Convert WeChatQRCode test model fixtures from Caffe to ONNX

[shimat]

Summary

• OpenCV 5.x's wechat_qrcode module (and OpenCvSharp's WeChatQRCode wrapper) only accepts single-file ONNX models now - the Caffe importer has been removed entirely (readNetFromONNX/readNet on a .caffemodel now fail with "Caffe importer has been removed"). The Caffe-format fixtures checked into this repo (detect.prototxt/.caffemodel, sr.prototxt/.caffemodel) were stale leftovers from the 4.x era, so every model-dependent case in WeChatQRCodeTest.cs was silently skipped via SkipIfModelFilesNotFound(), both locally and in CI.
• Replaced them with detect.onnx/sr.onnx, converted from the same published weights (opencv/opencv_3rdparty:wechat_qrcode_20210119, Apache License 2.0 - see the added NOTICE.md). No architecture or weight changes, just a format conversion.
• Related to Where can I download the onnx model required for WeChatQRCode? #2075, where a user asked where to get these ONNX models. As it turns out, opencv_extra's 5.x branch (not its default branch, which is why we missed it at first) does list an ONNX conversion of the same models via download_models.py, sourced from an unlicensed personal GitHub releases repo (omrope79/opencv-test-models). We verified it against our own conversion (equivalent detection output) but converted independently anyway, since our version carries clear attribution back to the original Apache-2.0 source and doesn't depend on an unlicensed third-party mirror.

Testing

• dotnet test --filter "FullyQualifiedName~WeChatQRCode": all 10 tests pass, 0 skipped (previously 7 of them were skipped for lack of model files).

🤖 Generated with Claude Code

Summary by CodeRabbit

• Documentation

  • Added licensing and attribution information for the QR code recognition model assets.

• Chores

  • Updated QR code recognition test resources to use ONNX model artifacts.
  • Removed obsolete Caffe model definition files.
  • Retained the existing recognition behavior while simplifying the packaged test assets.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: d93a54e4-46de-47be-bc13-aa32c4cdfb13

📥 Commits

Reviewing files that changed from the base of the PR and between e1baa1d and 665a043.

📒 Files selected for processing (1)

• test/OpenCvSharp.Tests/_data/wechat_qrcode/NOTICE.md

---

📝 Walkthrough

Walkthrough

The WeChat QR code test assets replace the Caffe detection and super-resolution prototxt files with an ONNX super-resolution model and add NOTICE documentation covering model provenance, licensing, and conversion details.

Changes

WeChat QR model assets

Layer / File(s)	Summary
ONNX model and asset documentation test/OpenCvSharp.Tests/_data/wechat_qrcode/sr.onnx, test/OpenCvSharp.Tests/_data/wechat_qrcode/NOTICE.md	Adds the serialized super-resolution ONNX graph with embedded parameters and documents its source, Apache 2.0 licensing, and Caffe-to-ONNX conversion. The detect.prototxt and sr.prototxt Caffe definitions are removed.

Estimated code review effort: 2 (Simple) | ~10 minutes

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly and accurately summarizes the main change: migrating WeChatQRCode test fixtures from Caffe to ONNX.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch wechat-qrcode-onnx-models-2075

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@test/OpenCvSharp.Tests/_data/wechat_qrcode/NOTICE.md`:
- Around line 1-12: Update the prose paragraphs in NOTICE.md to use one source
line per paragraph, removing mid-sentence hard wraps while preserving the
existing wording, links, code spans, and paragraph separation.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 50cbfc6a-0b36-43b5-8509-08fbe12bdd6e

📥 Commits

Reviewing files that changed from the base of the PR and between 4847ff4 and a3c46e8.

📒 Files selected for processing (5)

• test/OpenCvSharp.Tests/_data/wechat_qrcode/NOTICE.md
• test/OpenCvSharp.Tests/_data/wechat_qrcode/detect.onnx
• test/OpenCvSharp.Tests/_data/wechat_qrcode/detect.prototxt
• test/OpenCvSharp.Tests/_data/wechat_qrcode/sr.onnx
• test/OpenCvSharp.Tests/_data/wechat_qrcode/sr.prototxt

💤 Files with no reviewable changes (2)

• test/OpenCvSharp.Tests/_data/wechat_qrcode/detect.prototxt
• test/OpenCvSharp.Tests/_data/wechat_qrcode/sr.prototxt