cog

1d69171d - feat: add model test harness for validating cog SDK releases (#2851)

feat: add model test harness for validating cog SDK releases (#2851) * feat: add model test harness for validating cog SDK releases Add a declarative test harness (tools/test-harness/) that automates building and running cog models against new SDK versions. Designed for testing cog-examples against RC releases but extensible to any model in any repo. Models and their expected inputs/outputs are defined in manifest.yaml. The harness clones repos, patches cog.yaml with the target sdk_version, runs cog build + cog predict, and validates outputs using pluggable validators (exact match, contains, file_exists, json_match, etc.). Includes all 10 cog-examples models in the manifest and fixture images for blur/resnet tests. * docs: remove internal reference from README * Potential fix for pull request finding 'Empty except' Co-authored-by: Copilot Autofix powered by AI <223894421+github-code-quality[bot]@users.noreply.github.com> Signed-off-by: Mark Phelps <209477+markphelps@users.noreply.github.com> * feat: auto-resolve latest stable cog CLI and SDK versions in test harness Default to downloading the latest stable cog CLI from GitHub releases and resolving the latest stable SDK version from PyPI, skipping any alpha/beta/rc tags. Both can be overridden via --cog-version and --sdk-version CLI flags, or pinned in manifest.yaml defaults. --------- Signed-off-by: Mark Phelps <209477+markphelps@users.noreply.github.com> Co-authored-by: Copilot Autofix powered by AI <223894421+github-code-quality[bot]@users.noreply.github.com>