feat(cli): add video2yaml command to generate test scripts from video#2169
Open
feat(cli): add video2yaml command to generate test scripts from video#2169
Conversation
Deploying midscene with
|
| Latest commit: |
ada2091
|
| Status: | ✅ Deploy successful! |
| Preview URL: | https://9750da8a.midscene.pages.dev |
| Branch Preview URL: | https://feat-video-to-script.midscene.pages.dev |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: aa8f72b4d7
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| format: VideoScriptFormat, | ||
| ): string { | ||
| const ext = format === 'playwright' ? '.test.ts' : '.yaml'; | ||
| return inputPath.replace(/\.[^.]+$/, ext); |
There was a problem hiding this comment.
For extensionless video paths (for example midscene video2yaml ./recording), inputPath.replace(/\.[^.]+$/, ext) returns the original path unchanged, so outputPath becomes the same file as the input. The subsequent writeFileSync(outputPath, result.content, 'utf-8') then overwrites the source video with generated script content, which is destructive data loss in a realistic CLI usage scenario.
Useful? React with 👍 / 👎.
…n recordings Add a new CLI subcommand `midscene video2yaml` that extracts frames from a video file using ffmpeg and sends them to a VLM to generate runnable Midscene test scripts in YAML or Playwright format. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The npm packages @ffmpeg-installer/ffmpeg and @ffprobe-installer/ffprobe may install binaries without execute permission. Add chmod +x before use and verify executability with a test spawn, falling back to system binaries when the npm binary is not usable. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Long videos (>20 frames) are now automatically split into segments using ffmpeg scene detection, each segment analyzed independently by the VLM, then merged into a single coherent script. - Add segment-frames.ts with detectSceneChanges and segmentFrames - Add segment/merge prompts in video-to-yaml.ts - Short videos (<= 20 frames) still use single-call path unchanged - Add --max-frames-per-segment and --scene-threshold CLI options - Add 19 unit tests covering segmentation and CLI parsing - Extract prependWebConfigIfMissing to eliminate duplicate code - Add MAX_TOTAL_FRAMES=600 cap to prevent OOM on very long videos Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
quanru
force-pushed
the
feat/video-to-script
branch
from
9f9f47d to
eb7eff4
Compare
…LI args - Make maxFrames the global budget for both short and long video paths, capped by MAX_TOTAL_FRAMES (600) - Add validation for all numeric CLI parameters (fps, maxFrames, maxFramesPerSegment, sceneThreshold) with clear error messages - Add 5 new test cases for numeric validation edge cases Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Remove unused `format` parameter from `generateFromVideoSegment` - Remove redundant second sort in `detectSceneChanges` (already sorted in `parseSceneTimestamps`) Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
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.
Summary
midscene video2yamlCLI subcommand that generates runnable Midscene test scripts (YAML or Playwright) from screen recording videos@ffmpeg-installer/ffmpegwith system fallback)--format yaml|playwrightNew files
packages/core/src/ai-model/prompt/video-to-yaml.tspackages/cli/src/video/extract-frames.tspackages/cli/src/video/index.tsvideo2yaml()orchestrator: extract → analyze → writepackages/cli/src/video/cli.tsvideo2yamlsubcommandpackages/cli/tests/unit-test/video-cli.test.tspackages/core/tests/unit-test/video-to-yaml.test.tsUsage
Test plan
pnpm run lintpassesnpx nx build @midscene/clibuilds successfullynpx vitest --run tests/unit-test/video-cli.test.ts— 8 tests passnpx vitest --run tests/unit-test/video-to-yaml.test.ts— 2 tests passpage.goto()for navigation🤖 Generated with Claude Code