diff options
| author | amankumarsingh77 <amankumarsingh7702@gmail.com> | 2026-07-08 21:24:32 +0530 |
|---|---|---|
| committer | amankumarsingh77 <amankumarsingh7702@gmail.com> | 2026-07-08 21:24:32 +0530 |
| commit | 68bf7fd2988e42343833be09e4115cd7175029d0 (patch) | |
| tree | 52b8eaf99fa7e4d383ff169dd3bdccfe13911217 /test/deep-research/fixtures/article-capture.json | |
| parent | c5c42df5976ed63faac2adcf3b04a4ad2f519b32 (diff) | |
feat: add browser_read_page reader-mode tool
URL (or owned targetId) → clean main-article text with nav/ads/footer
boilerplate stripped, via a dependency-free readability heuristic. DOM walk
lives in an in-page capture expression; scoring/selection is a pure, fixture-
tested function (link-density + boilerplate-ancestor filtering, body-text
fallback for structure-less pages). Reuses the isolated-attach lifecycle;
registered unserialized (safe concurrent read). 10 extractor unit scenarios.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Diffstat (limited to 'test/deep-research/fixtures/article-capture.json')
| -rw-r--r-- | test/deep-research/fixtures/article-capture.json | 17 |
1 files changed, 17 insertions, 0 deletions
diff --git a/test/deep-research/fixtures/article-capture.json b/test/deep-research/fixtures/article-capture.json new file mode 100644 index 0000000..df9c67c --- /dev/null +++ b/test/deep-research/fixtures/article-capture.json @@ -0,0 +1,17 @@ +{ + "url": "https://blog.example.com/how-cdp-works", + "title": "How the Chrome DevTools Protocol Works", + "blocks": [ + { "kind": "other", "text": "Home Products Blog About Contact", "linkTextLength": 33, "inBoilerplate": true }, + { "kind": "heading", "text": "How the Chrome DevTools Protocol Works", "linkTextLength": 0, "inBoilerplate": false }, + { "kind": "paragraph", "text": "The Chrome DevTools Protocol lets tools instrument, inspect, and control Chromium-based browsers over a WebSocket connection.", "linkTextLength": 0, "inBoilerplate": false }, + { "kind": "paragraph", "text": "Each command targets a domain such as Page, Runtime, or Network, and receives a structured JSON response that the client correlates by id.", "linkTextLength": 0, "inBoilerplate": false }, + { "kind": "heading", "text": "Sessions and Targets", "linkTextLength": 0, "inBoilerplate": false }, + { "kind": "paragraph", "text": "A target is a page or worker; attaching to one yields a session id that scopes subsequent commands to that target without disturbing others.", "linkTextLength": 0, "inBoilerplate": false }, + { "kind": "listitem", "text": "Attach with Target.attachToTarget", "linkTextLength": 0, "inBoilerplate": false }, + { "kind": "listitem", "text": "Drive with Runtime.evaluate", "linkTextLength": 0, "inBoilerplate": false }, + { "kind": "paragraph", "text": "Subscribe here for our newsletter and follow us on social media for updates", "linkTextLength": 55, "inBoilerplate": true }, + { "kind": "other", "text": "Copyright 2026 Example Inc. Privacy Terms Cookies", "linkTextLength": 24, "inBoilerplate": true } + ], + "bodyText": "Home Products Blog How the Chrome DevTools Protocol Works The Chrome DevTools Protocol lets tools instrument inspect and control browsers Copyright 2026" +} |