fix: Prevent iterate_items from terminating early under server-side filters

[vdusek]

Summary

• DatasetClient.iterate_items (both sync and async) terminated early whenever skip_empty, skip_hidden, or clean=True caused a server-side page to contain fewer than 1000 items, silently dropping the remaining dataset items.
• The API applies those filters after scanning the [offset, offset+limit) window, so a short page does not imply the dataset is exhausted. The previous current_page_item_count < cache_size shortcut was incorrect under any server-side filter.
• The request offset was also advanced by the number of items returned rather than scanned, causing subsequent requests to re-scan overlapping windows whenever a filter was active.

Fix

• Track the scanned window explicitly via a new scanned_items counter and use it to compute the next request offset.
• Advance scanned_items by effective_limit (the size of the requested window), not by the count of returned items.
• Terminate only when the server returns zero items (or the caller-supplied limit has been reached).

Tests

• Added regression tests (sync and async) that simulate an API post-filtering half of every scanned 1000-item window across a 2500-item dataset. They fail on master (loop exits after the first ~500 items) and pass on this branch (all ~1250 non-empty items are yielded).

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 95.27%. Comparing base (6c3b84d) to head (1596ed9).
⚠️ Report is 7 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #747      +/-   ##
==========================================
- Coverage   95.36%   95.27%   -0.10%     
==========================================
  Files          45       45              
  Lines        5118     5120       +2     
==========================================
- Hits         4881     4878       -3     
- Misses        237      242       +5     

Flag	Coverage Δ
integration	95.27% <100.00%> (-0.10%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[vdusek]

Closing as this part of codebase will be refactored during #539