Conversation
Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Run Compose-defined services through the sandbox API with explicit provider capability validation and configurable networking, storage, and runtime requirements. Preserve dependency and health-check ordering and clean up collections safely, including short-lived services and shared providers. Co-Authored-By: Codex <noreply@openai.com> Signed-off-by: Hemil Desai <hemild@nvidia.com> (cherry picked from commit c5c15ee) Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Serialize named service handles through ConnectableProvider and reconnect without provisioning. Keep managed-volume cleanup with the creating server and preserve provider close semantics for connected services. Keep Compose-specific operations on the optional provider interfaces instead of adding public forwarding methods to the sandbox facade. Co-Authored-By: Codex <noreply@openai.com> Signed-off-by: Hemil Desai <hemild@nvidia.com> (cherry picked from commit 91750e0) Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Show a two-service Compose setup, Python execution, and cross-server serialization with explicit lifecycle ownership. Link the guide from the sandbox overview and describe configurable provider capabilities. Co-Authored-By: Codex <noreply@openai.com> Signed-off-by: Hemil Desai <hemild@nvidia.com> (cherry picked from commit d7af208) Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Use docker image inspect as the default metadata command and document pulling images for the sandbox platform before startup. Preserve image startup and Docker-specific health and shell defaults without requiring crane. Keep the inspection command configurable. Co-Authored-By: Codex <noreply@openai.com> Signed-off-by: Hemil Desai <hemild@nvidia.com> (cherry picked from commit 180b66a) Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Read image manifests and configuration directly from registries, preserving platform selection and Docker health metadata. Use standalone official Compose for normalization without requiring a Docker daemon. Co-Authored-By: Codex <noreply@openai.com> Signed-off-by: Hemil Desai <hemild@nvidia.com> (cherry picked from commit 63b86b9) Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Allow deployments to supply complete image defaults and skip registry lookup for those images. Co-Authored-By: Codex <noreply@openai.com> Signed-off-by: Hemil Desai <hemild@nvidia.com> (cherry picked from commit 3377b7c) Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Remove the public metadata cache and image health, shell, and environment emulation. Skip registry access for explicit YAML entrypoints. Validate all 11 TB4 Compose gold tasks using their official images and verifiers. Co-Authored-By: Codex <noreply@openai.com> Signed-off-by: Hemil Desai <hemild@nvidia.com> (cherry picked from commit 0aebb36) Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Remove image inspection, metadata options, and the ORAS dependency. Consume startup commands, working directories, and ports from YAML and reject missing commands before provisioning. Co-Authored-By: Codex <noreply@openai.com> Signed-off-by: Hemil Desai <hemild@nvidia.com> (cherry picked from commit e6555f7) Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Co-Authored-By: Codex <noreply@openai.com> Signed-off-by: Hemil Desai <hemild@nvidia.com> (cherry picked from commit 9286eed) Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Mark the two synthetic basic-auth test URLs as false positives and clarify that x-sandbox is an optional Gym-specific extension. Co-Authored-By: Codex <noreply@openai.com> Signed-off-by: Hemil Desai <hemild@nvidia.com> (cherry picked from commit f97f4b7) Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Use existing PyYAML and validate upstream-resolved inputs before provisioning. Document the supported input boundary and remove Compose executable configuration. Co-Authored-By: Codex <noreply@openai.com> Signed-off-by: Hemil Desai <hemild@nvidia.com> (cherry picked from commit bb03de6) Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
…acts Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Signed-off-by: Frederic Lafrance <flafrance@nvidia.com>
Contributor
|
🌿 Preview your docs: https://nvidia-preview-flafrance-terminal-bench-4.docs.buildwithfern.com/nemo/gym Here are the markdown pages you've updated: |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This is based on #3332
It adds TB4 through the Harbor integration and includes several sub-changes to make all tasks run and be observable.
WIP:
Known limitations: