxB,IAAY,EACZ,YAAoB,EACpB,QAAqB;IAErB,MAAM,QAAQ,GAAc;QAC1B,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,YAAY,EAAE;QACzC,EAAE,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,0DAA0D,IAAI,EAAE,EAAE;KAC5F,CAAC;IAEF,MAAM,MAAM,GAAG,MAAM,QAAQ,CAAC,IAAI,CAAC,QAAQ,EAAE;QAC3C,WAAW,EAAE,GAAG;QAChB,IAAI,EAAE,IAAI;QACV,SAAS,EAAE,IAAI;KAChB,CAAC,CAAC;IAEH,OAAO,EAAE,IAAI,EAAE,MAAM,CAAC,IAAI,EAAE,QAAQ,EAAE,MAAM,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC;AAC3D,CAAC;AAED;;;;;;;;;;GAUG;AACH,SAAS,aAAa,CAAC,GAAW;IAChC,IAAI,CAAC;QACH,IAAI,IAAI,GAAG,GAAG,CAAC,IAAI,EAAE,CAAC;QAEtB,6BAA6B;QAC7B,IAAI,GAAG,IAAI,CAAC,OAAO,CAAC,mBAAmB,EAAE,EAAE,CAAC,CAAC,OAAO,CAAC,aAAa,EAAE,EAAE,CAAC,CAAC,IAAI,EAAE,CAAC;QAE/E,iCAAiC;QACjC,MAAM,KAAK,GAAG,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,CAAC;QAChC,MAAM,GAAG,GAAG,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;QAClC,IAAI,KAAK,KAAK,CAAC,CAAC,IAAI,GAAG,KAAK,CAAC,CAAC;YAAE,OAAO,IAAI,CAAC;QAE5C,MAAM,OAAO,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,GAAG,GAAG,CAAC,CAAC,CAAC;QAC3C,MAAM,MAAM,GAAY,IAAI,CAAC,KAAK,CAAC,OAAO,CAAC,CAAC;QAE5C,gEAAgE;QAChE,MAAM,OAAO,GAAG,uBAAuB,CAAC,SAAS,CAAC,MAAM,CAAC,CAAC;QAC1D,IAAI,CAAC,OAAO,CAAC,OAAO;YAAE,OAAO,IAAI,CAAC;QAElC,oEAAoE;QACpE,yDAAyD;QACzD,OAAO,OAAO,CAAC,IAAI,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,UAAU,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,OAAO,CAAW,CAAC;IACrF,CAAC;IAAC,MAAM,CAAC;QACP,OAAO,IAAI,CAAC;IACd,CAAC;AACH,CAAC;AAED;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,OAAO,CAC3B,MAA8C,EAC9C,QAAqB,EACrB,MAAqB,EACrB,MAAc;IAEd,MAAM,OAAO,GAAoB,EAAE,CAAC;IACpC,KAAK,MAAM,GAAG,IAAI,MAAM,EAAE,CAAC;QACzB,OAAO,CAAC,IAAI,CAAC,MAAM,YAAY,CAAC,GAAG,CAAC,KAAK,EAAE,GAAG,CAAC,IAAI,EAAE,QAAQ,EAAE,MAAM,EAAE,MAAM,CAAC,CAAC,CAAC;IAClF,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC"}

package/dist/distill/stages/06-reduce.d.ts

@@ -0,0 +1,16 @@
+/**
+ * Stage 6 — Reduce (deterministic)
+ *
+ * Merge atoms across chunks by hash(normalized_text).
+ * When duplicates exist, keep the highest-importance instance.
+ */
+import type { Atom } from '../prompts/extract.js';
+export interface ReducedAtom extends Atom {
+    /** Hash of the normalized (lowercased, trimmed, whitespace-collapsed) text */
+    contentHash: string;
+}
+/**
+ * Merge a flat array of atoms, deduplicating by content hash.
+ * On collision, keep the highest-importance instance.
+ */
+export declare function reduce(atoms: Atom[]): ReducedAtom[];

package/dist/distill/stages/06-reduce.js

@@ -0,0 +1,30 @@
+/**
+ * Stage 6 — Reduce (deterministic)
+ *
+ * Merge atoms across chunks by hash(normalized_text).
+ * When duplicates exist, keep the highest-importance instance.
+ */
+import { createHash } from 'node:crypto';
+/**
+ * Merge a flat array of atoms, deduplicating by content hash.
+ * On collision, keep the highest-importance instance.
+ */
+export function reduce(atoms) {
+    const byHash = new Map();
+    for (const atom of atoms) {
+        const hash = contentHash(atom.text);
+        const existing = byHash.get(hash);
+        if (!existing || atom.importance > existing.importance) {
+            byHash.set(hash, { ...atom, contentHash: hash });
+        }
+    }
+    return [...byHash.values()];
+}
+function contentHash(text) {
+    const normalized = text.trim().toLowerCase().replace(/\s+/g, ' ');
+    // Full 64-hex (256-bit) SHA-256 — the previous 16-hex (64-bit) slice had
+    // birthday-bound collision risk around 2^32 atoms which is too low for a
+    // long-running memory corpus.
+    return createHash('sha256').update(normalized, 'utf8').digest('hex');
+}
+//# sourceMappingURL=06-reduce.js.map

