@opencraw/studio 0.0.11 → 0.0.13

package/dist/index.esm.js

@@ -1,12 +1,12 @@
 import { join, dirname, basename, extname, normalize, sep, resolve } from 'node:path';
 import { createServer } from 'node:http';
+import { promisify, stripVTControlCharacters } from 'node:util';
 import { fileURLToPath, pathToFileURL } from 'node:url';
 import { z } from 'zod';
-import { BODY_KINDS, HttpClient, BrowserClient, NODE_ID_ATTRIBUTE, rowsOfCells, parseXml, findDeckTables, findGridTables, analyzeTables, parseRegion, regionText, deckRegionText, isDeckDocument, readRecipeSource, parseInputRecipe, parseOutputRecipe, bindRecipeSet, memorySink, createCrawler, traceLine, UnknownHookError, resolveMissingPolicy, isWorkbookDocument, RecipeBindingError, RecipeValidationError, countMatches, matchesOf, tryParseJson, isPdfDocument, BrowserProfiles } from '@opencraw/core';
+import { BODY_KINDS, HttpClient, BrowserClient, NODE_ID_ATTRIBUTE, sessionStateOf, rowsOfCells, parseXml, findDeckTables, findGridTables, analyzeTables, parseRegion, regionText, deckRegionText, isDeckDocument, readRecipeSource, parseInputRecipe, parseOutputRecipe, bindRecipeSet, memorySink, createCrawler, traceLine, UnknownHookError, resolveMissingPolicy, isWorkbookDocument, RecipeBindingError, RecipeValidationError, countMatches, matchesOf, tryParseJson, isPdfDocument, BrowserProfiles } from '@opencraw/core';
 import { randomBytes, createHash } from 'node:crypto';
 import { execFile } from 'node:child_process';
 import { createRequire } from 'node:module';
-import { promisify } from 'node:util';
 import { load } from 'cheerio';
 import { readFile, writeFile, mkdir } from 'node:fs/promises';
 import { describeJson } from '@opencraw/probe';
@@ -1119,12 +1119,30 @@ async function takeSnapshot(input, stepPath, browser) {
   if (stepPath !== 'start') throw new Error(`takeSnapshot: "${stepPath}" is not supported yet — only "start" (the first start point, before any step runs) is; see this function's doc comment`);
   const point = input.start[0];
   if (point === undefined) throw new Error(`recipe "${input.id}" has no start point`);
-  return input.mode === 'web' ? snapshotWeb(point.url, browser) : snapshotApi(point.url);
+  const storageState = await loginState(input, browser);
+  return input.mode === 'web' ? snapshotWeb(point.url, storageState, browser) : snapshotApi(point.url, storageState);
 }
-async function snapshotWeb(url, browser) {
+/** What the recipe's own login leaves behind, so the snapshot shows the page as the crawl sees it (issue #185); nothing when it declares no login. */
+async function loginState(input, browser) {
+  if (input.session === undefined) return undefined;
+  let client;
+  try {
+    return await sessionStateOf(input, {
+      browser: async () => {
+        client = await ensureBrowserLaunch(browser, () => BrowserClient.launch(browser));
+        return client;
+      }
+    });
+  } finally {
+    await client?.close();
+  }
+}
+async function snapshotWeb(url, storageState, browser) {
   const client = await ensureBrowserLaunch(browser, () => BrowserClient.launch(browser));
   try {
-    const session = await client.newSession();
+    const session = await client.newSession({
+      storageState
+    });
     try {
       await session.page.goto(url);
       await session.page.evaluate(hiddenMarksScript(HIDDEN_ATTRIBUTE));
@@ -1143,8 +1161,10 @@ async function snapshotWeb(url, browser) {
     await client.close();
   }
 }
-async function snapshotApi(url) {
-  const client = await HttpClient.open({});
+async function snapshotApi(url, storageState) {
+  const client = await HttpClient.open({
+    storageState
+  });
   try {
     const response = await client.send({
       url
@@ -2105,9 +2125,9 @@ async function runToResult(crawler, set, recipeId, records, rejectedRecords, ste
 }
 /** What a run that could not start reports. A hook is the author's own code, which Studio never loads (issue #150): say so, instead of leaving "unknown hook" to read as a typo. */
 function failureMessage(error) {
-  if (error instanceof UnknownHookError) return `${error.message}. Studio does not load hooks, so a recipe that uses one cannot be sample-run here; run it with the cli or the mcp server, or temporarily replace the hook step to preview the rest.`;
-  if (error instanceof Error) return error.message;
-  return String(error);
+  if (error instanceof UnknownHookError) return `${stripVTControlCharacters(error.message)}. Studio does not load hooks, so a recipe that uses one cannot be sample-run here; run it with the cli or the mcp server, or temporarily replace the hook step to preview the rest.`;
+  if (error instanceof Error) return stripVTControlCharacters(error.message);
+  return stripVTControlCharacters(String(error));
 }
 
 /** @returns A fresh, empty cache. */
@@ -5070,7 +5090,7 @@ async function handlePdfBytesRequest(response, state, query) {
     response.end(Buffer.from(bytes));
   } catch (error) {
     respondJson(response, 500, {
-      error: error instanceof Error ? error.message : String(error)
+      error: messageOf(error)
     });
   }
 }
@@ -5088,7 +5108,7 @@ async function handleCommand(request, response, state) {
     command = parsed.data;
   } catch (error) {
     respondJson(response, 400, {
-      error: error instanceof Error ? error.message : String(error)
+      error: messageOf(error)
     });
     return;
   }
@@ -5096,7 +5116,7 @@ async function handleCommand(request, response, state) {
     respondJson(response, 200, await dispatch(command, state));
   } catch (error) {
     respondJson(response, 500, {
-      error: error instanceof Error ? error.message : String(error)
+      error: messageOf(error)
     });
   }
 }
@@ -5226,6 +5246,10 @@ function handleUpgrade(request, socket, token, state) {
     state.sockets.delete(connection);
   });
 }
+/** An error's message as plain text: Playwright colours its call logs with ANSI codes, which a browser would show as raw escapes (issue #187). */
+function messageOf(error) {
+  return stripVTControlCharacters(error instanceof Error ? error.message : String(error));
+}
 
 /**
  * Runs `opencraw-studio`: starts the server, opens `folder` (if given) as

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@opencraw/studio",
-  "version": "0.0.11",
+  "version": "0.0.13",
   "type": "module",
   "main": "./dist/index.esm.js",
   "module": "./dist/index.esm.js",
@@ -24,8 +24,8 @@
     "!**/*.js.map"
   ],
   "dependencies": {
-    "@opencraw/core": "^0.1.22",
-    "@opencraw/probe": "^0.0.5",
+    "@opencraw/core": "^0.1.24",
+    "@opencraw/probe": "^0.0.7",
     "cheerio": "^1.2.0",
     "playwright": "^1.63.0",
     "zod": "^4.6.5",