pi-unsloth-webtools 0.2.4 → 0.2.5

package/README.md

@@ -45,6 +45,8 @@ Port of Studio's `_fetch_page_text` / `_fetch_url_raw` pipeline:
   validation and fetch.
 - GitHub repo root pages are rewritten to the unauthenticated README API
   (`Accept: application/vnd.github.raw+json`), falling back to the HTML page on failure.
+- HTTP 4xx/5xx responses are reported as errors with the status reason instead of
+  returning error-page content.
 - Up to 4 redirect hops, each re-validated and re-resolved against the same rules.
 - 512 KiB download cap (10 MiB for PDFs), overall deadline + per-hop socket timeouts, abort-aware
   (`signal` cancels mid-flight).
@@ -53,7 +55,7 @@ Port of Studio's `_fetch_page_text` / `_fetch_url_raw` pipeline:
   pymupdf4llm-style markdown layer (headings, bold/italic, code fences, links, tables)
   with Studio's corrupted/incomplete fallback to plain text.
 - Content sniffing: MIME allow/deny, binary magic signatures, PDF magic detection, and charset
-  decoding (declared charset, BOM sniffing for UTF-8/16/32, `<meta charset>` sniffing for
+  decoding (BOM sniffing for UTF-8/16/32 first, then the declared charset, `<meta charset>` sniffing for
   CJK and Windows/ISO encodings, cp1252 rescue for mislabeled single-byte pages).
 - HTML → Markdown conversion ported from Studio's dependency-free `_html_to_md.py`: headings,
   links, emphasis, lists, tables, blockquotes, code fences, entity decoding; hidden-element
@@ -78,7 +80,9 @@ Port of Studio's `_fetch_page_text` / `_fetch_url_raw` pipeline:
   database.
 - Empty sweeps: ddgs 9.14.4 raises the last engine exception; this port reports a
   timeout whenever any engine timed out, so the timeout message is not masked by later
-  generic engine failures.
+  generic engine failures. The timeout budget bounds the entire sweep: per-engine
+  timeouts shrink as the budget is consumed, so the reported timeout matches the
+  worst-case wall time.
 - Proxies: Studio routes through environment proxies; this port always connects
   directly with DNS pinning (deliberately out of scope).
 

package/engines.ts

@@ -510,6 +510,7 @@ async function httpFetch(
     Accept: "*/*",
     ...options.headers,
   };
+  if (options.method === "POST") headers["Content-Type"] = "application/x-www-form-urlencoded";
   const cookie = options.cookies
     ? Object.entries(options.cookies)
         .map(([key, value]) => `${key}=${value}`)
@@ -695,6 +696,7 @@ const WIKIPEDIA: Engine = {
   provider: "wikipedia",
   priority: 2,
   async search(query, ctx, timeoutMs, signal) {
+    const started = Date.now();
     const lang = ctx.region.toLowerCase().split("-")[1] ?? "en";
     const encoded = encodeURIComponent(query);
     const opensearchUrl =
@@ -715,7 +717,7 @@ const WIKIPEDIA: Engine = {
     const extractUrl =
       `https://${lang}.wikipedia.org/w/api.php?action=query&format=json&prop=extracts` +
       `&titles=${encodeURIComponent(title)}&explaintext=0&exintro=0&redirects=1`;
-    const extract = await httpGet(extractUrl, {}, { timeoutMs, signal });
+    const extract = await httpGet(extractUrl, {}, { timeoutMs: Math.max(1, timeoutMs - (Date.now() - started)), signal });
     if (extract) {
       try {
         const pageData = JSON.parse(extract) as {
@@ -815,6 +817,7 @@ export async function autoTextSearch(
   signal?: AbortSignal,
 ): Promise<SearchResult[]> {
   const engines = shuffledEngines();
+  const deadline = Date.now() + timeoutMs;
   const seenProviders = new Set<string>();
   const aggregator = new ResultsAggregator();
   const ctx: EngineContext = { region: "us-en", safesearch: "moderate" };
@@ -825,8 +828,9 @@ export async function autoTextSearch(
   let i = 0;
   let pending: Promise<void>[] = [];
   const run = async (engine: Engine) => {
+    const remaining = Math.max(1, deadline - Date.now());
     try {
-      const results = await engine.search(query, ctx, timeoutMs, signal);
+      const results = await engine.search(query, ctx, remaining, signal);
       if (results && results.length) {
         aggregator.extend(results);
         seenProviders.add(engine.provider);

