# Battery ERP — Material, cell, pack, and supply-chain management with real-time
# commodity pricing and Fabric Lakehouse analytics.
#
# Covers the full battery value chain: lithium, cobalt, nickel, manganese, graphite
# through cell chemistries (NMC-811, NCA, LFP, LMO) to battery packs with BOM costing,
# supplier scoring, inventory management, and what-if cost scenarios.

[![Python](https://img.shields.io/badge/python-3.10%2B-blue)](https://www.python.org/)
[![PyPI](https://img.shields.io/pypi/v/battery-erp)](https://pypi.org/project/battery-erp/)
[![License](https://img.shields.io/badge/license-MIT-green)](LICENSE)
[![Tests](https://img.shields.io/badge/tests-48%2B%20passing-brightgreen)](tests/)
[![Fabric](https://img.shields.io/badge/Microsoft-Fabric-0078D4)](https://fabric.microsoft.com)

mcp-name: io.github.icohangar-ops/battery-erp

---

## What this is

Battery ERP manages the **complete battery value chain** — from raw material sourcing
through cell manufacturing to pack assembly. Every cost is traceable to a specific
material, supplier, and price point.

| Layer | Role |
|---|---|
| **Data Models** | RawMaterial, CellChemistry, BatteryCell, BatteryPack, BOMItem, Supplier, InventoryRecord, PurchaseOrder, ManufacturingBatch |
| **Business Rules** | BOM cost rollups, inventory status management, supplier scoring (composite A-D grade), manufacturing yield tracking, price trend analysis, what-if cost scenarios |
| **Pricing Engine** | Default material price table (20+ materials), AlphaVantage integration for live commodity prices, FRED macro overlay |
| **Analytics** | Inventory health reports, supply chain reports, manufacturing yield reports, chemistry cost comparison dashboards |
| **Fabric Lakehouse** | 11 Delta tables for persistent storage and SQL analytics |

---

## Quick start

```bash
# From PyPI — https://pypi.org/project/battery-erp/
python3 -m pip install 'battery-erp[mcp]'   # MCP tools
python3 -m pip install 'battery-erp[api]'   # REST adapter

# From source
git clone https://github.com/icohangar-ops/battery-erp.git
cd battery-erp
python3 -m pip install -e '.[dev]'
PYTHONPATH=src python3 -m pytest tests/ -v

# Domain modules
python3 -c "
from battery_erp.pricing import calculate_cell_cost_summary, get_material_price_table
prices = get_material_price_table()
for chem in ['NMC-811', 'NMC-622', 'NCA', 'LFP', 'LMO']:
    r = calculate_cell_cost_summary(chem, 50.0, prices)
    print(f'{chem}: \${r[\"cost_per_kwh\"]:.1f}/kWh (BOM: \${r[\"bom_cost_usd\"]:.2f})')
"
```

---

## Architecture

```
                    ┌──────────────────────────────────────┐
                    │  Raw Materials (20+ tracked)          │
                    │  Lithium · Cobalt · Nickel · Mn · Gr  │
                    └──────────────┬───────────────────────┘
                                   │ BOM
                    ┌──────────────▼───────────────────────┐
                    │  Cell Chemistries                     │
                    │  NMC-811 · NMC-622 · NCA · LFP · LMO  │
                    └──────────────┬───────────────────────┘
                                   │ cells + components
                    ┌──────────────▼───────────────────────┐
                    │  Battery Packs                        │
                    │  EV · ESS · Consumer · Industrial     │
                    └──────────────────────────────────────┘

Side modules:
  ┌─────────────────┐  ┌──────────────────┐  ┌──────────────────┐
  │ Supplier Scoring │  │ Inventory Mgmt   │  │ Cost Scenarios    │
  │ Composite 0-100  │  │ Reorder logic    │  │ What-if analysis  │
  │ A/B/C/D grades   │  │ Status tracking  │  │ Price shock model │
  └─────────────────┘  └──────────────────┘  └──────────────────┘
```

