test: fetch iread() URL-handling test image from GitHub, not petercorke.com

[petercorke]

Summary

• `test_iread` and `test_iread_url_handling` (`tests/base/test_io.py`) both fetch `https://petercorke.com/files/images/monalisa.png\` over real HTTPS to exercise `iread()`'s URL-fetch path. This has been a repeated source of flaky CI failures this cycle -- PR fix: add socket timeout to iread() URL fetches #37 added a socket timeout so a slow/unavailable server fails fast instead of hanging, but that doesn't remove the dependency on that one personal server's uptime.
• Switched both to `raw.githubusercontent.com`, serving the identical image already tracked in this repo (`packages/mvtb-data/mvtbdata/images/monalisa.png` -- confirmed same `(700, 677, 3)` shape as the file the old URL served, so the existing assertions are unchanged). GitHub's raw content is CDN-backed, considerably more robust under CI load (many parallel matrix legs) than a single personal server, and since the file already lives in this repo there's no new dependency to maintain -- it's a staple MVTB test image that isn't going anywhere.

Test plan

• Both tests run and pass against the real new URL (not mocked)

[codacy-production]

Up to standards ✅

🟢 Issues 0 issues

Results:
0 new issues

View in Codacy

🟢 Metrics 0 complexity · 0 duplication

Metric	Results
Complexity	0
Duplication	0

View in Codacy

_{NEW Get contextual insights on your PRs based on Codacy's metrics, along with PR and Jira context, without leaving GitHub. Enable AI reviewer
TIP This summary will be updated as you push new changes.}