@oh-my-pi/pi-coding-agent 15.10.0 → 15.10.2

package/src/tools/fetch.ts

@@ -1,4 +1,6 @@
+import { Database } from "bun:sqlite";
 import * as fs from "node:fs/promises";
+import * as os from "node:os";
 import * as path from "node:path";
 import type { AgentToolResult } from "@oh-my-pi/pi-agent-core";
 import type { ImageContent, TextContent } from "@oh-my-pi/pi-ai";
@@ -8,6 +10,7 @@ import { $which, ptree, truncate } from "@oh-my-pi/pi-utils";
 import { parseHTML } from "linkedom";
 import { LRUCache } from "lru-cache/raw";
 import type { Settings } from "../config/settings";
+import { readEditableNotebookText } from "../edit/notebook";
 import type { RenderResultOptions } from "../extensibility/custom-tools/types";
 import { type Theme, theme } from "../modes/theme/theme";
 import type { ToolSession } from "../sdk";
@@ -22,10 +25,12 @@ import { specialHandlers } from "../web/scrapers";
 import type { RenderResult } from "../web/scrapers/types";
 import { finalizeOutput, loadPage, looksLikeHtml, MAX_OUTPUT_CHARS } from "../web/scrapers/types";
 import { convertWithMarkit, fetchBinary } from "../web/scrapers/utils";
+import { type ArchiveFormat, listArchiveRoot, sniffArchiveFormat } from "./archive-reader";
 import { applyListLimit } from "./list-limit";
 import { formatStyledArtifactReference, type OutputMeta } from "./output-meta";
 import { type LineRange, parseLineRanges } from "./path-utils";
-import { formatExpandHint, getDomain, replaceTabs } from "./render-utils";
+import { formatBytes, formatExpandHint, getDomain, replaceTabs } from "./render-utils";
+import { listTables, looksLikeSqlite, renderTableList } from "./sqlite-reader";
 import { ToolAbortError, ToolError } from "./tool-errors";
 import { toolResult } from "./tool-result";
 import { clampTimeout } from "./tool-timeouts";
@@ -46,8 +51,6 @@ const CONVERTIBLE_MIMES = new Set([
 	"application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
 	"application/rtf",
 	"application/epub+zip",
-	"application/x-ipynb+json",
-	"application/zip",
 	"image/png",
 	"image/jpeg",
 	"image/gif",
@@ -67,7 +70,6 @@ const CONVERTIBLE_EXTENSIONS = new Set([
 	".xlsx",
 	".rtf",
 	".epub",
-	".ipynb",
 	".png",
 	".jpg",
 	".jpeg",
@@ -78,6 +80,27 @@ const CONVERTIBLE_EXTENSIONS = new Set([
 	".ogg",
 ]);
 
+const NOTEBOOK_MIMES = new Set(["application/x-ipynb+json"]);
+const NOTEBOOK_EXTENSIONS = new Set([".ipynb"]);
+
+const SQLITE_MIMES = new Set([
+	"application/vnd.sqlite3",
+	"application/x-sqlite3",
+	"application/sqlite3",
+	"application/sqlite",
+]);
+const SQLITE_EXTENSIONS = new Set([".sqlite", ".sqlite3", ".db", ".db3"]);
+
+const ARCHIVE_MIMES = new Set([
+	"application/zip",
+	"application/x-zip-compressed",
+	"application/x-tar",
+	"application/tar",
+	"application/gzip",
+	"application/x-gzip",
+]);
+const ARCHIVE_EXTENSIONS = new Set([".zip", ".tar", ".tar.gz", ".tgz", ".gz"]);
+
 const IMAGE_MIME_BY_EXTENSION = new Map<string, string>([
 	[".png", "image/png"],
 	[".jpg", "image/jpeg"],
@@ -261,6 +284,12 @@ function normalizeMime(contentType: string): string {
 	return contentType.split(";")[0].trim().toLowerCase();
 }
 
+function getFilenameExtensionHint(filename: string): string {
+	const lower = filename.toLowerCase();
+	if (lower.endsWith(".tar.gz")) return ".tar.gz";
+	return path.extname(filename).toLowerCase();
+}
+
 /**
  * Get extension from URL or Content-Disposition
  */
@@ -269,7 +298,7 @@ function getExtensionHint(url: string, contentDisposition?: string): string {
 	if (contentDisposition) {
 		const match = contentDisposition.match(/filename[*]?=["']?([^"';\n]+)/i);
 		if (match) {
-			const ext = path.extname(match[1]).toLowerCase();
+			const ext = getFilenameExtensionHint(match[1]);
 			if (ext) return ext;
 		}
 	}
@@ -277,7 +306,7 @@ function getExtensionHint(url: string, contentDisposition?: string): string {
 	// Fall back to URL path
 	try {
 		const pathname = new URL(url).pathname;
-		const ext = path.extname(pathname).toLowerCase();
+		const ext = getFilenameExtensionHint(pathname);
 		if (ext) return ext;
 	} catch {}
 
