@sisu-ai/tool-summarize-text 2.0.0

package/README.md

@@ -0,0 +1,18 @@
+# Summarize Text Tool
+
+A utility tool that asks the active model to summarize large text bodies into concise outputs. Designed for use alongside `webFetch` or other content tools.
+
+## Tool
+- `summarizeText` — Summarize a given `text` with optional controls for length and style.
+
+## Args
+- `text` (string, required): The input to summarize.
+- `target` ("short"|"medium"|"long", optional): Coarse target length. Default `medium`.
+- `maxChars` (number, optional): Hard cap for output characters.
+- `bullets` (boolean, optional): Prefer bullet-point output.
+- `includeCitations` (boolean, optional): Keep/collect URLs in the summary when present.
+- `focus` (string, optional): Guidance on what to emphasize.
+
+## Notes
+- Uses `ctx.model.generate` internally with `toolChoice:'none'` to avoid nested tool calls.
+- For very long inputs, applies a simple map-reduce (chunk→summary→combine) approach.

package/dist/index.d.ts

@@ -0,0 +1,16 @@
+import type { Tool } from '@sisu-ai/core';
+export type TargetLength = 'short' | 'medium' | 'long';
+export interface SummarizeArgs {
+    text: string;
+    target?: TargetLength;
+    maxChars?: number;
+    bullets?: boolean;
+    includeCitations?: boolean;
+    focus?: string;
+}
+export interface SummarizeResult {
+    summary: string;
+    urls?: string[];
+}
+export declare const summarizeText: Tool<SummarizeArgs>;
+export default summarizeText;

package/dist/index.js

@@ -0,0 +1,84 @@
+import { z } from 'zod';
+export const summarizeText = {
+    name: 'summarizeText',
+    description: 'Summarize a block of text using the current model. Useful for condensing large webFetch outputs while keeping key facts and URLs.',
+    schema: z.object({
+        text: z.string().min(1),
+        target: z.enum(['short', 'medium', 'long']).optional(),
+        maxChars: z.number().int().positive().max(50_000).optional(),
+        bullets: z.boolean().optional(),
+        includeCitations: z.boolean().optional(),
+        focus: z.string().optional(),
+    }),
+    handler: async ({ text, target = 'medium', maxChars, bullets, includeCitations, focus }, ctx) => {
+        const cap = Math.min(Math.max(Number(maxChars ?? 2000), 200), 50_000);
+        const chunks = chunkText(text, 10_000);
+        const chunkSummaries = [];
+        for (const [i, ch] of chunks.entries()) {
+            // Allocate characters per chunk, allowing a buffer of 200, but not exceeding the cap
+            const charsPerChunk = Math.min(Math.floor(cap / Math.max(chunks.length, 1)) + 200, cap);
+            const prompt = buildPrompt(ch, target, charsPerChunk, bullets, includeCitations, focus, `Part ${i + 1}/${chunks.length}`);
+            const res = await ctx.model.generate(prompt, { toolChoice: 'none', signal: ctx.signal });
+            const s = String(res?.message?.content ?? '').trim();
+            if (s)
+                chunkSummaries.push(s.slice(0, cap));
+        }
+        const combinedInput = chunkSummaries.join('\n\n');
+        const finalPrompt = buildPrompt(combinedInput || text, target, cap, bullets, includeCitations, focus, 'Final Synthesis');
+        const res = await ctx.model.generate(finalPrompt, { toolChoice: 'none', signal: ctx.signal });
+        const summary = String(res?.message?.content ?? '').slice(0, cap);
+        const urls = includeCitations ? extractUrls([summary]) : undefined;
+        return { summary, ...(urls && urls.length ? { urls } : {}) };
+    }
+};
+export default summarizeText;
+function buildPrompt(text, target, maxChars, bullets, includeCitations, focus, label) {
+    const aims = [];
+    aims.push(`Keep under ${maxChars} characters.`);
+    if (bullets)
+        aims.push('Prefer concise bullet points.');
+    if (includeCitations)
+        aims.push('Preserve any URLs as citations when relevant.');
+    if (focus)
+        aims.push(`Emphasize: ${focus}`);
+    const density = target === 'short' ? 'High compression' : target === 'long' ? 'Low compression' : 'Balanced compression';
+    const sys = `You are a careful summarizer. ${density}. Do not invent facts. If URLs are present, keep them.`;
+    const usr = `${label ? `[${label}] ` : ''}Summarize the following text.\n\n${text}`;
+    const guide = aims.length ? `Guidelines: ${aims.join(' ')}\n` : '';
+    return [
+        { role: 'system', content: sys },
+        { role: 'user', content: `${guide}${usr}` }
+    ];
+}
+function chunkText(s, size = 10_000) {
+    if (s.length <= size)
+        return [s];
+    const out = [];
+    let i = 0;
+    while (i < s.length) {
+        out.push(s.slice(i, i + size));
+        i += size;
+    }
+    return out;
+}
+function extractUrls(contents) {
+    /**
+     * Regular expression to match URLs starting with "http://" or "https://".
+     *
+     * This pattern matches any substring that:
+     * - Begins with "http://" or "https://"
+     * - Is followed by any sequence of characters except whitespace, closing parentheses, square brackets, double quotes, single quotes, greater-than signs, or angle brackets.
+     *
+     * Excluded characters (`\s)\]"'>`) are commonly found at the end of URLs in text (such as punctuation or delimiters) and are not considered part of the URL.
+     *
+     * @remarks
+     * While this regex works for many common cases, URL parsing can be complex and edge cases may not be handled correctly.
+     * For more robust and accurate URL extraction, we should consider using a dedicated URL parsing library.
+     */
+    const urlRe = /https?:\/\/[^\s)\]"'>]+/gi;
+    const out = new Set();
+    for (const c of contents)
+        for (const m of c.matchAll(urlRe))
+            out.add(m[0]);
+    return Array.from(out);
+}

package/package.json

@@ -0,0 +1,28 @@
+{
+  "name": "@sisu-ai/tool-summarize-text",
+  "version": "2.0.0",
+  "type": "module",
+  "main": "dist/index.js",
+  "types": "dist/index.d.ts",
+  "files": [
+    "dist"
+  ],
+  "scripts": {
+    "build": "tsc -b"
+  },
+  "dependencies": {
+    "zod": "^3.23.8"
+  },
+  "peerDependencies": {
+    "@sisu-ai/core": "0.3.0"
+  },
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/finger-gun/sisu",
+    "directory": "packages/tools/summarize-text"
+  },
+  "homepage": "https://github.com/finger-gun/sisu#readme",
+  "bugs": {
+    "url": "https://github.com/finger-gun/sisu/issues"
+  }
+}