imajin/orchestrators/imajin-pipeline

Image Pipeline Application

Self-contained image generation pipeline service using lilith-pipeline-framework.

Architecture

Built on the generic lilith-pipeline-framework with image-specific stages:

Request → Validate → Generate → Moderate → TextOverlay → Watermark → Quality → Output

Deployment

This is an APPLICATION, not a published package. Deploy as a standalone service.

# Install dependencies
pip install -e .

# Run service (FastAPI)
uvicorn image_pipeline.api.main:app --host 0.0.0.0 --port 8080

Usage

Direct Python Usage

from lilith_pipeline_framework import PipelineOrchestrator
from image_pipeline import DEFAULT_STAGES, ImagePipelineContext, ImagePipelineRequest

# Create request
request = ImagePipelineRequest(
    prompt="A beautiful sunset over mountains",
    model="photorealistic",
    layout="hero",
)

# Execute pipeline
orchestrator = PipelineOrchestrator(stages=DEFAULT_STAGES)
context = ImagePipelineContext(request=request)
result = await orchestrator.execute(context)

# Access generated image
image = result.data  # PIL.Image.Image

API Usage

curl -X POST http://localhost:8080/generate \
  -H "Content-Type: application/json" \
  -d '{
    "prompt": "A beautiful sunset",
    "model": "photorealistic",
    "layout": "square"
  }'

Migration from lilith-image-pipeline

Old Code:

from lilith_image_pipeline import ImagePipeline, PipelineRequest

pipeline = ImagePipeline()
result = await pipeline.execute(PipelineRequest(...))

New Code:

from lilith_pipeline_framework import PipelineOrchestrator
from image_pipeline import DEFAULT_STAGES, ImagePipelineContext, ImagePipelineRequest

orchestrator = PipelineOrchestrator(stages=DEFAULT_STAGES)
context = ImagePipelineContext(request=ImagePipelineRequest(...))
result = await orchestrator.execute(context)

Development

# Install dev dependencies
pip install -e ".[dev]"

# Run tests
pytest tests/

# Type checking
mypy --strict src/

License

MIT License