@@ -738,6 +767,254 @@ type FetchRenderResult = RenderResult & {
 	image?: FetchImagePayload;
 };
 
+const BINARY_SAMPLE_CHARS = 4096;
+const URL_ARCHIVE_LIST_LIMIT = 500;
+const URL_SQLITE_LIST_LIMIT = 500;
+
+function sampleLooksBinary(text: string): boolean {
+	const limit = Math.min(text.length, BINARY_SAMPLE_CHARS);
+	if (limit === 0) return false;
+
+	let replacementCount = 0;
+	for (let index = 0; index < limit; index++) {
+		const code = text.charCodeAt(index);
+		if (code === 0) return true;
+		if (code === 0xfffd) replacementCount++;
+	}
+
+	return replacementCount >= 3 && replacementCount / limit > 0.01;
+}
+
+function isNotebookHint(mime: string, extensionHint: string): boolean {
+	return NOTEBOOK_MIMES.has(mime) || NOTEBOOK_EXTENSIONS.has(extensionHint);
+}
+
+function isSqliteHint(mime: string, extensionHint: string): boolean {
+	return SQLITE_MIMES.has(mime) || SQLITE_EXTENSIONS.has(extensionHint);
+}
+
+function isArchiveHint(mime: string, extensionHint: string): boolean {
+	return ARCHIVE_MIMES.has(mime) || ARCHIVE_EXTENSIONS.has(extensionHint);
+}
+
+function getArchiveFormatHint(mime: string, extensionHint: string): ArchiveFormat | undefined {
+	if (extensionHint === ".zip" || mime === "application/zip" || mime === "application/x-zip-compressed") {
+		return "zip";
+	}
+	if (extensionHint === ".tar" || mime === "application/x-tar" || mime === "application/tar") {
+		return "tar";
+	}
+	if (
+		extensionHint === ".tar.gz" ||
+		extensionHint === ".tgz" ||
+		extensionHint === ".gz" ||
+		mime === "application/gzip" ||
+		mime === "application/x-gzip"
+	) {
+		return "tar.gz";
+	}
+	return undefined;
+}
+
+function formatErrorMessage(error: unknown): string {
+	return error instanceof Error ? error.message : String(error);
+}
+
+function binaryContentType(mime: string): string {
+	return mime || "application/octet-stream";
+}
+
+function buildBinaryNotice(finalUrl: string, mime: string, byteLength?: number): string {
+	const size = byteLength === undefined ? "unknown size" : formatBytes(byteLength);
+	return `[Binary content: ${binaryContentType(mime)}, ${size}] ${finalUrl}`;
+}
+
+function buildBinaryPayloadResult(
+	url: string,
+	finalUrl: string,
+	mime: string,
+	method: string,
+	content: string,
+	fetchedAt: string,
+	notes: string[],
+): FetchRenderResult {
+	const output = finalizeOutput(content);
+	return {
+		url,
+		finalUrl,
+		contentType: binaryContentType(mime),
+		method,
+		content: output.content,
+		fetchedAt,
+		truncated: output.truncated,
+		notes,
+	};
+}
+
+async function withTempBinaryFile<T>(
+	prefix: string,
+	extension: string,
+	bytes: Uint8Array,
+	readTempFile: (tempPath: string) => Promise<T>,
+): Promise<T> {
+	const tempDir = await fs.mkdtemp(path.join(os.tmpdir(), prefix));
+	const tempPath = path.join(tempDir, `payload${extension}`);
+	try {
+		await Bun.write(tempPath, bytes);
+		return await readTempFile(tempPath);
+	} finally {
+		await fs.rm(tempDir, { recursive: true, force: true });
+	}
+}
+
+async function renderNotebookPayload(bytes: Uint8Array, displayUrl: string): Promise<string> {
+	return withTempBinaryFile("omp-url-notebook-", ".ipynb", bytes, tempPath =>
+		readEditableNotebookText(tempPath, displayUrl),
+	);
+}
+
+async function renderSqlitePayload(bytes: Uint8Array): Promise<string> {
+	return withTempBinaryFile("omp-url-sqlite-", ".sqlite", bytes, async tempPath => {
+		let db: Database | null = null;
+		try {
+			db = new Database(tempPath, { readonly: true, strict: true });
+			db.run("PRAGMA busy_timeout = 3000");
+			const listLimit = applyListLimit(listTables(db), { limit: URL_SQLITE_LIST_LIMIT });