---

## Core modules

### `battery_erp.core.models`
All domain dataclasses:
- `RawMaterial` — material catalog with pricing, HS codes, hazards
- `CellChemistry` — NMC-111/622/811, NCA, LFP, LMO with energy density and cycle life
- `BatteryCell` — cell specs (capacity, voltage, form factor, weight)
- `BatteryPack` — pack assembly (cells + BMS + thermal)
- `BOMItem` — bill of materials line item with waste factor
- `Supplier` — supplier catalog with quality rating, lead time, certifications
- `InventoryRecord` — warehouse positions with reorder logic
- `PurchaseOrder` — PO lifecycle tracking
- `ManufacturingBatch` — production batch yield tracking
- `PriceHistory` — commodity price time series

### `battery_erp.core.rules`
Deterministic business rules:
- `rollup_bom_cost()` — total BOM cost with material breakdown and waste cost
- `calculate_cell_bom()` — generate representative BOM for any chemistry
- `calculate_pack_bom()` — pack-level BOM (cells + casing + BMS + cooling)
- `update_inventory_status()` — recalculate in_stock/low/out_of_stock
- `check_reorder_suggestions()` — generate PO suggestions
- `calculate_batch_metrics()` — aggregate manufacturing yield
- `analyze_price_history()` — price trend analysis with volatility
- `estimate_cell_cost_impact()` — what-if cost scenario modeling
- `calculate_pack_metrics()` — pack energy density and efficiency

### `battery_erp.supply_chain`
Supply chain management:
- `score_supplier()` — composite score (quality 35%, OTD 35%, lead time 20%, certs 10%)
- `rank_suppliers()` — rank by score, filter by material
- `create_purchase_order()` — PO creation from supplier data
- `analyze_po_pipeline()` — PO pipeline analysis (overdue detection, lead time tracking)
- `suggest_dual_sourcing()` — dual-sourcing strategy recommendation

### `battery_erp.pricing`
Commodity pricing:
- `get_material_price_table()` — default prices for 20+ battery materials
- `calculate_cell_cost_summary()` — quick cost estimate per chemistry
- `update_prices_from_alpha_vantage()` — live commodity price fetch
- `update_prices_from_fred()` — macro economic indicators

### `battery_erp.analytics`
Reporting:
- `generate_inventory_report()` — full inventory health dashboard
- `generate_supply_chain_report()` — supplier + PO pipeline report
- `generate_manufacturing_report()` — yield metrics
- `generate_pricing_report()` — chemistry cost comparison + price trends

---

## Chemistry cost comparison (default prices, 50Ah cell)

| Chemistry | BOM Cost | $/kWh | Cathode % | Key feature |
|---|---|---|---|---|
| **LFP** | Lowest | ~$50-55 | ~35% | No Co/Ni, ultra-safe, 4000+ cycles |
| **LMO** | Low | ~$55-60 | ~40% | Low cost, power tools |
| **NMC-111** | Medium | ~$70-80 | ~50% | Balanced, legacy |
| **NMC-622** | Medium | ~$75-85 | ~48% | Good energy-cost balance |
| **NMC-811** | Higher | ~$80-90 | ~52% | High energy, EV dominant |
| **NCA** | Highest | ~$85-95 | ~55% | Tesla flagship, 270 Wh/kg |

---

## Microsoft Fabric Integration

### Fabric Notebooks

| Notebook | Purpose |
|---|---|
| `fabric_setup_lakehouse.py` | Create all 11 Delta tables with seed data |
| `fabric_cost_dashboard.py` | Full cost analytics dashboard (chemistry comparison, pack costing, inventory, suppliers, price trends, scenarios) |

