fatcrapinmybutt/SINGULARITY-MBP-GENESIS

SINGULARITY-MBP-GENESIS — Architectural DNA of THEMANBEARPIG

Tier 0 — GENESIS: This is the foundational schema from which every other MBP skill derives. No visualization, no combat overlay, no emergence signal exists without this ontology. Every node, every link, every layer traces back to definitions in this file.

1. THEMANBEARPIG Overview

THEMANBEARPIG is a 13-layer interactive D3.js force-directed graph that renders the entire Pigors v. Watson litigation universe as an explorable intelligence visualization. It transforms 786+ database tables containing 1.3M+ rows into a living, breathing graph of actors, evidence, authorities, filings, events, weapons, and emergent patterns.

Design Philosophy

PRINCIPLE 1: Every node is a DB row. Every link is a DB relationship.
PRINCIPLE 2: Layers are semantic groupings, not visual z-indexes.
PRINCIPLE 3: Force simulation parameters are per-layer, not global.
PRINCIPLE 4: The graph is constructed from data, never hardcoded.
PRINCIPLE 5: Performance scales to 10,000+ nodes via LOD + culling.
PRINCIPLE 6: Interactions reveal intelligence — click = deep dive.
PRINCIPLE 7: The graph is the case. If it's not in the graph, it's not real.

Target Metrics

| Metric | Target | Stretch | |--------|--------|---------| | Total nodes | 2,500+ | 10,000+ | | Total links | 5,000+ | 25,000+ | | Layers | 13 | 13 (fixed) | | Frame rate | 30 FPS | 60 FPS | | Initial load | < 3s | < 1s | | Node types | 20+ | 30+ | | Link types | 17+ | 25+ |

---

2. Node Taxonomy — Complete Type Registry

Every node in THEMANBEARPIG has a type (primary category), a subtype (specific kind), and a set of properties that vary by type. The type system is hierarchical:

NODE
├── PERSON
│   ├── adversary
│   ├── judicial
│   ├── witness
│   ├── attorney
│   ├── law_enforcement
│   ├── agency_official
│   ├── child
│   ├── family_member
│   └── plaintiff
├── EVIDENCE
│   ├── document
│   ├── recording
│   ├── police_report
│   ├── court_order
│   ├── communication
│   ├── photo
│   ├── video
│   ├── medical_record
│   └── financial_record
├── AUTHORITY
│   ├── statute
│   ├── court_rule
│   ├── case_law
│   ├── constitutional
│   └── federal_statute
├── FILING
│   ├── motion
│   ├── brief
│   ├── complaint
│   ├── petition
│   ├── affidavit
│   ├── exhibit
│   ├── proposed_order
│   └── certificate_of_service
├── EVENT
│   ├── hearing
│   ├── order_entered
│   ├── violation
│   ├── incident
│   ├── contact
│   ├── withholding
│   ├── arrest
│   └── filing_event
├── WEAPON
│   ├── false_allegation
│   ├── ppo_weaponization
│   ├── contempt_abuse
│   ├── evidence_suppression
│   ├── ex_parte_abuse
│   ├── medication_coercion
│   └── incarceration_weapon
├── INSTITUTION
│   ├── court
│   ├── agency
│   ├── law_firm
│   ├── school
│   ├── employer
│   ├── medical_facility
│   └── law_enforcement_agency
└── CLAIM
    ├── cause_of_action
    ├── defense
    ├── counterclaim
    └── affirmative_defense

2.1 PERSON Nodes

Person nodes represent human actors in the litigation universe. Every person has a threat level, connection count, and role classification.

/**
 * PERSON node schema.
 * Source tables: evidence_quotes (actors), timeline_events (actors),
 *                impeachment_matrix (target), berry_mcneill_intelligence
 */
const PersonNode = {
  // Identity
  id: "person_andrew_pigors",           // Stable ID: type_subtype_name
  type: "PERSON",
  subtype: "plaintiff",                  // adversary|judicial|witness|attorney|...
  label: "Andrew J. Pigors",            // Display name
  short_label: "Pigors",                // Compact label for zoomed-out view

  // Classification
  party_role: "plaintiff",              // plaintiff|defendant|judge|witness|third_party
  alignment: "allied",                  // allied|adverse|neutral|unknown
  threat_level: 0,                      // 0-10 (0=self, 10=primary adversary)

