fix(console): honor gpu model filters in bid screening - #106
Merged
Merged
Conversation
troian
force-pushed
the
fix/console-screen-gpu-model
branch
from
September 24, 2026 15:08
ccada1e to
529b63a
Compare
troian
approved these changes
Sep 24, 2026
Encode GPU models in resource attribute keys so Console can filter providers. Keep empty screening results as arrays in JSON and YAML. Add wire-request and output regression tests. Console automatic funding from issue #103 is already covered by #94. Closes #103 Signed-off-by: Joseph Chalabi <chalabi.joseph@gmail.com>
chalabi2
force-pushed
the
fix/console-screen-gpu-model
branch
from
September 25, 2026 03:59
529b63a to
11603e3
Compare
This branch was successfully deployed
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
console screen --gpu-modelsent an attribute shape the API did not use for model filtering, so even nonexistent models could return providers. Sendvendor/nvidia/model/<model>=trueand preserve an empty provider array in JSON/YAML. Pretty output keeps its existing no-match message.Add wire-request regression coverage for RTX 5090, A100, unknown models, and SDL overrides, plus empty-result checks for all three output formats. Update SPEC.md, DESIGN.md, and the changelog fragment.
The deposit half of #103 was already fixed by #94. Existing tests confirm Console deployments accept the default/auto deposit and omit the deprecated API field.
Validation:
Reproduced the GPU attribute and empty structured-output failures before the fix.
Console CLI/client, transport, and workflow test suites pass.
Screening and output-error race tests pass; Console CLI lint reports no issues.
Release-style build and changelog check pass.
Public API check with a native macOS build: three A100s matched 4 providers; three RTX 5090s and the fake model matched 0 at test time. The fake model returned empty arrays in JSON and YAML. The native build was used because the release build's Go DNS resolver timed out against this machine's Tailscale DNS.
GitHub CI passes, including the live Console sandbox lifecycle, coverage reports, and required gate.
Closes #103