@oai404iao/pi-codex-web-search 0.1.0-alpha.1

package/src/tools/web-search/activity.ts

@@ -0,0 +1,195 @@
+import { glyphs } from "@oai404iao/pi-codex-runtime/internal/glyphs";
+import { type StreamEventShape } from "@oai404iao/pi-codex-runtime/internal/providers/openai-codex/types";
+
+export const WEB_SEARCH_ACTIVITY_MESSAGE_TYPE = "codex-web-search-activity";
+
+export interface SurfacedWebSearch {
+	callId: string;
+	status?: string;
+	completed?: boolean;
+	actionType?: string;
+	query?: string;
+	queries: string[];
+	url?: string;
+	pattern?: string;
+	sources: Array<{ title?: string; url: string }>;
+	responseItem?: Record<string, unknown>;
+}
+
+export function extractWebSearch(
+	item: StreamEventShape["item"],
+	options?: { completed?: boolean },
+): SurfacedWebSearch | undefined {
+	if (!item || item.type !== "web_search_call") return undefined;
+	const callId = typeof item.id === "string" ? item.id : typeof item.call_id === "string" ? item.call_id : undefined;
+	if (!callId) return undefined;
+
+	const action = typeof item.action === "object" && item.action !== null ? (item.action as Record<string, unknown>) : undefined;
+	const actionType = typeof action?.type === "string" ? action.type : undefined;
+	const query = typeof action?.query === "string" ? action.query : typeof item.query === "string" ? item.query : undefined;
+	const queries = [
+		...(Array.isArray(action?.queries) ? action.queries : []),
+		...(Array.isArray(item.queries) ? item.queries : []),
+	].filter((value): value is string => typeof value === "string" && value.trim().length > 0);
+	const url = typeof action?.url === "string" && action.url.trim()
+		? action.url.trim()
+		: typeof item.url === "string" && item.url.trim()
+			? item.url.trim()
+			: undefined;
+	const pattern = typeof action?.pattern === "string" && action.pattern.trim() ? action.pattern.trim() : undefined;
+
+	const asRecordArray = (value: unknown): Record<string, unknown>[] => Array.isArray(value)
+		? value
+				.map((entry) => typeof entry === "object" && entry !== null ? entry as Record<string, unknown> : undefined)
+				.filter((entry): entry is Record<string, unknown> => !!entry)
+		: [];
+	const sourceCandidates = [
+		...asRecordArray(action?.sources),
+		...asRecordArray(action?.results),
+		...asRecordArray(item.results),
+	];
+	if (typeof item.url === "string") sourceCandidates.push(item as Record<string, unknown>);
+
+	const seenUrls = new Set<string>();
+	const sources: Array<{ title?: string; url: string }> = [];
+	for (const source of sourceCandidates) {
+		const url = typeof source.url === "string" && source.url.trim() ? source.url.trim() : undefined;
+		if (!url || seenUrls.has(url)) continue;
+		seenUrls.add(url);
+		const title = typeof source.title === "string" && source.title.trim() ? source.title.trim() : undefined;
+		sources.push({ ...(title ? { title } : {}), url });
+	}
+
+	return {
+		callId,
+		...(typeof item.status === "string" ? { status: item.status } : {}),
+		...(options?.completed !== undefined ? { completed: options.completed } : {}),
+		...(actionType ? { actionType } : {}),
+		...(query ? { query } : {}),
+		queries,
+		...(url ? { url } : {}),
+		...(pattern ? { pattern } : {}),
+		sources,
+		...(options?.completed ? { responseItem: item as Record<string, unknown> } : {}),
+	};
+}
+
+export function extractWebSearchProgress(event: StreamEventShape): SurfacedWebSearch | undefined {
+	const status = event.type === "response.web_search_call.in_progress"
+		? "in_progress"
+		: event.type === "response.web_search_call.searching"
+			? "searching"
+			: event.type === "response.web_search_call.completed"
+				? "completed"
+				: undefined;
+	if (!status || typeof event.item_id !== "string" || !event.item_id) return undefined;
+	return {
+		callId: event.item_id,
+		status,
+		completed: status === "completed",
+		queries: [],
+		sources: [],
+	};
+}
+
+export function mergeWebSearchActivity(
+	previous: SurfacedWebSearch | undefined,
+	next: SurfacedWebSearch,
+): SurfacedWebSearch {
+	if (!previous) return next;
+	const seenUrls = new Set<string>();
+	const sources = [...next.sources, ...previous.sources].filter((source) => {
+		if (seenUrls.has(source.url)) return false;
+		seenUrls.add(source.url);
+		return true;
+	});
+	const completed = Boolean(previous.completed || next.completed);
+	return {
+		callId: next.callId,
+		status: previous.completed && !next.completed ? previous.status : (next.status ?? previous.status),
+		completed,
+		actionType: next.actionType ?? previous.actionType,
+		query: next.query ?? previous.query,
+		queries: next.queries.length > 0 ? next.queries : previous.queries,
+		url: next.url ?? previous.url,
+		pattern: next.pattern ?? previous.pattern,
+		sources,
+		responseItem: next.responseItem ?? previous.responseItem,
+	};
+}
+
+export function webSearchActivityDetail(search: SurfacedWebSearch): string {
+	if (search.actionType === "open_page") return search.url ?? "";
+	if (search.actionType === "find_in_page") {
+		if (search.pattern && search.url) return `'${search.pattern}' in ${search.url}`;
+		if (search.pattern) return `'${search.pattern}'`;
+		return search.url ?? "";
+	}
+	const query = search.query?.trim();
+	if (query) return query;
+	const first = search.queries[0]?.trim() ?? "";
+	return search.queries.length > 1 && first ? `${first} ...` : first;
+}
+
+export function webSearchActivityHosts(search: SurfacedWebSearch): string[] {
+	const seen = new Set<string>();
+	const hosts: string[] = [];
+	for (const source of search.sources) {
+		try {
+			const host = new URL(source.url).hostname.replace(/^www\./i, "");
+			const key = host.toLowerCase();
+			if (!host || seen.has(key)) continue;
+			seen.add(key);
+			hosts.push(host);
+		} catch {
+			// Sources without a valid URL do not produce a host tag.
+		}
+	}
+	return hosts;
+}
+
+export function buildWebSearchStatusText(search: SurfacedWebSearch): string {
+	const completed = search.completed ?? search.status === "completed";
+	const detail = webSearchActivityDetail(search);
+	if (completed) return `Searched the web${detail ? ` for ${detail}` : ""}`;
+	return `Searching the web${detail ? ` ${detail}` : ""}`;
+}
+
+export function buildWebSearchInlineText(search: SurfacedWebSearch, cwd?: string): string {
+	const completed = search.completed ?? search.status === "completed";
+	const header = completed ? "Searched the web" : "Searching the web";
+	const detail = webSearchActivityDetail(search);
+	const separator = detail ? (completed ? " for " : " ") : "";
+	return `${glyphs(cwd).bullet}**${header}**${separator}${detail}`;
+}
+
+export function buildWebSearchActivityMessage(searches: SurfacedWebSearch[]): string {
+	const sections = searches.map((search, index) => {
+		const heading = searches.length > 1
+			? `${index + 1}. ${buildWebSearchStatusText(search)}`
+			: buildWebSearchStatusText(search);
+		const lines = [heading, `Call: ${search.callId}${search.status ? ` (${search.status})` : ""}`];
+		const queries = search.queries.length > 0 ? search.queries : search.query ? [search.query] : [];
+		if (queries.length > 0) {
+			lines.push(`Query: ${queries.join(" | ")}`);
+		}
+		if (search.sources.length > 0) {
+			lines.push("Sources:");
+			for (const source of search.sources.slice(0, 8)) {
+				lines.push(`- ${source.title ? `${source.title}: ` : ""}${source.url}`);
+			}
+		}
+		return lines.join("\n");
+	});
+
+	return sections.join("\n\n");
+}
+
+export function buildWebSearchSummaryText(searches: SurfacedWebSearch[]): string {
+	if (searches.length === 0) return "Web search";
+	if (searches.length === 1) return buildWebSearchStatusText(searches[0]!);
+	const completed = searches.filter((search) => search.completed ?? search.status === "completed").length;
+	if (completed === searches.length) return `Searched the web ${searches.length} times`;
+	if (completed === 0) return `Searching the web (${searches.length} calls)`;
+	return `Web search activity (${completed}/${searches.length} completed)`;
+}

