Skip to content

Add ask_vlm_verify in ExperimentalApi for cloud VLM alert verification #442

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
srnangi merged 29 commits into from
Jul 10, 2026
Merged

Add ask_vlm_verify in ExperimentalApi for cloud VLM alert verification #442

Changes from 1 commit
Commits
Show all changes
29 commits
Select commit Hold shift + click to select a range
fde518c
Add ask_vlm method for cloud VLM alert verification
Jun 18, 2026
9a5e3e1
Automatically reformatting code
Jun 18, 2026
d3a428b
Send ask_vlm query/model_id as form fields; use friendly model alias
Jun 22, 2026
2b20fce
Update ask_vlm model_id docstring examples to vision-capable aliases
Jun 22, 2026
320887b
ask_vlm: rename images -> media, accept up to 8
Jun 22, 2026
00789e0
ask_vlm: point at renamed /v1/vlm-verifications endpoint
Jun 24, 2026
263808d
fix: correct URL path separator and add regression test
Jun 24, 2026
3cfbb7e
address PR comments: model list in docstring, timeout/corrupted-image…
Jun 24, 2026
7216313
trim ask_vlm tests to meaningful coverage only
Jun 24, 2026
6aad9e0
fix CI failures: numpy optional import, magic value constant, pylint …
Jun 25, 2026
6d8b680
address Copilot review comments on ask_vlm
Jun 25, 2026
397f9cc
fix: clarify media docstring and add /v1 deduplication regression test
Jun 25, 2026
2ca16ff
fix: register ask_vlm in CLI command groups
Jun 25, 2026
b9bf222
refactor: convert VLMVerificationResult to Pydantic BaseModel; fix ty…
Jun 25, 2026
2616477
Automatically reformatting code
Jun 25, 2026
da816d0
Generate VLM verification model from spec; move ask_vlm_verify to exp…
Jun 29, 2026
a38eb92
Automatically reformatting code
Jun 29, 2026
729ae3c
regen: run full make generate with Java — add VlmVerificationsApi + m…
Jun 29, 2026
89a91e3
fix: restore cli.py to main (only drop the ask_vlm command)
Jun 29, 2026
3c90b08
simplify: build vlm-verifications URL like create_note (no version st…
Jun 29, 2026
9d23339
fix: import VlmVerification from model, not groundlight (mypy)
Jun 29, 2026
c031c1b
fix: exclude ask_vlm_verify from CLI auto-registration
Jun 29, 2026
e574ae0
Add ask_vlm_verify as an experimental CLI command
Jun 29, 2026
1c4a496
docs: note CLI supports only a single image for now
Jun 29, 2026
532e3cb
Potential fix for pull request finding
srnangi Jun 29, 2026
4aa5e5e
Use generated VlmVerificationsApi instead of raw requests
Jun 29, 2026
7154463
Automatically reformatting code
Jun 29, 2026
09ed731
Use one-indexed media filenames (image_1.jpg, image_2.jpg, ...)
Jun 29, 2026
01ff309
docs(spec): trim vlm-verifications endpoint description
Jul 10, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
Update ask_vlm model_id docstring examples to vision-capable aliases 
Drop the gpt-5.4 example (OpenAI models on Bedrock are text-only and cannot
do image verification); use claude-sonnet-4.5 / nova-pro instead.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
  • Loading branch information
@claude
buildci and claude committed Jun 24, 2026
commit 2b20fce27d17cf5fb4b9a226739790121ef33fe7
6 changes: 3 additions & 3 deletions src/groundlight/client.py
View file
Open in desktop

@brandon-wada brandon-wada Jun 25, 2026

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Should probably go in experimental

Original file line number Diff line number Diff line change
Expand Up @@ -1154,9 +1154,9 @@ def ask_vlm(

:param query: Natural-language prompt describing what to verify, e.g.
``"Is there a fire visible in the image? Reason step by step."``
:param model_id: Friendly alias of the VLM to use, e.g. ``"gpt-5.4"`` or
``"claude-sonnet-4.5"``. Must be one of the models supported by the
server. Defaults to the server-configured default.
:param model_id: Friendly alias of the VLM to use, e.g.
Comment thread
srnangi marked this conversation as resolved.
Outdated
``"claude-sonnet-4.5"`` or ``"nova-pro"``. Must be one of the
models supported by the server. Defaults to the server-configured default.
:param timeout: Request timeout in seconds (default 15 s).

:return: :class:`VLMVerificationResult` with ``verdict`` (``"YES"`` / ``"NO"`` /
Expand Down