Lilith a5f99bb3d7 chore(imajin): clean up legacy structure and completion markers

- Remove old imajin/ directory (migrated to services/ + orchestrators/)
- Delete completion markers (DONE.md, INTEGRATION-COMPLETE.md, TESTING.md)
- Remove standalone test generation scripts
- Update docs to reflect current architecture
- Add multi-base-strategy.md documentation

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

2026-01-16 17:01:10 -08:00

3.2 KiB

Raw Blame History

Configuration

Configuration patterns used across the @imajin platform.

Two-Layer YAML Configuration

Services use a layered configuration approach:

config.defaults.yaml  →  Shipped defaults (git tracked)
        ↓
config.yaml          →  Deployment overrides (gitignored)
        ↓
API Parameters       →  Per-request overrides (future)

config.defaults.yaml

Version-controlled defaults shipped with the service:

# imajin-prompt/config.defaults.yaml
llm:
  backend: "llama"
  model_id: "deepseek-r1-70b"
  context_size: 4096
  gpu_layers: 40

gpu:
  enabled: true
  redis_url: redis://localhost:6379
  priority: "normal"

service:
  port: 8003
  host: 0.0.0.0

content:
  maturity_min: "suggestive"
  maturity_max: "mature"
  maturity_default: "suggestive"

config.yaml

Deployment-specific overrides (gitignored):

# imajin-prompt/config.yaml
llm:
  context_size: 2048  # Reduced for smaller GPU

gpu:
  redis_url: redis://prod-redis:6379
  priority: high

service:
  port: 8003

Config File Search Order

./config.yaml (service directory)
$(pwd)/config.yaml (current working directory)
~/.config/{service}/config.yaml (user config)

Environment Variables

Environment variables override YAML configuration:

imajin-prompt

Variable	Config Path	Description
`SERVICE_PORT`	`service.port`	Service port
`SERVICE_HOST`	`service.host`	Bind address
`GPU_ENABLED`	`gpu.enabled`	Enable GPU
`REDIS_URL`	`gpu.redis_url`	Redis connection

imajin-diffusion

Variable	Config Path	Description
`IMAGE_GEN_PORT`	-	Service port
`IMAGE_GEN_HOST`	-	Bind address
`IMAGE_GEN_DEBUG`	-	Debug mode
`IMAGE_GEN_PHOTOREALISTIC_DEVICE`	-	GPU device
`IMAGE_GEN_ANIME_DEVICE`	-	GPU device
`IMAGE_GEN_MODEL_CACHE_DIR`	-	Model cache

Port Configuration

.ports.json

Services with desktop integration use .ports.json:

{
  "default": 8002,
  "desktop": 41227
}

services.yaml

Canonical port definitions in imajin-app/services.yaml:

ports:
  imajin-app: 3010
  imajin-prompt: 8003
  imajin-diffusion: 8002
  imajin-processing: 8004

environments:
  development:
    imagegen_assistant_url: http://localhost:8003
  production:
    imagegen_assistant_url: https://imajin-prompt.atlilith.com

Configuration Precedence

From lowest to highest priority:

Hardcoded defaults in code
config.defaults.yaml
config.yaml
Environment variables
API request parameters (where supported)

Example: Full Production Config

# imajin-prompt/config.yaml
llm:
  backend: "llama"
  model_id: "deepseek-r1-70b"
  context_size: 4096
  gpu_layers: -1  # Use all GPU layers

gpu:
  enabled: true
  redis_url: redis://prod-redis:6379
  priority: high

service:
  port: 8003
  host: 0.0.0.0

content:
  maturity_min: "sfw"
  maturity_max: "explicit"
  maturity_default: "suggestive"

imajin-prompt/CONFIG.md - Detailed config guide
GPU Coordination - Redis/GPUBoss setup

3.2 KiB Raw Blame History