  // Metrics (drive node sizing)
  evidence_count: 0,                    // Evidence nodes linked to this person
  impeachment_count: 0,                 // Impeachment entries targeting this person
  contradiction_count: 0,               // Contradictions involving this person
  violation_count: 0,                   // Judicial violations (if judge)
  mention_count: 0,                     // Total mentions across all tables

  // Visual
  layer: 0,                             // Primary layer assignment
  color: null,                          // Computed from alignment + threat_level
  radius: null,                         // Computed from evidence_count
  icon: "user",                         // Icon glyph
  glow: false,                          // Pulsing glow effect for high-threat

  // Metadata
  case_lanes: ["A"],                    // Which case lanes this person appears in
  address: null,                        // Physical address (for service tracking)
  bar_number: null,                     // Bar number (attorneys only)
  relationship_to_child: null,          // father|mother|none (custody relevance)

  // D3 force simulation
  fx: null,                             // Fixed x position (null = free)
  fy: null,                             // Fixed y position (null = free)
  charge_multiplier: 1.0,              // Per-node charge override
};

Person Subtypes — Complete Registry

| Subtype | Alignment | Threat | Icon | Examples | |---------|-----------|--------|------|----------| | plaintiff | allied | 0 | ⚖️ | Andrew J. Pigors | | adversary | adverse | 7-10 | 🎯 | Emily A. Watson | | judicial | adverse | 8-10 | 🔨 | Hon. Jenny L. McNeill, Hon. Kenneth Hoopes | | witness | varies | 1-5 | 👁️ | Officer Randall, HealthWest evaluator | | attorney | varies | 3-6 | 📜 | Jennifer Barnes P55406 (WITHDREW) | | law_enforcement | neutral | 1-3 | 🛡️ | NSPD officers | | agency_official | varies | 4-7 | 🏛️ | Pamela Rusco (FOC) | | child | protected | 0 | 👶 | L.D.W. (initials ONLY — MCR 8.119(H)) | | family_member | adverse | 5-8 | 👥 | Albert Watson, Lori Watson, Ronald Berry |

Canonical Person Registry — Hardcoded Core Actors

These persons are ALWAYS present in the graph. They form Layer 0 (CORE).

const CORE_PERSONS = [
  // LAYER 0: CORE
  {
    id: "person_plaintiff_andrew_pigors",
    label: "Andrew J. Pigors",
    subtype: "plaintiff",
    alignment: "allied",
    threat_level: 0,
    layer: 0,
    case_lanes: ["A", "B", "C", "D", "E", "F", "CRIMINAL"],
    relationship_to_child: "father",
    fixed_position: { x: 0, y: 0 },  // Graph center
  },
  {
    id: "person_adversary_emily_watson",
    label: "Emily A. Watson",
    subtype: "adversary",
    alignment: "adverse",
    threat_level: 9,
    layer: 0,
    case_lanes: ["A", "D"],
    relationship_to_child: "mother",
  },
  {
    id: "person_child_ldw",
    label: "L.D.W.",
    subtype: "child",
    alignment: "protected",
    threat_level: 0,
    layer: 0,
    case_lanes: ["A"],
    relationship_to_child: "self",
    icon: "child",
    // NEVER display full name — MCR 8.119(H)
  },
  {
    id: "person_judicial_mcneill",
    label: "Hon. Jenny L. McNeill",
    subtype: "judicial",
    alignment: "adverse",
    threat_level: 10,
    layer: 0,
    case_lanes: ["A", "D", "E"],
    bar_number: "P58235",
  },

  // LAYER 1: ADVERSARY NETWORK
  {
    id: "person_family_albert_watson",
    label: "Albert Watson",
    subtype: "family_member",
    alignment: "adverse",
    threat_level: 7,
    layer: 1,
    case_lanes: ["A"],
  },
  {
    id: "person_family_lori_watson",
    label: "Lori Watson",
    subtype: "family_member",
    alignment: "adverse",
    threat_level: 5,
    layer: 1,
    case_lanes: ["A"],
  },
  {
    id: "person_family_ronald_berry",
    label: "Ronald Berry",
    subtype: "family_member",
    alignment: "adverse",
    threat_level: 6,
    layer: 1,
    case_lanes: ["A"],
    // NON-ATTORNEY — never display "Esq." or bar number
  },