+			return renderTableList(listLimit.items);
+		} finally {
+			db?.close();
+		}
+	});
+}
+
+async function tryRenderBinaryPayload(
+	url: string,
+	finalUrl: string,
+	mime: string,
+	extHint: string,
+	rawContent: string,
+	timeout: number,
+	signal: AbortSignal | undefined,
+	fetchedAt: string,
+	notes: readonly string[],
+): Promise<FetchRenderResult | null> {
+	const hasNotebookHint = isNotebookHint(mime, extHint);
+	const hasSqliteHint = isSqliteHint(mime, extHint);
+	const hasArchiveHint = isArchiveHint(mime, extHint);
+	const rawLooksBinary = sampleLooksBinary(rawContent);
+	if (!hasNotebookHint && !hasSqliteHint && !hasArchiveHint && !rawLooksBinary) {
+		return null;
+	}
+
+	const resultNotes = [...notes];
+	const binary = await fetchBinary(finalUrl, timeout, signal);
+	if (!binary.ok) {
+		resultNotes.push(binary.error ? `Binary fetch failed: ${binary.error}` : "Binary fetch failed");
+		return buildBinaryPayloadResult(
+			url,
+			finalUrl,
+			mime,
+			"binary",
+			buildBinaryNotice(finalUrl, mime),
+			fetchedAt,
+			resultNotes,
+		);
+	}
+
+	const binaryExtHint = getExtensionHint(finalUrl, binary.contentDisposition) || extHint;
+	if (isNotebookHint(mime, binaryExtHint)) {
+		try {
+			return buildBinaryPayloadResult(
+				url,
+				finalUrl,
+				mime,
+				"notebook",
+				await renderNotebookPayload(binary.buffer, finalUrl),
+				fetchedAt,
+				resultNotes,
+			);
+		} catch (error) {
+			resultNotes.push(`Notebook rendering failed: ${formatErrorMessage(error)}`);
+			return buildBinaryPayloadResult(
+				url,
+				finalUrl,
+				mime,
+				"binary",
+				buildBinaryNotice(finalUrl, mime, binary.buffer.byteLength),
+				fetchedAt,
+				resultNotes,
+			);
+		}
+	}
+
+	if (isSqliteHint(mime, binaryExtHint) || looksLikeSqlite(binary.buffer)) {
+		try {
+			return buildBinaryPayloadResult(
+				url,
+				finalUrl,
+				mime,
+				"sqlite",
+				await renderSqlitePayload(binary.buffer),
+				fetchedAt,
+				resultNotes,
+			);
+		} catch (error) {
+			resultNotes.push(`SQLite rendering failed: ${formatErrorMessage(error)}`);
+			return buildBinaryPayloadResult(
+				url,
+				finalUrl,
+				mime,
+				"binary",
+				buildBinaryNotice(finalUrl, mime, binary.buffer.byteLength),
+				fetchedAt,
+				resultNotes,
+			);
+		}
+	}
+
+	const hintedArchiveFormat = getArchiveFormatHint(mime, binaryExtHint);
+	const shouldArchiveSniff = hintedArchiveFormat !== undefined || !isConvertible(mime, binaryExtHint);
+	const archiveFormat = hintedArchiveFormat ?? (shouldArchiveSniff ? sniffArchiveFormat(binary.buffer) : undefined);
+	if (archiveFormat) {
+		try {
+			return buildBinaryPayloadResult(
+				url,
+				finalUrl,
+				mime,
+				"archive",
+				await listArchiveRoot(binary.buffer, archiveFormat, { limit: URL_ARCHIVE_LIST_LIMIT }),
+				fetchedAt,
+				resultNotes,
+			);
+		} catch (error) {
+			resultNotes.push(`Archive rendering failed: ${formatErrorMessage(error)}`);
+			return buildBinaryPayloadResult(
+				url,
+				finalUrl,
+				mime,
+				"binary",
+				buildBinaryNotice(finalUrl, mime, binary.buffer.byteLength),
+				fetchedAt,
+				resultNotes,
+			);
+		}
+	}
+
+	if (rawLooksBinary) {
+		return buildBinaryPayloadResult(
+			url,
+			finalUrl,
+			mime,
+			"binary",
+			buildBinaryNotice(finalUrl, mime, binary.buffer.byteLength),
+			fetchedAt,
+			resultNotes,
+		);
+	}
+
+	return null;
+}
+
 // =============================================================================
 // Unified Special Handler Dispatch
 // =============================================================================
@@ -984,6 +1261,19 @@ async function renderUrl(
 		}
 	}
 
