Platform Overview
The AESOP strategic decision support platform — five Django microservices: the CALCHAS trilogy (intell · modelisation · simulation) plus a shared energy currency (aesop_energy) and a structural knowledge graph (aesop_tree).
Architecture Diagram
Services Detail
aesop_intell (XXX)
OSINT intelligence collection and synthesis. Gathers, enriches, and structures open-source intelligence for downstream consumption.
28 Django apps
PostgreSQL Apache AGE pgvector sentence-transformers Mistral AI H3 Celery+Redis HTMX Leaflet.js
aesop_modelisation (MIDAS)
Critical infrastructure modeling. Maps and simulates dependencies across vital sectors to evaluate cascade effects.
8 Django apps
app_midas_core— Core models and utilitiesapp_midas_pnaiv— 5 PNAIV sectorsapp_midas_is— IS graphs (nodes, edges, dependencies)app_midas_geo— H3 hex tiling + altitude layersapp_midas_propagation— Cascade engineapp_midas_injection— 3 disruption modesapp_midas_dashboard— Visualization dashboardapp_midas_api— REST API
PostgreSQL Apache AGE H3 Celery+Redis Leaflet HTMX
aesop_simulation (AESOP)
Military wargame simulation. Models belligerents, units, orders, and combat through a 10-step OODA turn engine with Monte Carlo resolution.
11 Django apps
app_sim_core— Core simulation engineapp_sim_belligerant— Will + Capability modelingapp_sim_unit— 6 VE + 8 VC tactical unitsapp_sim_tabord— TABORD ordersapp_sim_turn— 10-step OODA loopapp_sim_combat— Monte Carlo combat resolutionapp_sim_infops— Information warfareapp_sim_c2— Command chain (AGE graph)app_sim_map— H3 battlefield mapapp_sim_montecarlo— Monte Carlo engineapp_sim_api— REST + WebSocket API
PostgreSQL Apache AGE H3 Celery+Redis Django Channels NumPy/SciPy Leaflet HTMX
Port, Database & Redis Allocation
| Service | Codename | Port | PostgreSQL Database | Redis Databases | Django Apps |
|---|---|---|---|---|---|
| aesop_intell | XXX | 8000 |
aesop_intell |
DB 0 (Celery broker), DB 1 (cache) |
28 |
| aesop_modelisation | MIDAS | 8001 |
aesop_modelisation |
DB 2 (Celery broker), DB 3 (cache) |
8 |
| aesop_simulation | AESOP | 8002 |
aesop_simulation |
DB 4 (Celery broker), DB 5 (cache), DB 6 (Channels) |
11 |
| aesop_energy | — | 8008 |
aesop_energy |
DB 7 (Celery broker), DB 8 (results) |
4 |
| aesop_tree | — | 8003 |
aesop_tree (AGE + pgvector + pg_search) |
DB 9 (Celery broker), DB 10 (results) |
9 |
The CALCHAS Trilogy
CALCHAS is the overarching strategic concept that binds the three AESOP microservices into a coherent decision-support pipeline. Named after the mythological Greek seer who guided the Achaeans during the Trojan War, CALCHAS represents the platform's ambition: to see clearly through the fog of complex, multi-domain crises.
Stage 1 — XXX (aesop_intell) serves as the eyes and ears of the platform. It continuously collects, enriches, and synthesizes open-source intelligence (OSINT) from diverse feeds. Using sentence-transformers for semantic embedding (via pgvector), Mistral AI for natural-language synthesis, and Apache AGE for knowledge-graph construction, XXX transforms raw information into structured, geolocated, and interconnected intelligence products. These enriched OSINT datasets are then pushed downstream to MIDAS.
Stage 2 — MIDAS (aesop_modelisation) receives intelligence from XXX and uses it to build and maintain a living model of critical infrastructure. Organized around the 5 PNAIV sectors (Political, Normative, Administrative, Infrastructure, Vital), MIDAS constructs dependency graphs of information systems — nodes, edges, and weighted dependencies. Its cascade propagation engine can simulate how a disruption in one infrastructure element ripples through the network. The resulting "infrastructure chessboard" — a complete, scored, and geographically tiled model — is made available to AESOP for wargaming.
Stage 3 — AESOP (aesop_simulation) sits at the apex of the trilogy. It takes the infrastructure chessboard from MIDAS and overlays a full military wargame engine on top of it. Belligerents are modeled with Will (political resolve) and Capability (military strength). Tactical units — 6 types of VE (live forces) and 8 types of VC (cyber/virtual forces) — receive TABORD orders and execute them through a 10-step OODA turn cycle. Combat is resolved via Monte Carlo simulation. Information operations (INFOPS) model influence and perception warfare. The command chain is represented as a graph in Apache AGE, and the battlefield is projected onto H3 hexagonal tiles. Real-time updates flow through Django Channels and WebSocket connections.
The feedback loop: simulation results from AESOP can feed back into XXX, creating a closed analytical loop. Wargame outcomes inform intelligence priorities, which in turn refine the infrastructure models, which produce better simulation scenarios. This cyclical flow is the core strength of the CALCHAS trilogy — each service amplifies the value of the others.
Supporting Services
aesop_energy provides a shared energy currency — a watt-hour scalar per H3 cell that blends infrastructure, force and information into one comparable measure of a cell's importance, while preserving provenance. It ships both a pure roll-up library (scale↔resolution, imported by the other services via
-e ../aesop_energy) and a COP Django project with a Leaflet H3 map and a military energy calculator in Joules.
aesop_tree captures the structural knowledge of a "system considered": the technology, energy, political and ideology trees, in one Apache AGE property graph. It joins abstract "science" recipe links (C4 needs RDX needs nitrogen) to concrete "logistic" sourcing links (instances ingested from aesop_intell) via the
satisfied_by bridge. Where a requirement has no known supplier, the gap engine emits an intelligence orientation (RFI) back to aesop_intell (POST /api/intell/orientation/) — closing the OODA loop in the other direction: structure drives collection, collection fills structure. See the aesop_tree page for detail.