package/src/tools/web-search/capture.ts

@@ -0,0 +1,44 @@
+import type { AssistantMessage } from "@earendil-works/pi-ai/compat";
+import type { ProviderEventContext } from "@oai404iao/pi-codex-runtime/internal/providers/openai-codex/stream-effects";
+import type { StreamEventShape } from "@oai404iao/pi-codex-runtime/internal/providers/openai-codex/types";
+import { encodeWebSearchActivityTextSignature } from "@oai404iao/pi-codex-runtime/internal/providers/responses/signatures";
+import {
+	buildWebSearchInlineText, extractWebSearch, extractWebSearchProgress,
+	mergeWebSearchActivity, type SurfacedWebSearch,
+} from "./activity.js";
+
+export function createWebSearchCapture({ cwd, output, stream }: ProviderEventContext) {
+	type TextBlock = Extract<AssistantMessage["content"][number], { type: "text" }>;
+	const states = new Map<string, { search: SurfacedWebSearch; block: TextBlock; contentIndex: number }>();
+	const updateActivity = (search: SurfacedWebSearch) => {
+		const existing = states.get(search.callId);
+		const merged = mergeWebSearchActivity(existing?.search, search);
+		const text = buildWebSearchInlineText(merged, cwd);
+		const textSignature = encodeWebSearchActivityTextSignature(merged.callId, merged.responseItem);
+		if (existing) {
+			existing.search = merged;
+			existing.block.text = text;
+			existing.block.textSignature = textSignature;
+			stream.push({ type: "text_delta", contentIndex: existing.contentIndex, delta: "", partial: output });
+			return;
+		}
+		const block: TextBlock = { type: "text", text: "", textSignature };
+		output.content.push(block);
+		const contentIndex = output.content.length - 1;
+		states.set(search.callId, { search: merged, block, contentIndex });
+		stream.push({ type: "text_start", contentIndex, partial: output });
+		block.text = text;
+		stream.push({ type: "text_delta", contentIndex, delta: text, partial: output });
+	};
+	return (event: StreamEventShape): void => {
+		if (
+			(event.type === "response.output_item.added" || event.type === "response.output_item.done")
+			&& event.item?.type === "web_search_call"
+		) {
+			const search = extractWebSearch(event.item, { completed: event.type === "response.output_item.done" });
+			if (search) updateActivity(search);
+		}
+		const progress = extractWebSearchProgress(event);
+		if (progress) updateActivity(progress);
+	};
+}

package/src/tools/web-search/render.ts

@@ -0,0 +1,48 @@
+import type { ExtensionAPI } from "@earendil-works/pi-coding-agent";
+import { Container, Text } from "@earendil-works/pi-tui";
+import { glyphs } from "@oai404iao/pi-codex-runtime/internal/glyphs";
+import { themeBold, themeFg } from "@oai404iao/pi-codex-runtime/internal/utils/theme";
+import { WEB_SEARCH_ACTIVITY_MESSAGE_TYPE, buildWebSearchSummaryText, webSearchActivityDetail, webSearchActivityHosts, type SurfacedWebSearch } from "./activity.js";
+
+export function registerWebSearchActivityRenderer(pi: ExtensionAPI): void {
+
+
+	pi.registerMessageRenderer<{ searches?: SurfacedWebSearch[] }>(WEB_SEARCH_ACTIVITY_MESSAGE_TYPE, (message, options, theme) => {
+		const searches = message.details?.searches ?? [];
+		const container = new Container();
+		if (searches.length > 0) {
+			searches.forEach((search, index) => {
+					const completed = search.completed ?? search.status === "completed";
+					const header = completed ? "Searched the web" : "Searching the web";
+					const detail = webSearchActivityDetail(search);
+					const separator = detail ? (completed ? " for " : " ") : "";
+					const bullet = themeFg(theme, completed ? "muted" : "accent", glyphs().bullet);
+					const lines = [
+						`${bullet}${themeFg(theme, "text", themeBold(theme, header))}${themeFg(theme, "dim", `${separator}${detail}`)}`,
+					];
+					const hosts = webSearchActivityHosts(search);
+					if (hosts.length > 0) {
+						const shown = hosts.slice(0, 8);
+						const hostLine = shown.map((host) => themeFg(theme, "accent", host));
+						if (hosts.length > shown.length) {
+							hostLine.push(themeFg(theme, "dim", `+${hosts.length - shown.length}`));
+						}
+						lines.push(`  ${hostLine.join(themeFg(theme, "dim", glyphs().dot))}`);
+					}
+					container.addChild(new Text(`${index > 0 ? "\n" : ""}${lines.join("\n")}`, 0, 0));
+				});
+		} else {
+			container.addChild(new Text(themeFg(theme, "text", themeBold(theme, buildWebSearchSummaryText(searches))), 0, 0));
+		}
+		if (options.expanded) {
+			const content = typeof message.content === "string"
+				? message.content
+				: message.content
+						.filter((item) => item.type === "text")
+						.map((item) => item.text)
+						.join("\n");
+			container.addChild(new Text(`\n${themeFg(theme, "dim", content)}`, 0, 0));
+		}
+		return container;
+	});
+}