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pi-unsloth-webtools",
-  "version": "0.2.4",
+  "version": "0.2.5",
   "type": "module",
   "description": "Pi extension: web_search and web_fetch tools ported from the Unsloth Studio codebase (DuckDuckGo search, SSRF-safe direct fetching, HTML-to-Markdown extraction)",
   "main": "index.ts",
@@ -51,7 +51,8 @@
     "test:unit": "vitest run --exclude **/smoke.test.ts",
     "test:smoke": "vitest run test/smoke.test.ts",
     "typecheck": "tsc --noEmit",
-    "prepublishOnly": "npm run typecheck && npm run test:unit"
+    "check:package": "node -e \"const fs=require('fs');const pkg=require('./package.json');const listed=new Set(pkg.files);const bad=[];for(const f of pkg.files){if(!fs.existsSync(f))bad.push('missing file: '+f)}for(const f of fs.readdirSync('.').filter(f=>f.endsWith('.ts'))){if(!listed.has(f))bad.push('unlisted source: '+f)}if(bad.length){console.error(bad.join('\\n'));process.exit(1)}\"",
+    "prepublishOnly": "npm run typecheck && npm run test:unit && npm run check:package"
   },
   "devDependencies": {
     "@earendil-works/pi-coding-agent": "^0.84.0",

package/web-fetch.ts

@@ -462,6 +462,7 @@ export function requestHop(opts: HopOptions): Promise<HopResponse> {
     const request = transport.request(options, (res: IncomingMessage) => {
       const chunks: Buffer[] = [];
       let total = 0;
+      let head = Buffer.alloc(0);
       const declaredPdf = String(res.headers["content-type"] ?? "").toLowerCase().includes("pdf");
       let limit = declaredPdf ? opts.maxPdfBytes : opts.maxBytes;
       let extendedForPdf = false;
@@ -488,8 +489,12 @@ export function requestHop(opts: HopOptions): Promise<HopResponse> {
           finish("timed out", Buffer.concat(chunks));
           return;
         }
+        if (head.length < 1024) {
+          const need = 1024 - head.length;
+          head = Buffer.concat([head, chunk.subarray(0, Math.min(need, chunk.length))]);
+        }
         if (!declaredPdf && !extendedForPdf && total + chunk.length > opts.maxBytes) {
-          if (hasPdfMagic(Buffer.concat(chunks))) {
+          if (hasPdfMagic(head)) {
             limit = opts.maxPdfBytes;
             extendedForPdf = true;
           }
@@ -628,6 +633,15 @@ export async function fetchUrlRaw(
     budgetError = fetchBudgetExceeded(deadline, signal, now);
     if (budgetError !== null) return { error: budgetError, body: "", contentType: "" };
 
+    if (response.status >= 400) {
+      const reason = http.STATUS_CODES[response.status] ?? "";
+      return {
+        error: `Failed to fetch URL: HTTP ${response.status}${reason ? ` ${reason}` : ""}`,
+        body: "",
+        contentType: "",
+      };
+    }
+
     const contentTypeHeader = response.headers["content-type"];
     const contentType = contentTypeHeader
       ? (/^[\w.+-]+\/[\w.+-]+/.exec(String(contentTypeHeader).toLowerCase()) ?? [""])[0]
@@ -679,7 +693,7 @@ export async function fetchUrlRaw(
     const bomCodec = bomCodecFor(response.body);
     const rawHtml = decodeWithCodec(
       response.body,
-      declaredCodec ?? bomCodec ?? sniffMetaCharsetForHtml(response.body, contentType) ?? "utf-8",
+      bomCodec ?? declaredCodec ?? sniffMetaCharsetForHtml(response.body, contentType) ?? "utf-8",
     );
 
     if (looksBinary(rawHtml)) {

package/web-search.ts

@@ -93,7 +93,7 @@ export function searchFailureMessage(exc: unknown, timeoutMs = SEARCH_TIMEOUT_MS
 export function formatSearchResults(results: SearchResult[]): string {
   const parts = results.map((result) => {
     const title = String(result.title ?? "").replace(/\s+/g, " ");
-    const href = String(result.href ?? "").trim();
+    const href = String(result.href ?? "").replace(/\s+/g, " ").trim();
     const snippet = String(result.body ?? "").replace(/\s+/g, " ");
     return `Title: ${title}\nURL: ${href}\nSnippet: ${snippet}`;
   });