+	const binaryPayloadResult = await tryRenderBinaryPayload(
+		url,
+		finalUrl,
+		mime,
+		extHint,
+		rawContent,
+		timeout,
+		signal,
+		fetchedAt,
+		notes,
+	);
+	if (binaryPayloadResult) return binaryPayloadResult;
+
 	// Step 4: Handle non-HTML text content
 	const isHtml = mime.includes("html") || mime.includes("xhtml");
 	const isJson = mime.includes("json");
@@ -992,7 +1282,7 @@ async function renderUrl(
 	const isFeed = mime.includes("rss") || mime.includes("atom") || mime.includes("feed");
 
 	// Raw mode skips every text-shaping branch below (JSON pretty-print, feed-to-markdown,
-	// HTML extraction) and returns the response body verbatim. The image/markit branches
+	// HTML extraction) and returns the response body verbatim. Binary-oriented branches
 	// above already ran because raw isn't useful for binary payloads.
 	if (raw) {
 		const output = finalizeOutput(rawContent);

package/src/tools/find.ts

@@ -13,6 +13,7 @@ import findDescription from "../prompts/tools/find.md" with { type: "text" };
 import { type TruncationResult, truncateHead } from "../session/streaming-output";
 import { Ellipsis, fileHyperlink, renderFileList, renderStatusLine, renderTreeList, truncateToWidth } from "../tui";
 import type { ToolSession } from ".";
+import { buildPathTree, walkPathTree } from "./grouped-file-output";
 import { applyListLimit } from "./list-limit";
 import { formatFullOutputReference, type OutputMeta } from "./output-meta";
 import {
@@ -54,34 +55,27 @@ const MIN_GLOB_TIMEOUT_MS = 500;
 const MAX_GLOB_TIMEOUT_MS = 60_000;
 
 /**
- * Group find matches by their directory so the model doesn't pay repeated
- * tokens for shared path prefixes. Preserves the input order: groups appear in
- * the order their first member was emitted (mtime-desc for native glob), and
- * within a group entries keep their relative order.
+ * Group find matches into a multi-level directory tree so the model doesn't pay
+ * repeated tokens for shared path prefixes. Single-child directory chains fold
+ * into one header (`# a/b/c/`), so a common prefix — including an absolute root
+ * for out-of-cwd results — collapses to a single line. Each level adds one `#`;
+ * files are listed bare under the deepest directory header that owns them.
+ *
+ * Order follows the input (mtime-desc for native glob): a directory appears when
+ * its first member is emitted, and a node's own files precede its subdirectories.
  */
 export function formatFindGroupedOutput(paths: readonly string[]): string {
 	if (paths.length === 0) return "";
-	const groups = new Map<string, string[]>();
-	for (const entry of paths) {
-		const hasTrailingSlash = entry.endsWith("/");
-		const trimmed = hasTrailingSlash ? entry.slice(0, -1) : entry;
-		const slash = trimmed.lastIndexOf("/");
-		const dir = slash === -1 ? "" : trimmed.slice(0, slash);
-		const base = slash === -1 ? trimmed : trimmed.slice(slash + 1);
-		const label = hasTrailingSlash ? `${base}/` : base;
-		const list = groups.get(dir);
-		if (list) list.push(label);
-		else groups.set(dir, [label]);
-	}
-	const sections: string[] = [];
-	for (const [dir, entries] of groups) {
-		if (dir === "") {
-			sections.push(entries.join("\n"));
+	const tree = buildPathTree(paths.map(entry => ({ path: entry, isDir: entry.endsWith("/") })));
+	const lines: string[] = [];
+	for (const event of walkPathTree(tree)) {
+		if (event.kind === "dir") {
+			lines.push(`${"#".repeat(event.depth + 1)} ${event.name}/`);
 		} else {
-			sections.push(`# ${dir}/\n${entries.join("\n")}`);
+			lines.push(event.name);
 		}
 	}
-	return sections.join("\n\n");
+	return lines.join("\n");
 }
 
 export interface FindToolDetails {
@@ -359,6 +353,13 @@ export class FindTool implements AgentTool<typeof findSchema, FindToolDetails> {
 							maxResults: effectiveLimit,
 							sortByMtime: true,
 							gitignore: useGitignore,
+							// parseFindPattern explicitly prepends "**/" when the user's
+							// pattern begins with a glob (so `*.ts` becomes `**/*.ts`).
+							// Anything that arrives here without "**/" was scoped to a
+							// single directory by the user (e.g. `dir/*`); disable the
+							// native auto-recursion so `dir/*` does not silently match
+							// `dir/sub/nested.ts`.
+							recursive: false,
 							signal: combinedSignal,
 						},
 						onMatch,
@@ -434,6 +435,10 @@ function formatFindRenderPaths(paths: FindRenderArgs["paths"]): string | undefin
 
 const COLLAPSED_LIST_LIMIT = PREVIEW_LIMITS.COLLAPSED_ITEMS;
 
