fix: skip anti-bot detection for PDF responses

[nightcityblade]

Summary

Fixes #2135.

Treat responses declared as application/pdf like other binary content before anti-bot classification. This prevents PDFCrawlerStrategy's intentional 33-byte HTML placeholder from triggering retries, fallback fetching, and the final failure veto after PDF extraction succeeds.

List of files changed and why

• crawl4ai/async_webcrawler.py - detect PDF response content types and bypass HTML-specific anti-bot checks in both classification stages.
• tests/test_issue_2135_pdf_antibot.py - exercise the real PDFCrawlerStrategy through AsyncWebCrawler.arun() and verify one attempt, success, and no fallback.

How Has This Been Tested?

• .venv/bin/python -m pytest tests/test_issue_2135_pdf_antibot.py -q - 1 passed.
• .venv/bin/python tests/proxy/test_antibot_detector.py - 47 passed, 0 failed.
• .venv/bin/ruff check --isolated --select E9,F63,F7,F82 --ignore F821 crawl4ai/async_webcrawler.py tests/test_issue_2135_pdf_antibot.py - passed.
• .venv/bin/black --check --target-version py312 --fast tests/test_issue_2135_pdf_antibot.py - passed.
• .venv/bin/python -m compileall -q crawl4ai/async_webcrawler.py tests/test_issue_2135_pdf_antibot.py - passed.
• git diff --check upstream/develop...HEAD - passed.

Checklist:

• My code follows the style guidelines of this project
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas - N/A; the PDF-response condition is explicit and covered by a focused regression test.
• I have made corresponding changes to the documentation - N/A; this fixes the implementation to match the existing PDF parsing documentation.
• I have added/updated unit tests that prove my fix is effective or that my feature works
• New and existing relevant tests pass locally with my changes

[nightcityblade]

CI note: the only failing check is the repository's notify-discord job because DISCORD_WEBHOOK is empty for this fork PR. The job exits before any code validation; the focused PDF/anti-bot tests and lint checks reported in the PR body are green.