lightflow-text-to-image

# LightFlow Text To Image Use `lightflow.text_to_image` for the built-in preview image runtime. This workflow writes a deterministic preview PNG and is not a real diffusion model executor. ## Workflow - Workflow id: `lightflow.text_to_image` - Runtime ownership: this workflow crate owns its public `execute()` entry and exposes it through the `lightflow-text-to-image-runner` package binary. - Host engine: `package.command.v1`; discovery, help, and plan do not execute package source. - Input values are validated against their declared JSON types without string coercion. - Cargo package `lightflow-text-to-image` supplies this ID and its Cargo package version through `workflow!()`. - Input `prompt`: required positive prompt; widget `prompt`. - Input `negative`: optional negative prompt; default `""`; widget `textarea`. - Input `width`: optional integer; default `512`; range `64..2048`; step `8`; widget `number`. - Input `height`: optional integer; default `512`; range `64..2048`; step `8`; widget `number`. - Input `seed`: optional integer seed; widget `seed`. - Input `output_path`: optional destination PNG path; artifact kind `image`; widget `file_save`. - Input `model`: optional model variant id bound to `image_model`; choices `sdxl-gguf-q4` or `sdxl-safetensors`; widget `model_select`. - Outputs: `image` artifact metadata and `image_path`; artifact kind `image`; bound to `image_model`. - Runtime: `lightflow.image.generate` with engine `builtin.preview.v1`. - Default output directory: XDG Pictures under `lightflow/`. ## Usage ```bash lfw run lightflow.text_to_image \ -i prompt='"a quiet lake"' \ -i width=512 \ -i height=512 \ -i output_path='"out/lake.png"' ``` ## API Usage Start `lfw serve`, then call the workflow through the shared HTTP run contract. Adjust `inputs` to match the workflow contract above. ```bash curl -sS -X POST http://127.0.0.1:5174/workflows/lightflow.text_to_image/run \ -H 'content-type: application/json' \ -d '{"inputs":{}}' ```