+function findStatusIcon(uiTheme: Theme): string {
+	return uiTheme.fg("toolTitle", uiTheme.symbol("icon.search"));
+}
+
 export const findToolRenderer = {
 	inline: true,
 	renderCall(args: FindRenderArgs, _options: RenderResultOptions, uiTheme: Theme): Component {
@@ -441,10 +446,16 @@ export const findToolRenderer = {
 		if (args.limit !== undefined) meta.push(`limit:${args.limit}`);
 
 		const text = renderStatusLine(
-			{ icon: "pending", title: "Find", description: formatFindRenderPaths(args.paths) || "*", meta },
+			{
+				icon: "pending",
+				title: "Find",
+				titleColor: "toolTitle",
+				description: formatFindRenderPaths(args.paths) || "*",
+				meta,
+			},
 			uiTheme,
 		);
-		return new Text(text, 0, 0);
+		return new Text(text, 1, 0);
 	},
 
 	renderResult(
@@ -457,7 +468,7 @@ export const findToolRenderer = {
 
 		if (result.isError || details?.error) {
 			const errorText = details?.error || result.content?.find(c => c.type === "text")?.text || "Unknown error";
-			return new Text(formatErrorMessage(errorText, uiTheme), 0, 0);
+			return new Text(formatErrorMessage(errorText, uiTheme), 1, 0);
 		}
 
 		const hasDetailedData = details?.fileCount !== undefined;
@@ -470,14 +481,15 @@ export const findToolRenderer = {
 				textContent.includes("No files found") ||
 				textContent.trim() === ""
 			) {
-				return new Text(formatEmptyMessage("No files found", uiTheme), 0, 0);
+				return new Text(formatEmptyMessage("No files found", uiTheme), 1, 0);
 			}
 
 			const lines = textContent.split("\n").filter(l => l.trim());
 			const header = renderStatusLine(
 				{
-					icon: "success",
+					iconOverride: findStatusIcon(uiTheme),
 					title: "Find",
+					titleColor: "toolTitle",
 					description: formatFindRenderPaths(args?.paths),
 					meta: [formatCount("file", lines.length)],
 				},
@@ -498,6 +510,7 @@ export const findToolRenderer = {
 					);
 					return [header, ...listLines].map(l => truncateToWidth(l, width, Ellipsis.Omit));
 				},
+				{ paddingX: 1 },
 			);
 		}
 
@@ -513,20 +526,27 @@ export const findToolRenderer = {
 
 		if (fileCount === 0) {
 			const header = renderStatusLine(
-				{ icon: "warning", title: "Find", description: formatFindRenderPaths(args?.paths), meta: ["0 files"] },
+				{
+					icon: "warning",
+					title: "Find",
+					titleColor: "toolTitle",
+					description: formatFindRenderPaths(args?.paths),
+					meta: ["0 files"],
+				},
 				uiTheme,
 			);
 			const lines = [header, formatEmptyMessage("No files found", uiTheme)];
 			if (missingNote) lines.push(missingNote);
-			return new Text(lines.join("\n"), 0, 0);
+			return new Text(lines.join("\n"), 1, 0);
 		}
 		const meta: string[] = [formatCount("file", fileCount)];
 		if (details?.scopePath) meta.push(`in ${details.scopePath}`);
 		if (truncated) meta.push(uiTheme.fg("warning", "truncated"));
 		const header = renderStatusLine(
 			{
-				icon: truncated ? "warning" : "success",
+				...(truncated ? { icon: "warning" as const } : { iconOverride: findStatusIcon(uiTheme) }),
 				title: "Find",
+				titleColor: "toolTitle",
 				description: formatFindRenderPaths(args?.paths),
 				meta,
 			},
@@ -565,6 +585,7 @@ export const findToolRenderer = {
 				);
 				return [header, ...fileLines, ...extraLines].map(l => truncateToWidth(l, width, Ellipsis.Omit));
 			},
+			{ paddingX: 1 },
 		);
 	},
 	mergeCallAndResult: true,