No description
084a0b7ea0
Fix validation error where LLM returns nested dict for "reasoning" field instead of string. Add defensive coding to handle unexpected LLM response formats and fix URL path duplication. Changes: - services/*/src/api/main.py: Coerce dict reasoning to JSON string - services/*/src/config.py: Revert LLMConfig to BaseModel for proper env var merging in load_config() - packages/cli/imajin: Fix LLM_SERVICE_URL to pass base URL only Fixes: - Validation error: "power_reasoning should be string, got dict" - URL duplication: /v1/chat/completions/v1/chat/completions - Add fallback reasoning for error cases Note: Environment variable override for LLM_SERVICE_URL still needs investigation - config may be cached before subprocess env vars are set. Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com> |
||
|---|---|---|
| docs | ||
| imajin | ||
| infrastructure | ||
| node_modules | ||
| orchestrators/imajin-pipeline | ||
| packages | ||
| scripts | ||
| services | ||
| tests | ||
| tooling/claude | ||
| .npmrc | ||
| CLAUDE.md | ||
| package-lock.json | ||
| package.json | ||
| pnpm-lock.yaml | ||
| README.md | ||
| run | ||
Imajin (Imagen Djinni)
AI-powered image generation platform with multi-service architecture.
Architecture
imajin/
├── services/ # ML Services (GPU-accelerated)
│ ├── imajin-prompt/ # LLM prompt generation (Port 8003)
│ ├── imajin-diffusion/ # SDXL/Diffusers (Port 8002)
│ └── imajin-processing/ # CPU post-processing (Port 8004)
│
├── orchestrators/ # Composition Layer (no GPU)
│ └── imajin-pipeline/ # 7-stage pipeline (Port 8080)
│
├── packages/ # Published Libraries (@lilith/imajin-*)
│ ├── imajin-app/ # Core configs & types
│ ├── imajin-react/ # React components
│ ├── imajin-electron/ # Desktop integration
│ └── imajin-client/ # Unified HTTP client
│
├── tests/ # Integration tests
├── scripts/ # Build & dev scripts
├── tooling/ # Claude config
└── docs/ # Documentation
ML Construction Kit Pattern
Services own ONE model type. Orchestrators compose services via HTTP.
┌──────────────────────────────────────────────────────────────────┐
│ @packages/@ml/ ← Building blocks (libraries) │
│ ├── model-boss/ ← GPU coordination │
│ └── pipeline-framework/ ← Stage orchestration │
│ │
│ @applications/@ml/imajin/ ← This application │
│ ├── services/ ← ML services with model-boss │
│ └── orchestrators/ ← HTTP orchestration, no GPU │
└──────────────────────────────────────────────────────────────────┘
Quick Start
# Install dependencies
npm install
# Build all packages
npm run build
# Start services (requires GPU for diffusion/prompt services)
cd services/imajin-diffusion/service && uvicorn src.api.main:app --port 8002
cd services/imajin-prompt/service && uvicorn src.api.main:app --port 8003
cd services/imajin-processing && npm run dev
cd orchestrators/imajin-pipeline && uvicorn src.image_pipeline.api.main:app --port 8080
Package Names
| Package | Description |
|---|---|
@lilith/imajin-app |
Core configs, pipelines, prompts |
@lilith/imajin-react |
React UI components |
@lilith/imajin-electron |
Desktop integration |
@lilith/imajin-client |
Unified HTTP client |
@lilith/imajin-prompt-types |
Prompt service types |
@lilith/imajin-prompt-client |
Prompt service client |
@lilith/imajin-diffusion-types |
Diffusion service types |
@lilith/imajin-diffusion-client |
Diffusion service client |
@lilith/imajin-processing-types |
Processing service types |
@lilith/imajin-processing-client |
Processing service client |
Python Packages
| Package | Description |
|---|---|
imajin-prompt-service |
LLM prompt generation (FastAPI) |
imajin-diffusion-service |
SDXL image generation (FastAPI) |
imajin-pipeline |
7-stage orchestrator (FastAPI) |
Port Assignments
| Service | Port |
|---|---|
| imajin-diffusion | 8002 |
| imajin-prompt | 8003 |
| imajin-processing | 8004 |
| imajin-pipeline | 8080 |
License
MIT