git 730baa8733 Render 3D Q3 player head and blit to HUD
Add a new workflow step that renders the Quake3 MD3 player head to a small offscreen GPU render target each frame and exposes it to the overlay as "overlay.head_gpu_tex" with face rectangle metadata. Implement WorkflowQ3HudHeadStep (kHeadSz=64) and its header; register the step in the workflow registrar and add it to the CMake build and q3_overlay workflow JSON. Update q3 HUD drawing to use the live 3D portrait when available (fall back to the icon otherwise). Extend overlay.sw.end to blit the head texture into the HUD face rect (BlitHeadPortrait), with lazy allocation of a vertex buffer and sampler and proper cleanup. Also add resource release logic and small API includes where needed.
2026-05-04 19:50:17 +01:00
2026-03-09 22:30:41 +00:00
2026-03-09 22:30:41 +00:00
2026-03-09 22:30:41 +00:00
2026-03-09 22:30:41 +00:00
2026-03-09 22:30:41 +00:00
2026-03-09 22:30:41 +00:00
2026-03-09 22:30:41 +00:00
2026-03-09 22:30:41 +00:00
2026-01-14 01:41:59 +00:00
2026-03-09 22:30:41 +00:00

MetaBuilder

Philosophy: 95% JSON config, 5% TypeScript/C++ infrastructure Scale: 27,826+ files across 34 directories Status: Production-ready — Phase 2 complete


What's Running

Pastebin — Code Snippet Manager

Full-stack app: Next.js frontend + Flask auth backend + C++ DBAL API

http://localhost/pastebin          # UI
http://localhost/pastebin-api      # Flask auth (register, login, Python runner)
http://localhost:8080              # DBAL C++ REST API (entities)

Test accounts (seeded on first startup):

User Password Namespaces Snippets
demo demo1234 Default, Python Recipes, SQL Patterns, Utilities 11
alice alice1234 Default, React Components, CSS Tricks, JS Utilities 9
bob bob12345 Default, Go Patterns, Bash Scripts, API Design 8

Quick Start

# Deploy full stack
cd deployment
docker compose -f compose.yml up -d

# Build & deploy a specific app
python3 deployment.py build apps --force dbal pastebin

# Rebuild base images (rare)
python3 deployment.py build base

Architecture

Browser
  └── Next.js (React + Redux + IndexedDB)
        └── Flask (auth, Python runner, JWT)
              └── DBAL C++ daemon (REST API, 14 DB backends)
                    └── PostgreSQL (prod) / SQLite (dev)

DBAL event flow (user registration → automatic seeding):

POST /User  →  pastebin.User.created  →  WfEngine (detached thread)
                  └── on_user_created.json workflow
                        ├── dbal.uuid × 7
                        ├── dbal.timestamp
                        ├── dbal.entity.create → Namespace "Default"
                        ├── dbal.entity.create → Namespace "Examples"
                        └── dbal.entity.create × 5 → seed snippets

Key Directories

metabuilder/
├── dbal/
│   ├── production/          # C++ daemon (Drogon HTTP, 14 DB backends, JWT auth)
│   │   ├── src/workflow/    # Event-driven workflow engine (7 step types)
│   │   ├── src/auth/        # JWT validation + JSON ACL config
│   │   └── build-config/    # Dockerfile, CMakeLists, conanfile
│   └── shared/
│       ├── api/schema/      # JSON entity schemas (source of truth)
│       │   ├── entities/    # 39 entity definitions
│       │   ├── events/      # event_config.json → workflow mappings
│       │   └── workflows/   # on_user_created.json etc.
│       └── seeds/database/  # Declarative JSON seed data
├── frontends/pastebin/
│   ├── src/                 # Next.js app (React + Redux + FakeMUI)
│   └── backend/             # Flask auth + Python runner
├── workflow/                # DAG workflow engine (TS, Python, C++ plugins)
├── gameengine/              # SDL3/bgfx 2D/3D engine (27/27 tests)
├── fakemui/                 # Material Design clone (167 components)
├── packages/                # 62 feature packages
├── deployment/              # Docker compose stack + build scripts
└── CLAUDE.md                # AI assistant guide (read first)

DBAL Backends (14)

Adapter Backend
memory In-process (testing)
sqlite SQLite3
postgres PostgreSQL
mysql / mariadb MySQL / MariaDB
cockroachdb CockroachDB
mongodb MongoDB
redis Redis (cache layer)
elasticsearch Elasticsearch (search layer)
cassandra Cassandra
surrealdb SurrealDB
supabase Supabase REST
prisma Prisma ORM bridge

Switch adapter at runtime: DATABASE_URL=sqlite://:memory: DBAL_ADAPTER=sqlite


Development

# Frontend (pastebin)
cd frontends/pastebin && npm run dev

# DBAL logs
docker logs -f metabuilder-dbal

# Admin seed endpoint (force re-seed)
curl -X POST http://localhost:8080/admin/seed \
  -H "Authorization: Bearer $DBAL_ADMIN_TOKEN" \
  -d '{"force": true}'

# Search docs / reports
cd txt && python3 reports.py search "topic"
cd docs && python3 docs.py search "topic"

Schema-First Development

Entity schemas live in dbal/shared/api/schema/entities/ as JSON. They are the single source of truth — consumed by:

  • C++ DBAL daemon (table creation, CRUD validation)
  • gen_types.py (TypeScript + C++ type generation)
  • Seed loader (entity-aware bulk insert)
  • Workflow engine (entity creation steps)
# Regenerate types after schema changes
python3 dbal/shared/tools/codegen/gen_types.py

Last Updated: 2026-03-04

S
Description
No description provided
Readme 452 MiB
Languages
TypeScript 30.3%
JavaScript 19.1%
Jupyter Notebook 12.5%
C++ 10.3%
SCSS 8.7%
Other 18.7%