  // LAYER 2: JUDICIAL CARTEL
  {
    id: "person_judicial_hoopes",
    label: "Hon. Kenneth Hoopes",
    subtype: "judicial",
    alignment: "adverse",
    threat_level: 8,
    layer: 2,
    case_lanes: ["B", "E"],
  },
  {
    id: "person_judicial_ladas_hoopes",
    label: "Hon. Maria Ladas-Hoopes",
    subtype: "judicial",
    alignment: "adverse",
    threat_level: 7,
    layer: 2,
    case_lanes: ["CRIMINAL", "E"],
  },
  {
    id: "person_agency_rusco",
    label: "Pamela Rusco",
    subtype: "agency_official",
    alignment: "adverse",
    threat_level: 6,
    layer: 2,
    case_lanes: ["A"],
    institution: "FOC",
  },
  {
    id: "person_family_cavan_berry",
    label: "Cavan Berry",
    subtype: "family_member",
    alignment: "adverse",
    threat_level: 5,
    layer: 2,
    case_lanes: ["E"],
    // Attorney magistrate 60th District — McNeill's spouse
    // Office at 990 Terrace St = FOC address
  },
  {
    id: "person_attorney_barnes",
    label: "Jennifer Barnes",
    subtype: "attorney",
    alignment: "adverse",
    threat_level: 3,
    layer: 2,
    case_lanes: ["A", "D"],
    bar_number: "P55406",
    status: "WITHDREW",  // Withdrew March 2026
  },
];

2.2 EVIDENCE Nodes

Evidence nodes represent physical or digital proof items. They link to PERSON nodes (who created/discovered them), FILING nodes (where they're used), and AUTHORITY nodes (what rules govern their admissibility).

/**
 * EVIDENCE node schema.
 * Source tables: evidence_quotes, police_reports, documents
 */
const EvidenceNode = {
  id: "evidence_doc_ns2505044",
  type: "EVIDENCE",
  subtype: "police_report",             // document|recording|police_report|court_order|...
  label: "NSPD NS2505044",
  short_label: "NS2505044",

  // Evidence metadata
  source_file: "J:\\POLICE_REPORTS\\Albert calling police.pdf",
  bates_number: "PIGORS-A-000142",     // Bates stamp
  page_count: 3,
  date: "2025-08-07",
  lane: "A",                            // Primary case lane

  // Classification
  category: "police_report",           // From 20-type evidence taxonomy
  relevance_score: 0.95,               // 0-1 relevance to active claims
  impeachment_value: 9,                // 1-10 impeachment potential
  is_smoking_gun: true,                // Highlight flag

  // Content summary
  key_quote: "They want this documented so Emily can go tomorrow to get an Ex Parte order",
  actors: ["Albert Watson", "Emily A. Watson"],
  topics: ["premeditation", "ex_parte", "custody"],

  // Authentication
  mre_basis: "MRE 901(b)(7)",         // Authentication rule
  is_authenticated: true,
  chain_of_custody: true,

  // Visual
  layer: 3,
  color: null,                          // Computed from relevance_score
  radius: null,                         // Computed from impeachment_value
  icon: "file-text",
  pulse: false,                         // Pulse animation for smoking guns
};

Evidence Subtypes

| Subtype | Icon | Source Table | Key Properties | |---------|------|-------------|----------------| | document | 📄 | documents | file_path, page_count, sha256 | | recording | 🎙️ | evidence_quotes | duration, participants, sullivan_auth | | police_report | 🚔 | police_reports | incident_number, officer, outcome | | court_order | ⚖️ | timeline_events | order_type, judge, case_number | | communication | 💬 | evidence_quotes | platform (AppClose/email/text), sender, receiver | | photo | 📷 | evidence_quotes | location, timestamp, subjects | | video | 🎥 | evidence_quotes | duration, location, participants | | medical_record | 🏥 | evidence_quotes | provider, findings, date | | financial_record | 💰 | evidence_quotes | institution, amount, date_range |

2.3 AUTHORITY Nodes

Authority nodes represent legal citations — the weapons of litigation.