package/src/tools/web-search/schema.ts

@@ -0,0 +1,171 @@
+export interface SearchQuery {
+	q: string;
+	recency?: number;
+	domains?: string[];
+}
+
+export interface WebSearchInput {
+	search_query?: SearchQuery[];
+	image_query?: SearchQuery[];
+	open?: Array<{ ref_id: string; lineno?: number }>;
+	click?: Array<{ ref_id: string; id: number }>;
+	find?: Array<{ ref_id: string; pattern: string }>;
+	screenshot?: Array<{ ref_id: string; pageno: number }>;
+	finance?: Array<{
+		ticker: string;
+		type: "equity" | "fund" | "crypto" | "index";
+		market?: string;
+	}>;
+	weather?: Array<{ location: string; start?: string; duration?: number }>;
+	sports?: Array<{
+		tool?: "sports";
+		fn: "schedule" | "standings";
+		league: "nba" | "wnba" | "nfl" | "nhl" | "mlb" | "epl" | "ncaamb" | "ncaawb" | "ipl";
+		team?: string;
+		opponent?: string;
+		date_from?: string;
+		date_to?: string;
+		num_games?: number;
+		locale?: string;
+	}>;
+	time?: Array<{ utc_offset: string }>;
+	response_length?: "short" | "medium" | "long";
+}
+
+const searchQuerySchema = {
+	type: "object",
+	additionalProperties: false,
+	required: ["q"],
+	properties: {
+		q: { type: "string", minLength: 1, description: "Search query." },
+		recency: { type: "integer", minimum: 0, description: "Restrict results to this many recent days." },
+		domains: { type: "array", items: { type: "string", minLength: 1 }, description: "Restrict results to these domains." },
+	},
+};
+
+export const webSearchToolSchema = {
+	type: "object",
+	additionalProperties: false,
+	properties: {
+		search_query: {
+			type: "array",
+			maxItems: 4,
+			items: searchQuerySchema,
+			description: "Query the internet search engine.",
+		},
+		image_query: {
+			type: "array",
+			maxItems: 2,
+			items: searchQuerySchema,
+			description: "Query the image search engine.",
+		},
+		open: {
+			type: "array",
+			items: {
+				type: "object",
+				additionalProperties: false,
+				required: ["ref_id"],
+				properties: {
+					ref_id: { type: "string" },
+					lineno: { type: "integer", minimum: 0 },
+				},
+			},
+		},
+		click: {
+			type: "array",
+			items: {
+				type: "object",
+				additionalProperties: false,
+				required: ["ref_id", "id"],
+				properties: {
+					ref_id: { type: "string" },
+					id: { type: "integer", minimum: 0 },
+				},
+			},
+		},
+		find: {
+			type: "array",
+			items: {
+				type: "object",
+				additionalProperties: false,
+				required: ["ref_id", "pattern"],
+				properties: {
+					ref_id: { type: "string" },
+					pattern: { type: "string" },
+				},
+			},
+		},
+		screenshot: {
+			type: "array",
+			items: {
+				type: "object",
+				additionalProperties: false,
+				required: ["ref_id", "pageno"],
+				properties: {
+					ref_id: { type: "string" },
+					pageno: { type: "integer", minimum: 0 },
+				},
+			},
+		},
+		finance: {
+			type: "array",
+			items: {
+				type: "object",
+				additionalProperties: false,
+				required: ["ticker", "type"],
+				properties: {
+					ticker: { type: "string" },
+					type: { type: "string", enum: ["equity", "fund", "crypto", "index"] },
+					market: { type: "string" },
+				},
+			},
+		},
+		weather: {
+			type: "array",
+			items: {
+				type: "object",
+				additionalProperties: false,
+				required: ["location"],
+				properties: {
+					location: { type: "string" },
+					start: { type: "string" },
+					duration: { type: "integer", minimum: 1 },
+				},
+			},
+		},
+		sports: {
+			type: "array",
+			items: {
+				type: "object",
+				additionalProperties: false,
+				required: ["fn", "league"],
+				properties: {
+					tool: { type: "string", enum: ["sports"] },
+					fn: { type: "string", enum: ["schedule", "standings"] },
+					league: { type: "string", enum: ["nba", "wnba", "nfl", "nhl", "mlb", "epl", "ncaamb", "ncaawb", "ipl"] },
+					team: { type: "string" },
+					opponent: { type: "string" },
+					date_from: { type: "string" },
+					date_to: { type: "string" },
+					num_games: { type: "integer", minimum: 1 },
+					locale: { type: "string" },
+				},
+			},
+		},
+		time: {
+			type: "array",
+			items: {
+				type: "object",
+				additionalProperties: false,
+				required: ["utc_offset"],
+				properties: {
+					utc_offset: { type: "string" },
+				},
+			},
+		},
+		response_length: {
+			type: "string",
+			enum: ["short", "medium", "long"],
+		},
+	},
+};