pi-read-chunks 2.0.0 → 2.0.2

package/README.md

@@ -19,11 +19,11 @@ Replaces the built-in `read()` for text files via `registerTool`. Images and oth
 
 **Context-aware chunk labels** — Chunk labels are line ranges (`start-end`) with char offsets in parentheses, derived by binary-searching the file's `\n` positions. Labels match how the agent reasons (line numbers, not raw char offsets).
 
-**Per-tool-result compression** — The raw chunked-mode payload is JSON; a `tool_result` hook rewrites it into a compact, line-budgeted summary (~25 chunk entries × 120 chars each) before the model sees it, so a multi-chunk scan doesn't blow the context budget. Toggle with `/read-chunks`.
+**Per-tool-result compression** — The raw chunked-mode payload is JSON; a `tool_result` hook rewrites it into a human-readable summary before the model sees it, so a multi-chunk scan doesn't blow the context budget. Toggle with `/read-chunks`.
 
 **Optional per-invocation debug dump** — `/read-chunks debug` (toggles on/off) writes each invocation's LLM requests/responses and the final tool return to `/tmp/read-chunks_<YYMMDD-hhmmss>.json`. Disabled by default.
 
-**Summary budget** — `thresholdKB` also acts as a soft cap on total summary size. Each chunk's LLM prompt includes a budget hint: `thresholdKB / numChunks` KB per chunk. For a 160KB file with 50KB threshold (4 chunks), each chunk gets ~12KB of summary budget. A 500KB file (10 chunks) → 5KB per chunk. This keeps the total summary proportional to `thresholdKB` regardless of file size, preventing small files from wasting context while still allowing large files enough room for useful summaries.
+**Summary budget** — By default, summaries preserve names, places, events, dates, and key details without artificial truncation. Typical output stays under ~1.5KB for a 100KB file (~94% reduction from the original). If summaries exceed `thresholdKB`, they are sent back to the LLM for denser compression (up to 3 attempts) without losing factual content.
 
 
 ## Installation
@@ -67,7 +67,7 @@ Copy read-chunks.example.json to:
 
 | Key                 | Default | Notes                                                                                          |
 | ------------------- | ------- | ---------------------------------------------------------------------------------------------- |
-| `thresholdKB`       | `50`    | Files ≤ this size are returned verbatim; larger files are chunked using this same value as the target chunk size (KB). Also serves as the soft limit for total summary size — the per-chunk summary budget is `thresholdKB / numChunks`, so the total stays bounded to roughly one chunk's worth. |
+| `thresholdKB`       | `50`    | Files ≤ this size are returned verbatim; larger files are chunked using this same value as the target chunk size (KB). Also acts as the hard cap on total summary size — summaries exceeding this are sent back to the LLM for denser compression without losing factual content. |
 | `chunkOverlapChars` | `800`   | Backward overlap between consecutive chunks, in characters.                                    |
 
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pi-read-chunks",
-  "version": "2.0.0",
+  "version": "2.0.2",
   "description": "A Pi Agent extension that uses a subagent to enhance `read` functionality for large text files to reduce context bloat, rot and lost in the middle issues.",
   "keywords": [
     "pi-package",

package/read-chunks.ts

@@ -462,6 +462,45 @@ function formatReadChunksCall(args: any, theme: any, cwd: string): string {
 	return text;
 }
 
+// ---------- Summary compression ----------
+
+async function compressSummary(
+	text: string,
+	modelRegistry: any,
+	model: any,
+): Promise<string> {
+	if (!model) return text;
+	try {
+		if (!modelRegistry.hasConfiguredAuth(model)) return text;
+	} catch {
+		return text;
+	}
+
+	const prompt = `You are compressing a detailed summary into a denser, more concise version. Keep ALL factual content — names, places, events, dates, numbers, key details. Remove filler words, redundant phrases, unnecessary articles, and verbose constructions. Write in a telegraphic, information-dense style. Do not omit or dilute any facts.
+
+Compressed summary:
+"""
+${text}
+"""
+
+Compressed:`;
+
+	const requestPayload = {
+		model: model.id,
+		messages: [{ role: "user", content: [{ type: "text", text: prompt }] }],
+	};
+
+	try {
+		const response = await modelRegistry.complete(model, requestPayload);
+		const contentBlocks = response?.content || [];
+		const textBlocks = contentBlocks.filter((c: any) => c.type === "text");
+		if (textBlocks.length === 0) return text;
+		return textBlocks.map((c: any) => c.text).join(" ").trim();
+	} catch {
+		return text;
+	}
+}
+
 // ---------- Summary builder ----------
 
 function buildSummary(parsed: any): string {
@@ -721,6 +760,7 @@ export default function (pi: ExtensionAPI) {
 				mode: query ? "query" : "summary",
 				file: absolutePath,
 				kb_total: totalKB,
+				thresholdKB: config.thresholdKB,
 				chunks_scanned: chunks.length,
 				chunks_read: perChunk.length,
 				stop_reason: stopReason,
@@ -774,8 +814,21 @@ export default function (pi: ExtensionAPI) {
 			return;
 		}
 
+		let summaryText = buildSummary(parsed);
+
+		// Post-compression if summary exceeds thresholdKB
+		const model = _ctx.model;
+		const maxKB = parsed.thresholdKB;
+		if (model && maxKB) {
+			const summaryBytes = Buffer.byteLength(summaryText, "utf-8");
+			const summaryKB = summaryBytes / 1024;
+			if (summaryKB > maxKB) {
+				summaryText = await compressSummary(summaryText, _ctx.modelRegistry, model);
+			}
+		}
+
 		return {
-			content: [{ type: "text", text: buildSummary(parsed) }],
+			content: [{ type: "text", text: summaryText }],
 			details: event.details,
 		};
 	});