/**
 * AUTHORITY node schema.
 * Source tables: authority_chains_v2, michigan_rules_extracted, master_citations
 */
const AuthorityNode = {
  id: "authority_statute_mcl_722_23",
  type: "AUTHORITY",
  subtype: "statute",                   // statute|court_rule|case_law|constitutional|federal_statute
  label: "MCL 722.23",
  short_label: "§722.23",
  full_title: "Child Custody Act — Best Interest Factors",

  // Citation metadata
  citation: "MCL 722.23",
  pin_cite: "MCL 722.23(a)-(l)",
  jurisdiction: "Michigan",
  source_type: "statute",

  // Usage metrics
  chain_count: 0,                       // How many authority_chains reference this
  filing_count: 0,                      // How many filings cite this
  claim_count: 0,                       // How many claims rely on this

  // Classification
  domain: "custody",                    // custody|ppo|contempt|due_process|appellate|...
  strength: "binding",                  // binding|persuasive|advisory
  is_verified: true,                    // Verified in DB (not hallucinated)

  // Visual
  layer: 4,
  color: null,
  radius: null,                         // Computed from chain_count
  icon: "book",
};

2.4 FILING Nodes

Filing nodes represent documents in the filing pipeline (F1-F10).

/**
 * FILING node schema.
 * Source tables: filing_readiness, filing_packages, deadlines
 */
const FilingNode = {
  id: "filing_f09_coa_brief",
  type: "FILING",
  subtype: "brief",
  label: "F09: COA Brief 366810",
  short_label: "F09",

  // Filing metadata
  lane_id: "F9",                        // F1-F10 identifier
  case_number: "366810",
  court: "MI Court of Appeals",
  filing_type: "brief",
  status: "DRAFT",                      // DRAFT|QA_REVIEW|SERVICE_READY|FILED|DOCKETED

  // Readiness
  readiness_score: 0.85,               // 0-1 readiness percentage
  evidence_count: 42,                  // Supporting evidence nodes
  authority_count: 28,                 // Supporting authority nodes
  gap_count: 3,                        // Missing items

  // Deadlines
  deadline: "2026-04-30",
  days_remaining: null,                 // Computed dynamically
  urgency: "HIGH",                      // CRITICAL|HIGH|MEDIUM|LOW

  // Visual
  layer: 5,
  color: null,                          // Computed from status + urgency
  radius: null,                         // Computed from readiness_score
  icon: "file-plus",
};

2.5 EVENT Nodes

/**
 * EVENT node schema.
 * Source tables: timeline_events, docket_events
 */
const EventNode = {
  id: "event_hearing_20240717",
  type: "EVENT",
  subtype: "hearing",                   // hearing|order_entered|violation|incident|...
  label: "Custody Trial",
  short_label: "Trial",

  date: "2024-07-17",
  date_end: null,                       // For multi-day events
  actors: ["Andrew J. Pigors", "Emily A. Watson", "Hon. Jenny L. McNeill"],
  location: "14th Circuit Court",
  case_number: "2024-001507-DC",
  lane: "A",

  // Event classification
  event_category: "hearing",
  significance: 10,                     // 1-10 impact score
  outcome: "adverse",                   // favorable|adverse|neutral|mixed

  // Narrative
  description: "Sole custody awarded to Mother. All 12 MCL 722.23 factors found in Mother's favor.",
  key_quote: null,

  // Visual
  layer: 6,
  color: null,
  radius: null,                         // Computed from significance
  icon: "calendar",
};

2.6 WEAPON Nodes

Weapon nodes represent legal instruments that have been weaponized against Andrew.

/**
 * WEAPON node schema.
 * Source tables: judicial_violations, impeachment_matrix, contradiction_map
 */
const WeaponNode = {
  id: "weapon_false_allegation_arsenic",
  type: "WEAPON",
  subtype: "false_allegation",          // false_allegation|ppo_weaponization|contempt_abuse|...
  label: "Arsenic Poisoning Allegation",
  short_label: "Arsenic FA",

  // Weapon metadata
  wielder: "Emily A. Watson",          // Who used the weapon
  target: "Andrew J. Pigors",          // Target of the weapon
  date: "2024-03-15",
  lane: "A",

  // Classification
  weapon_category: "false_allegation",
  severity: 8,                          // 1-10 severity
  debunked: true,                       // Whether weapon has been neutralized
  rebuttal_evidence_count: 37,         // Evidence nodes that rebut this weapon

  // Impact
  resulted_in: ["custody_loss", "investigation"],
  harm_category: "reputational",