package/dist/distill/stages/06-reduce.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"06-reduce.js","sourceRoot":"","sources":["../../../src/distill/stages/06-reduce.ts"],"names":[],"mappings":"AAAA;;;;;GAKG;AAEH,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AAQzC;;;GAGG;AACH,MAAM,UAAU,MAAM,CAAC,KAAa;IAClC,MAAM,MAAM,GAAG,IAAI,GAAG,EAAuB,CAAC;IAE9C,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,IAAI,GAAG,WAAW,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QACpC,MAAM,QAAQ,GAAG,MAAM,CAAC,GAAG,CAAC,IAAI,CAAC,CAAC;QAElC,IAAI,CAAC,QAAQ,IAAI,IAAI,CAAC,UAAU,GAAG,QAAQ,CAAC,UAAU,EAAE,CAAC;YACvD,MAAM,CAAC,GAAG,CAAC,IAAI,EAAE,EAAE,GAAG,IAAI,EAAE,WAAW,EAAE,IAAI,EAAE,CAAC,CAAC;QACnD,CAAC;IACH,CAAC;IAED,OAAO,CAAC,GAAG,MAAM,CAAC,MAAM,EAAE,CAAC,CAAC;AAC9B,CAAC;AAED,SAAS,WAAW,CAAC,IAAY;IAC/B,MAAM,UAAU,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC,WAAW,EAAE,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC;IAClE,yEAAyE;IACzE,yEAAyE;IACzE,8BAA8B;IAC9B,OAAO,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,UAAU,EAAE,MAAM,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;AACvE,CAAC"}

package/dist/distill/stages/07-memory-normalize.d.ts

@@ -0,0 +1,27 @@
+/**
+ * Stage 7 — Memory-normalize (deterministic)
+ *
+ * Apply canonical text rules to each reduced atom:
+ * - Trim and collapse whitespace
+ * - Apply entity dictionary (case canonicalization)
+ * - Compute final SHA-256 hash of the normalized text
+ */
+import type { ReducedAtom } from './06-reduce.js';
+export interface NormalizedMemory extends ReducedAtom {
+    normalizedText: string;
+    finalHash: string;
+}
+/**
+ * Normalize a single atom text to canonical form.
+ */
+export declare function normalizeText(text: string): string;
+/**
+ * Compute a stable hash for a normalized memory text.
+ * This is the dedup key used when inserting into the memories table.
+ */
+export declare function computeHash(normalizedText: string): string;
+/**
+ * Apply memory normalization to all reduced atoms.
+ * Returns atoms with `normalizedText` and `finalHash` added.
+ */
+export declare function memoryNormalize(atoms: ReducedAtom[]): NormalizedMemory[];

package/dist/distill/stages/07-memory-normalize.js

@@ -0,0 +1,65 @@
+/**
+ * Stage 7 — Memory-normalize (deterministic)
+ *
+ * Apply canonical text rules to each reduced atom:
+ * - Trim and collapse whitespace
+ * - Apply entity dictionary (case canonicalization)
+ * - Compute final SHA-256 hash of the normalized text
+ */
+import { createHash } from 'node:crypto';
+/**
+ * Entity dictionary: patterns → canonical replacement.
+ * Applied as case-insensitive whole-word replacements.
+ * Extend here as more entities are discovered.
+ */
+const ENTITY_DICT = [
+    // sqlite-vec MUST precede bare sqlite so hyphenated form is protected first
+    [/\bsqlite-vec\b/gi, 'sqlite-vec'],
+    [/\bsqlite\b(?!-vec)/gi, 'SQLite'],
+    [/\bpostgres(?:ql)?\b/gi, 'PostgreSQL'],
+    [/\bmysql\b/gi, 'MySQL'],
+    [/\bjavascript\b/gi, 'JavaScript'],
+    [/\btypescript\b/gi, 'TypeScript'],
+    [/\bnodejs\b|\bnode\.js\b/gi, 'Node.js'],
+    [/\bnpm\b/g, 'npm'],
+    [/\bbun\b/gi, 'Bun'],
+    [/\bfastify\b/gi, 'Fastify'],
+    [/\bzod\b/gi, 'Zod'],
+    [/\bvitest\b/gi, 'Vitest'],
+    [/\bgithub\b/gi, 'GitHub'],
+    [/\bwindows\b/gi, 'Windows'],
+    [/\bmacos\b|mac os x\b/gi, 'macOS'],
+    [/\blinux\b/gi, 'Linux'],
+];
+/**
+ * Normalize a single atom text to canonical form.
+ */
+export function normalizeText(text) {
+    let t = text.trim();
+    // Collapse internal whitespace runs to single space
+    t = t.replace(/\s+/g, ' ');
+    // Apply entity dictionary
+    for (const [pattern, canonical] of ENTITY_DICT) {
+        t = t.replace(pattern, canonical);
+    }
+    return t;
+}
+/**
+ * Compute a stable hash for a normalized memory text.
+ * This is the dedup key used when inserting into the memories table.
+ */
+export function computeHash(normalizedText) {
+    return createHash('sha256').update(normalizedText, 'utf8').digest('hex');
+}
+/**
+ * Apply memory normalization to all reduced atoms.
+ * Returns atoms with `normalizedText` and `finalHash` added.
+ */
+export function memoryNormalize(atoms) {
+    return atoms.map(atom => {
+        const normalizedText = normalizeText(atom.text);
+        const finalHash = computeHash(normalizedText);
+        return { ...atom, normalizedText, finalHash };
+    });
+}
+//# sourceMappingURL=07-memory-normalize.js.map

