pi-blackhole 0.3.4 → 0.3.7

package/src/om/runtime.ts

@@ -10,7 +10,7 @@
  * - markConsolidationError sets 30s retry gate for failed runs.
  */
 import { type Config, type ConfiguredModel, DEFAULTS, loadConfig } from "./config.js";
-import { isCooldownActive, recordCooldown, expireCooldowns, modelKey } from "./cooldown.js";
+import { isCooldownActive, getCooldownEntry, recordCooldown, expireCooldowns, modelKey } from "./cooldown.js";
 
 export type ResolveResult =
 	| { ok: true; model: any; apiKey: string; headers?: Record<string, string>; cooldownApplied?: boolean }
@@ -94,13 +94,17 @@ export class Runtime {
 	 * Tries the candidate list in order:
 	 * 1. Primary stage model → 2. Stage fallbacks → 3. Base config.model → 4. Session model.
 	 *
+	 * Session model fallback can be disabled via config.sessionFallback: false.
+	 * When disabled, returns { ok: false } instead of using the session model,
+	 * allowing the stage to be skipped entirely when all configured OM models fail.
+	 *
 	 * Skips models that are currently in a cooldown window.
 	 * On retryable error (after the agent runs), the model that failed is cooled down
 	 * and the next candidate is tried.  The caller must call `recordRetryableError`
 	 * after the API attempt to mark the failed model.
 	 *
 	 * Returns `ok: true` with the resolved model, or `ok: false` with a reason
-	 * if all candidates (including session model) are exhausted or unavailable.
+	 * if all candidates (including session model, if enabled) are exhausted or unavailable.
 	 */
 	async resolveModel(ctx: ResolveCtx): Promise<ResolveResult> {
 		const candidates = this.buildCandidateList(ctx.stageModel, ctx.stageFallbacks);
@@ -123,8 +127,10 @@ export class Runtime {
 
 			if (isCooldownActive(candidate)) {
 				if (ctx.hasUI && ctx.ui) {
+					const entry = getCooldownEntry(candidate);
+					const reason = entry ? `: ${entry.reason}` : "";
 					ctx.ui.notify(
-						`Observational memory: ${stageName} skipping ${key} (cooldown active)`,
+						`Observational memory: ${stageName} skipping ${key} (cooldown${reason})`,
 						"info",
 					);
 				}
@@ -162,25 +168,40 @@ export class Runtime {
 			};
 		}
 
-		// Fall back to session model
-		const sessionModel = ctx.model;
-		if (!sessionModel) {
-			return { ok: false, reason: `no model available for ${stageName} (all candidates exhausted, no session model)` };
+		// Fall back to session model (if enabled)
+		if (this.config.sessionFallback !== false) {
+			const sessionModel = ctx.model;
+			if (!sessionModel) {
+				return { ok: false, reason: `no model available for ${stageName} (all candidates exhausted, no session model)` };
+			}
+
+			const auth = await ctx.modelRegistry.getApiKeyAndHeaders(sessionModel);
+			if (!auth.ok || !auth.apiKey) {
+				const provider = (sessionModel as { provider?: string }).provider ?? "unknown";
+				return { ok: false, reason: `no API key for session model provider "${provider}"` };
+			}
+
+			return {
+				ok: true,
+				model: sessionModel,
+				apiKey: auth.apiKey as string,
+				headers: auth.headers as Record<string, string> | undefined,
+				cooldownApplied: false,
+			};
 		}
 
-		const auth = await ctx.modelRegistry.getApiKeyAndHeaders(sessionModel);
-		if (!auth.ok || !auth.apiKey) {
-			const provider = (sessionModel as { provider?: string }).provider ?? "unknown";
-			return { ok: false, reason: `no API key for session model provider "${provider}"` };
+		// All configured candidates exhausted and session fallback disabled —
+		// skip the stage entirely.  Info-level to match cooldown-disabled pattern.
+		// Set resolveFailureNotified so the consolidation layer doesn't duplicate.
+		if (ctx.hasUI && ctx.ui) {
+			ctx.ui.notify(
+				`Observational memory: ${stageName} skipped — all candidates failed (sessionFallback disabled, won't use main model)`,
+				"info",
+			);
 		}
+		this.resolveFailureNotified = true;
 
-		return {
-			ok: true,
-			model: sessionModel,
-			apiKey: auth.apiKey as string,
-			headers: auth.headers as Record<string, string> | undefined,
-			cooldownApplied: false,
-		};
+		return { ok: false, reason: `no model available for ${stageName} (all candidates exhausted, sessionFallback disabled)` };
 	}
 