  // Visual
  layer: 7,
  color: "#ff0000",                     // Weapons are always red-spectrum
  radius: null,
  icon: "alert-triangle",
  pulse: true,                          // Weapons pulse to draw attention
};

Weapon Subtypes — The Arsenal of Abuse

| Subtype | Severity Range | Count | Key Pattern | |---------|---------------|-------|-------------| | false_allegation | 6-9 | 7+ | Suicidal → arsenic → assault → drugs → threats | | ppo_weaponization | 8-10 | 7+ | PPO filed 2 days after recanting | | contempt_abuse | 9-10 | 3+ | SC#5 (14d) + SC#6+7 (45d) = 59 days | | evidence_suppression | 7-9 | 5+ | HealthWest eval excluded | | ex_parte_abuse | 9-10 | 8+ | 5 orders on single day (Aug 8, 2025) | | medication_coercion | 8-9 | 1+ | Medication as condition for parenting time | | incarceration_weapon | 10 | 2+ | Jailed for birthday messages |

2.7 INSTITUTION Nodes

const InstitutionNode = {
  id: "institution_court_14th_circuit",
  type: "INSTITUTION",
  subtype: "court",
  label: "14th Circuit Court — Muskegon",
  short_label: "14th Circuit",

  // Metadata
  address: "Muskegon County, MI",
  jurisdiction: "Michigan",
  role: "trial_court",

  // Connections
  judges: ["person_judicial_mcneill", "person_judicial_hoopes"],
  cases: ["2024-001507-DC", "2023-5907-PP", "2025-002760-CZ"],

  layer: 2,
  icon: "building",
};

2.8 CLAIM Nodes

const ClaimNode = {
  id: "claim_coa_parental_alienation",
  type: "CLAIM",
  subtype: "cause_of_action",
  label: "Parental Alienation — MCL 722.23(j)",
  short_label: "Factor (j)",

  // Claim metadata
  legal_basis: "MCL 722.23(j)",
  claim_type: "cause_of_action",
  lane: "A",
  status: "active",

  // Strength metrics
  evidence_count: 2404,                // Supporting evidence
  authority_count: 12,                 // Supporting authorities
  impeachment_count: 45,              // Cross-exam ammunition
  strength_score: 0.92,               // 0-1 overall strength

  // Elements
  elements: [
    "Willingness to facilitate close relationship",
    "Pattern of interference with parenting time",
    "False allegations to limit contact",
  ],
  elements_met: 3,
  elements_total: 3,

  layer: 10,
  icon: "target",
};

---

3. Link Taxonomy — Complete Relationship Registry

Links connect nodes across and within layers. Every link has a type, strength, and directional semantics (source → target).

/**
 * Link schema — universal for all link types.
 */
const GraphLink = {
  id: "link_filed_by_motion_pigors",
  source: "filing_f09_coa_brief",      // Source node ID
  target: "person_plaintiff_andrew_pigors", // Target node ID
  type: "FILED_BY",                     // Relationship type
  strength: 0.8,                        // 0-1 link strength (affects force simulation)
  label: null,                          // Optional link label
  date: "2026-04-15",                  // When this relationship was established
  bidirectional: false,                // One-way or two-way
  layer: null,                          // Inherit from source node's layer
  style: "solid",                       // solid|dashed|dotted|animated
  color: null,                          // Computed from type
  width: null,                          // Computed from strength
};

3.1 Complete Link Type Registry

Procedural Links (Filing ↔ Person)

| Type | Direction | Meaning | Style | Color | |------|-----------|---------|-------|-------| | FILED_BY | Filing → Person | Person filed this document | solid | #4a90d9 | | AGAINST | Filing → Person | Filing targets this person | solid | #e74c3c | | REGARDING | Filing → Person | Filing mentions this person | dashed | #95a5a6 | | PRESIDES_OVER | Person → Filing/Event | Judge presides over case | solid | #8e44ad | | ASSIGNED_TO | Filing → Institution | Filed in this court | solid | #3498db |

