feat: add GetVideoLastFrame node (CORE-123)

[bigcat88]

No description provided.

[coderabbitai]

📝 Walkthrough

Walkthrough

This pull request adds a new GetVideoLastFrame node to the Comfy video extension. The node extracts and returns the last visible frame from a video file or video components as a normalized image tensor. The implementation includes a fast path for VideoFromFile inputs that performs seeking and direct frame decoding, with fallback to the existing get_components() path for trimmed or complex cases. Comprehensive unit tests validate frame extraction across various video formats, trimming behavior, tensor contract compliance, and error handling.

🚥 Pre-merge checks | ✅ 3 | ❌ 2

❌ Failed checks (1 warning, 1 inconclusive)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 40.74% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.
Description check	❓ Inconclusive	No pull request description was provided by the author, making it impossible to assess relevance to the changeset.	Add a description explaining the purpose and use case for the GetVideoLastFrame node to provide context for reviewers.

✅ Passed checks (3 passed)

Check name	Status	Explanation
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Title check	✅ Passed	The title clearly and concisely describes the main change: adding a new GetVideoLastFrame node to the codebase.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 1

🧹 Nitpick comments (1)

tests-unit/comfy_extras_test/nodes_video_test.py (1)

298-301: Tighten the corrupt-input expectation.

pytest.raises(Exception) is very broad and can pass on unrelated regressions. Prefer asserting a narrower failure condition (specific error text or bounded exception set) so this test catches the right failure mode.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@tests-unit/comfy_extras_test/nodes_video_test.py` around lines 298 - 301,
Change the test to assert a narrower failure: replace pytest.raises(Exception)
in test_corrupt_bytes_raises with a specific exception type thrown by
GetVideoLastFrame.execute for invalid input (e.g., pytest.raises(ValueError))
and optionally add a match argument to assert the error message contains a
distinguishing substring like "corrupt" or "decode"; locate this in the test
function test_corrupt_bytes_raises which constructs the bad input via
InputImpl.VideoFromFile(io.BytesIO(...)) and calls
GetVideoLastFrame.execute(video).

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@comfy_extras/nodes_video.py`:
- Around line 200-203: The fixed 0.003s tolerance used when comparing
raw_duration_us/av.time_base to reported_duration_s can mis-detect trims for
high-FPS sources; replace the magic constant with a frame-rate-aware tolerance
by querying the stream frame rate (e.g., use av.average_rate or av.rate),
compute frame_duration_s = 1.0 / float(frame_rate) (fallback to a sensible
default if unavailable), then set threshold = min(0.003, 0.5 * frame_duration_s)
and use that threshold in the existing comparison using raw_duration_us,
av.time_base and reported_duration_s so tiny per-frame offsets at high FPS are
handled correctly.

---

Nitpick comments:
In `@tests-unit/comfy_extras_test/nodes_video_test.py`:
- Around line 298-301: Change the test to assert a narrower failure: replace
pytest.raises(Exception) in test_corrupt_bytes_raises with a specific exception
type thrown by GetVideoLastFrame.execute for invalid input (e.g.,
pytest.raises(ValueError)) and optionally add a match argument to assert the
error message contains a distinguishing substring like "corrupt" or "decode";
locate this in the test function test_corrupt_bytes_raises which constructs the
bad input via InputImpl.VideoFromFile(io.BytesIO(...)) and calls
GetVideoLastFrame.execute(video).

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: ece8b7bb-bfd0-46b1-85ce-ef74d8c60228

📥 Commits

Reviewing files that changed from the base of the PR and between 24de8dc and ddb52f4.

📒 Files selected for processing (2)

• comfy_extras/nodes_video.py
• tests-unit/comfy_extras_test/nodes_video_test.py

[alexisrolland]

As discussed, this can be managed with a couple of nodes packaged as blueprints. We will add it to the next blueprints updates: #13611

Example:

Blueprint: