Repository navigation
feat(live): add ZIP_STORED visual frame pack and unpack helpers - #7562
Merged
Merged
Conversation
Live video reaches ADK as `image/*` frames, alongside `audio/*` chunks. Saving each frame as its own artifact costs one storage operation per frame, while adding a batch method to `BaseArtifactService` would grow the abstract storage interface for every backend. This change adds `google.adk.live._visual_frames`, which packs a frame sequence into one uncompressed (`zipfile.ZIP_STORED`) `.zip` archive that the existing `save_artifact()` / `load_artifact()` methods store as is: - `VisualFrame`: a Pydantic model pairing a frame's `types.Blob` with the wall-clock time it was received; a NaN or infinite time is rejected when the frame is created. - `pack_visual_frames()`: validates the frames (non-empty data, non-decreasing timestamps), writes `frames/frame_NNNN.<ext>` members and a versioned `metadata.json` manifest whose millisecond fields all derive from the same rounded times, and returns `(archive_bytes, artifact_metadata)`, where the metadata is the manifest without its per-frame index, small enough for an artifact's `custom_metadata`. - `unpack_visual_frames()`: the inverse; restores `(list[VisualFrame], manifest)` from the archive bytes, with each time to the nearest millisecond. It reads only stored, unencrypted members, each listed frame once, and rejects an unknown `formatVersion`, so every failure is a `ValueError`. - `VISUAL_FRAMES_ZIP_MIME_TYPE`: the archive's MIME type, `application/zip`. PiperOrigin-RevId: 1001035243
copybara-service
Bot
force-pushed
the
copybara/982888971
branch
from
October 16, 2026 19:44
bf745ac to
17d8a5a
Compare
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
1 participant
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.
feat(live): add ZIP_STORED visual frame pack and unpack helpers
Live video reaches ADK as
image/*frames, alongsideaudio/*chunks. Saving each frame as its own artifact costs one storage operation per frame, while adding a batch method toBaseArtifactServicewould grow the abstract storage interface for every backend.This change adds
google.adk.live._visual_frames, which packs a frame sequence into one uncompressed (zipfile.ZIP_STORED).ziparchive that the existingsave_artifact()/load_artifact()methods store as is:VisualFrame: a Pydantic model pairing a frame'stypes.Blobwith its capture timestamp.pack_visual_frames(): validates the frames (non-empty data, finite and non-decreasing timestamps), writesframes/frame_NNNN.<ext>members and ametadata.jsontiming manifest, and returns(archive_bytes, manifest).unpack_visual_frames(): the inverse; restores(list[VisualFrame], manifest)from the archive bytes.summarize_manifest(): the manifest without its per-frame index, small enough for an artifact'scustom_metadata.VISUAL_FRAMES_ZIP_MIME_TYPE: the archive's MIME type,application/zip.NO_UNIT_GUIDE=Internal pack/unpack helper module for live visual frames