Datasource: File Stream
Use File Stream to replay existing media as datasource events.
This is ideal for deterministic testing, regression validation, and demos where live camera input is not required.
Best Use Cases
- QA and repeatable model testing
- Offline playback of captured incidents
- Development without live camera dependencies
Configuration Fields
| Field | Required | Type | Default | Purpose |
|---|---|---|---|---|
sourceSelector | UI helper | media-source-selector | - | Frontend selector for source media |
sourceType | Hidden | string | - | Internal source classification |
bucketId | Yes | string | - | Bucket containing media object |
objectKey | Yes | string | - | Path/key to selected media object |
fps | No | number | 5 | Frame rate (1-8) |
resolution | No | select | 640x480 | Output frame size |
iterations | No | number | 1 | Repeat count (-1 for infinite) |
sleepBetweenIterationsSec | No | number | 0 | Delay between loops |
How to Configure Correctly
- Select File Stream datasource type.
- Use media selector to choose bucket/object.
- Set FPS and resolution to match model input needs.
- Set iterations for test or continuous playback.
- Save and trigger workflow with datasource node.
- Validate reproducibility across repeated runs.
Example Config
{
"bucketId": "1766071881123456",
"objectKey": "media/test/scenario_a.mp4",
"fps": 4,
"resolution": "640x480",
"iterations": 2,
"sleepBetweenIterationsSec": 1
}
Practical Guidance
- Use fixed clips as benchmark fixtures
- Keep test clips versioned with known outcomes
- Avoid infinite loops in shared environments unless explicitly required
Common Problems
- No frames: incorrect bucket key or inaccessible media object
- Inconsistent output: variable source files across test runs
- Resource spikes: high FPS with large resolution and long iterations
Related Pages
- Continue with Storage Buckets for media organization practices.
- Continue with Workflows for regression and validation patterns.