### Delta Table Schema

| Table | Key Columns |
|---|---|
| `raw_materials` | material_id, name, category, unit_price_usd, price_source, hs_code |
| `cell_chemistries` | chemistry_id, name, cathode_type, energy_density_wh_per_kg, cycle_life |
| `battery_cells` | cell_id, sku, chemistry, form_factor, nominal_capacity_ah, energy_wh, weight_kg |
| `battery_packs` | pack_id, sku, cell_sku, total_cells, nominal_capacity_kwh, pack_weight_kg |
| `bill_of_materials` | bom_id, parent_sku, material_name, quantity_per_unit, unit_cost_usd, waste_factor_pct |
| `suppliers` | supplier_id, name, country, materials_supplied, quality_rating, lead_time_days |
| `inventory` | record_id, sku, material_name, quantity_on_hand, quantity_reserved, reorder_point |
| `purchase_orders` | po_id, po_number, supplier_name, quantity, total_usd, status, expected_delivery |
| `price_history` | material_name, price_usd, as_of, source |
| `manufacturing_batches` | batch_id, product_sku, chemistry, quantity_produced, quantity_pass, yield_pct |
| `cost_scenarios` | scenario_id, scenario_name, material_name, current_price_usd, scenario_price_usd, pct_change |

### Fabric Quick Start

1. Run `fabric_setup_lakehouse.py` to create all 11 Delta tables
2. Run `fabric_cost_dashboard.py` for the full analytics dashboard
3. Dashboard covers: chemistry cost comparison, pack-level costing, inventory health, supplier scorecard, price trends, manufacturing yield, cost scenarios

---

## Inventory adapters (REST + MCP)

Battery ERP is a **domain library**, not a hosted ERP. For SMS / text-line and
agent workflows, a shared `InventoryService` sits under both a thin REST API and
an MCP server. Full detail: **[docs/INTEGRATION.md](docs/INTEGRATION.md)**.

```
SMS / ClickSend          Cursor / Claude (stdio MCP)
        │                         │
        ▼                         ▼
  REST API (port 8088)      battery_erp.mcp
        └──────────┬──────────────┘
                   ▼
           InventoryService
                   ▼
        InMemoryInventoryStore (demo)
                   ▼
          Human bin confirmation
```

**Do not** call MCP from a browser. Keep **REST** and **MCP** in separate terminals.

### Install (macOS / zsh)

Use `python3`. Quote pip extras so zsh does not glob them:

```bash
cd ~/Desktop/icohangar-repos/battery-erp
python3 -m pip install -e '.[dev]'   # api + mcp + pytest
```

### Terminal A — REST (text-line backend)

```bash
cd ~/Desktop/icohangar-repos/battery-erp
export BATTERY_ERP_CONFIRM_TOKEN=dev-secret
export BATTERY_ERP_AUDIT_LOG=/tmp/battery-erp-audit.jsonl
PYTHONPATH=src python3 -m battery_erp.api
# Uvicorn → http://127.0.0.1:8088
```

```bash
curl -s http://127.0.0.1:8088/health
curl -s http://127.0.0.1:8088/inventory/lookup/lithium
```

### Terminal B — MCP (agents / Cursor only)

Leave Terminal A running. In a **new** terminal:

```bash
cd ~/Desktop/icohangar-repos/battery-erp
export BATTERY_ERP_CONFIRM_TOKEN=dev-secret
PYTHONPATH=src python3 -m battery_erp.mcp
```

That process stays quiet on stdio — normal for MCP hosts. Do not paste this into the API terminal.