package/dist/distill/stages/07-memory-normalize.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"07-memory-normalize.js","sourceRoot":"","sources":["../../../src/distill/stages/07-memory-normalize.ts"],"names":[],"mappings":"AAAA;;;;;;;GAOG;AAEH,OAAO,EAAE,UAAU,EAAE,MAAM,aAAa,CAAC;AAQzC;;;;GAIG;AACH,MAAM,WAAW,GAAgD;IAC/D,4EAA4E;IAC5E,CAAC,kBAAkB,EAAE,YAAY,CAAC;IAClC,CAAC,sBAAsB,EAAE,QAAQ,CAAC;IAClC,CAAC,uBAAuB,EAAE,YAAY,CAAC;IACvC,CAAC,aAAa,EAAE,OAAO,CAAC;IACxB,CAAC,kBAAkB,EAAE,YAAY,CAAC;IAClC,CAAC,kBAAkB,EAAE,YAAY,CAAC;IAClC,CAAC,2BAA2B,EAAE,SAAS,CAAC;IACxC,CAAC,UAAU,EAAE,KAAK,CAAC;IACnB,CAAC,WAAW,EAAE,KAAK,CAAC;IACpB,CAAC,eAAe,EAAE,SAAS,CAAC;IAC5B,CAAC,WAAW,EAAE,KAAK,CAAC;IACpB,CAAC,cAAc,EAAE,QAAQ,CAAC;IAC1B,CAAC,cAAc,EAAE,QAAQ,CAAC;IAC1B,CAAC,eAAe,EAAE,SAAS,CAAC;IAC5B,CAAC,wBAAwB,EAAE,OAAO,CAAC;IACnC,CAAC,aAAa,EAAE,OAAO,CAAC;CACzB,CAAC;AAEF;;GAEG;AACH,MAAM,UAAU,aAAa,CAAC,IAAY;IACxC,IAAI,CAAC,GAAG,IAAI,CAAC,IAAI,EAAE,CAAC;IAEpB,oDAAoD;IACpD,CAAC,GAAG,CAAC,CAAC,OAAO,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC;IAE3B,0BAA0B;IAC1B,KAAK,MAAM,CAAC,OAAO,EAAE,SAAS,CAAC,IAAI,WAAW,EAAE,CAAC;QAC/C,CAAC,GAAG,CAAC,CAAC,OAAO,CAAC,OAAO,EAAE,SAAS,CAAC,CAAC;IACpC,CAAC;IAED,OAAO,CAAC,CAAC;AACX,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,WAAW,CAAC,cAAsB;IAChD,OAAO,UAAU,CAAC,QAAQ,CAAC,CAAC,MAAM,CAAC,cAAc,EAAE,MAAM,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC;AAC3E,CAAC;AAED;;;GAGG;AACH,MAAM,UAAU,eAAe,CAAC,KAAoB;IAClD,OAAO,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,EAAE;QACtB,MAAM,cAAc,GAAG,aAAa,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;QAChD,MAAM,SAAS,GAAG,WAAW,CAAC,cAAc,CAAC,CAAC;QAC9C,OAAO,EAAE,GAAG,IAAI,EAAE,cAAc,EAAE,SAAS,EAAE,CAAC;IAChD,CAAC,CAAC,CAAC;AACL,CAAC"}

package/dist/distill/stages/08-embed-index.d.ts

@@ -0,0 +1,31 @@
+/**
+ * Stage 8 — Embed + Index (provider call + DB writes)
+ *
+ * For each normalized memory:
+ * 1. Insert into memories table via MemoryRepo.insert (hash dedup built in)
+ * 2. Embed the normalizedText via EmbedProvider
+ * 3. Upsert vec row via SqliteVecStore.upsert
+ * 4. Update embedding_provider/model/dim on the memories row
+ */
+import type { EmbedProvider } from '../../contracts/index.js';
+import type { NormalizedMemory } from './07-memory-normalize.js';
+import type { DB } from '../../storage/db.js';
+export interface IndexResult {
+    memoryId: string;
+    created: boolean;
+}
+export interface EmbedIndexContext {
+    db: DB;
+    embed: EmbedProvider;
+    sessionId: string | null;
+    repo: string | null;
+    project: string | null;
+    branch: string | null;
+    agent: string | null;
+    sourceHash: string | null;
+}
+/**
+ * Embed and index all normalized memories.
+ * Returns one IndexResult per input atom (some may be deduped by hash).
+ */
+export declare function embedAndIndex(memories: NormalizedMemory[], ctx: EmbedIndexContext): Promise<IndexResult[]>;