 	/**

package/src/sections.ts

@@ -4,7 +4,6 @@
  * Upstream: https://github.com/sting8k/pi-vcc (src/sections.ts)
  * Unmodified.
  */
-import type { TranscriptEntry } from "./core/brief";
 
 export interface SectionData {
   sessionGoal: string[];
@@ -13,6 +12,4 @@ export interface SectionData {
   commits: string[];
   userPreferences: string[];
   briefTranscript: string;
-  /** Structured transcript entries (verbose object format) */
-  transcriptEntries: TranscriptEntry[];
 }

package/src/tools/recall.ts

@@ -138,7 +138,11 @@ async function vccRecall(params: { query?: string; expand?: number[]; page?: num
 
 	// Merge expanded entries into full result set BEFORE pagination
 	// so pagination counts and positioning stay consistent
+	let appendedExpandCount = 0;
 	if (expandedFullEntries) {
+		// Track expand-only entries that don't match the query so the header can distinguish them
+		const existingIndices = new Set(allResults.map((r) => r.index));
+		appendedExpandCount = expandedFullEntries.filter((fe) => !existingIndices.has(fe.index)).length;
 		allResults = mergeExpandedIntoSearchResults(allResults, expandedFullEntries);
 	}
 
@@ -148,9 +152,10 @@ async function vccRecall(params: { query?: string; expand?: number[]; page?: num
 		const pageResults: SearchHit[] = allResults.slice(start, start + PAGE_SIZE);
 		const totalPages = Math.ceil(allResults.length / PAGE_SIZE);
 		const scopeSuffix = scope === "all" ? " (scope: all)" : "";
+		const matchCount = allResults.length - appendedExpandCount;
 		const header = totalPages > 1
-			? `Page ${page}/${totalPages} (${allResults.length} total matches${scopeSuffix})`
-			: `${allResults.length} matches${scopeSuffix}`;
+			? `Page ${page}/${totalPages} (${matchCount} matches${appendedExpandCount > 0 ? ` + ${appendedExpandCount} expanded` : ""}${scopeSuffix})`
+			: `${matchCount} matches${appendedExpandCount > 0 ? ` (+ ${appendedExpandCount} expanded)` : ""}${scopeSuffix}`;
 		const footer = page < totalPages
 			? `\n--- Use page:${page + 1}${scope === "all" ? " with scope:'all'" : ""} for more results ---`
 			: "";
@@ -233,13 +238,10 @@ export function registerRecallTool(pi: ExtensionAPI): void {
 			"Search session history and earlier lines omitted, file write/edit content by text/regex. " +
 			"Expand entries (#N), drill-down file content (#N:path) with paging, or aggregate touched files (mode:touched).",
 		promptSnippet:
-			"recall: Search session history + file write/edit content by text/regex. #N expand, #N:path drill-down with offset/limit paging, mode:file/transcript/touched.",
+			"recall: Search session history + file write/edit content by text/regex. #N expand, #N:path drill-down with optional :offset:limit or :full, mode:file/touched.",
 		promptGuidelines: [
-			"Use recall — search is literal text/regex matching, NOT semantic. If no results, try different terms or a regex pattern. Set scope:'all' to search the full session.",
-			"Use recall — mode:file to search only write/edit file content, mode:transcript for conversation-only, mode:touched for aggregate view of all files written/edited across the session.",
-			"Use recall — drill-down supports paging: #42:auth.ts shows first 30 lines, #42:auth.ts:30 shows next 30, #42:auth.ts:full shows everything. Note: edit diffs are not indexed for text search — drill-down reads them from raw JSONL.",
-			"Use recall — when a drill-down path matches multiple files, options are listed. Narrow with a more specific path substring.",
-			"Use recall — hex observation/reflection ids (12-char hex) link memory evidence to session entries with cross-references for navigation.",
+			"Use recall — literal text/regex search across session history and file write/edit content. #N expands an entry; #N:path with optional :offset:limit or :full drills down into file content; 12-char hex ids recover observation/reflection sources. mode:file for file-content-only, mode:touched for aggregated files-by-path. scope:'all' to search the full session. If no results, try fewer terms or a regex pattern.",
+			"Use recall — when a drill-down path matches multiple files, options are listed. Narrow with a more specific path substring. Only full-file writes are indexed for text search (edit diffs are not).",
 		],
 		parameters: Type.Object({
 			query: Type.Optional(
@@ -255,7 +257,7 @@ export function registerRecallTool(pi: ExtensionAPI): void {
 				StringEnum(["lineage", "all"] as const, { description: "Search scope. lineage = active lineage (default), all = entire session." }),
 			),
 			mode: Type.Optional(
-				StringEnum(["hybrid", "file", "transcript", "touched"] as const, { description: "What content to search. hybrid (default) = transcript + file indicators. file = write/edit file content only. transcript = conversation only. touched = aggregate files grouped by path (not per-entry search)." }),
+				StringEnum(["hybrid", "file", "touched"] as const, { description: "What content to search. hybrid (default) = all session content. file = file content only. touched = files-by-path summary with entry indices." }),
 			),
 		}),
 		async execute(_toolCallId, params, _signal, _onUpdate, ctx) {