### Cursor MCP config

Add to `~/.cursor/mcp.json` (absolute paths; `python3` not `python`):

```json
{
  "mcpServers": {
    "battery-erp": {
      "command": "python3",
      "args": ["-m", "battery_erp.mcp"],
      "env": {
        "PYTHONPATH": "/Users/YOU/Desktop/icohangar-repos/battery-erp/src",
        "BATTERY_ERP_CONFIRM_TOKEN": "replace-me",
        "BATTERY_ERP_AUDIT_LOG": "/tmp/battery-erp-audit.jsonl"
      }
    }
  }
}
```

Then reload MCP servers in Cursor. Tools exposed: `lookup_inventory`,
`get_inventory_status`, `get_inventory_record`, `list_inventory`,
`create_bin_check_request`, `record_bin_confirmation`.

| Surface | Entry | Shared layer |
|---|---|---|
| REST | `GET /inventory/lookup/{part}`, bin-check / bin-confirm | `battery_erp.services.InventoryService` |
| MCP | tools above (mcp SDK 2.x `MCPServer`) | same |

Mutating bin confirmation requires `BATTERY_ERP_CONFIRM_TOKEN` and optionally
writes JSONL audit to `BATTERY_ERP_AUDIT_LOG`.

Directory / registry packaging: see **[PUBLISH.md](PUBLISH.md)** (`glama.json`,
`Dockerfile`, `server.json`).

| Directory | Link |
|---|---|
| PyPI | https://pypi.org/project/battery-erp/ |
| MCP Registry | `io.github.icohangar-ops/battery-erp` |
| Glama | https://glama.ai/mcp/servers/@icohangar-ops/battery-erp |

---

## Tests

```bash
PYTHONPATH=src python3 -m pytest tests/ -v
# Domain tests + inventory service / REST / MCP scaffold
```

---

## Use cases

- **Cell manufacturers** — BOM cost tracking across chemistries, yield optimization
- **Pack integrators** — pack-level cost estimation, supplier selection
- **Procurement** — supplier scoring, dual-sourcing, PO pipeline management
- **Finance** — commodity price risk, what-if scenarios, inventory valuation
- **C-suite** — dashboard showing $/kWh trends, supply chain resilience, cost reduction opportunities

---

## License

MIT. See [LICENSE](LICENSE).

---

## CHP Governance

This repository is hardened with the [Consensus Hardening Protocol (CHP)](https://codeberg.org/cubiczan/consensus-hardening-protocol), Cubiczan's decision-governance layer for multi-agent AI systems.

### Protocol Layers
- **R0 Gate**: All decisions must pass Solvable, Scoped, Valid, Worth_it checks
- **Foundation Disclosure**: 1-3 weakest assumptions, 1-2 invalidation conditions, 1 key vulnerability
- **Adversarial Layer**: Mandatory devil's advocate at Phase 0 and Round 3
- **State Machine**: EXPLORING → PROVISIONAL → PROVISIONAL_LOCK → LOCKED
- **Third-Party Validation**: Independent CONFIRM/REJECT before lock

### Domain Configuration
- **Category**: Mining / Supply Chain
- **Foundation Threshold**: 75
- **CFO Accuracy Guard**: Disabled

### Compliance Artifacts
| File | Purpose |
|------|---------|
| `.chp/STATE_MACHINE.md` | Decision state transitions |
| `.chp/R0_CONFIG.yaml` | Domain-calibrated thresholds |
| `.chp/ADVERSARIAL_PROMPTS.md` | Standardized challenge templates |
| `.chp/CHP_COMPLIANCE.md` | Compliance tracking & audit trail |

### CHP Version
cognitive-mesh-orchestrator 0.1.0 | [Protocol Docs](https://codeberg.org/cubiczan/consensus-hardening-protocol)


## Demo

### Handheld — Parts Bin Text Line

Mobile capture from the live lead page
([parts-bin-text-line--icohangar.replit.app](https://parts-bin-text-line--icohangar.replit.app/)).
SMS → text-line → Battery ERP inventory check → human bin confirm.

<p align="center">
  <img src="docs/media/parts-bin-handheld.png" alt="Parts Bin Text Line handheld demo" width="360" />
</p>

[![Demo Video](https://img.shields.io/badge/Watch_Demo-3min-blue)](docs/media/battery-erp-demo.mp4)