Evidentiary Links (Evidence ↔ *)

| Type | Direction | Meaning | Style | Color | |------|-----------|---------|-------|-------| | SUPPORTS | Evidence → Claim/Filing | Evidence supports claim/filing | solid | #27ae60 | | CONTRADICTS | Evidence → Evidence | Two pieces of evidence conflict | animated | #e74c3c | | AUTHENTICATES | Person → Evidence | Person can authenticate evidence | dashed | #f39c12 | | CREATED_BY | Evidence → Person | Person created/obtained this evidence | solid | #3498db | | EXHIBITS | Filing → Evidence | Filing includes this as exhibit | solid | #9b59b6 |

Relational Links (Person ↔ Person)

| Type | Direction | Meaning | Style | Color | |------|-----------|---------|-------|-------| | MARRIED_TO | Person ↔ Person | Spousal relationship | solid, thick | #e91e63 | | RELATED_TO | Person ↔ Person | Family relationship | solid | #ff9800 | | LIVES_WITH | Person ↔ Person | Cohabitation | dashed | #ff9800 | | EMPLOYS | Institution → Person | Employment relationship | solid | #607d8b | | REPRESENTS | Person → Person | Attorney represents party | solid | #795548 |

Adversarial Links (Combat Intelligence)

| Type | Direction | Meaning | Style | Color | |------|-----------|---------|-------|-------| | RETALIATES | Person → Person/Event | Retaliatory action | animated, red | #ff0000 | | COORDINATES_WITH | Person ↔ Person | Coordinated adverse action | dashed, red | #ff4444 | | CONSPIRES_WITH | Person ↔ Person | Conspiracy link | animated, red | #cc0000 | | WIELDS | Person → Weapon | Person used this weapon | solid, red | #e74c3c | | TARGETS | Weapon → Person | Weapon was used against person | solid, red | #c0392b |

Authority Links (Legal Citations)

| Type | Direction | Meaning | Style | Color | |------|-----------|---------|-------|-------| | CITES | Filing/Claim → Authority | Document cites this authority | solid | #2ecc71 | | SUPERSEDES | Authority → Authority | Newer authority supersedes older | dashed | #e67e22 | | AMENDS | Authority → Authority | Amendment relationship | dashed | #f1c40f | | VIOLATES | Person/Event → Authority | Action violates this rule | animated | #e74c3c | | GOVERNS | Authority → Claim | Authority governs this claim | solid | #1abc9c |

Cartel Links (Judicial Intelligence)

| Type | Direction | Meaning | Style | Color | |------|-----------|---------|-------|-------| | FORMER_PARTNER | Person ↔ Person | Former law firm partners | thick, red | #ff0000 | | SPOUSE_OF | Person ↔ Person | Judicial spouse connection | thick, red | #e91e63 | | SHARES_ADDRESS | Person/Institution ↔ Person/Institution | Same physical address | dashed | #ff6600 | | EX_PARTE_WITH | Person → Person | Ex parte communication | animated, red | #ff0000 |

3.2 Link Construction Rules

/**
 * Link factory — creates properly typed links with computed visual properties.
 */
function createLink(source, target, type, metadata = {}) {
  const LINK_DEFAULTS = {
    // Procedural
    FILED_BY:         { strength: 0.7, color: "#4a90d9", style: "solid", width: 2 },
    AGAINST:          { strength: 0.8, color: "#e74c3c", style: "solid", width: 2 },
    REGARDING:        { strength: 0.3, color: "#95a5a6", style: "dashed", width: 1 },
    PRESIDES_OVER:    { strength: 0.9, color: "#8e44ad", style: "solid", width: 3 },
    ASSIGNED_TO:      { strength: 0.5, color: "#3498db", style: "solid", width: 1 },
    // Evidentiary
    SUPPORTS:         { strength: 0.6, color: "#27ae60", style: "solid", width: 2 },
    CONTRADICTS:      { strength: 0.9, color: "#e74c3c", style: "animated", width: 3 },
    AUTHENTICATES:    { strength: 0.4, color: "#f39c12", style: "dashed", width: 1 },
    CREATED_BY:       { strength: 0.5, color: "#3498db", style: "solid", width: 1 },
    EXHIBITS:         { strength: 0.6, color: "#9b59b6", style: "solid", width: 2 },
    // Relational
    MARRIED_TO:       { strength: 1.0, color: "#e91e63", style: "solid", width: 4 },
    RELATED_TO:       { strength: 0.8, color: "#ff9800", style: "solid", width: 3 },
    LIVES_WITH:       { strength: 0.7, color: "#ff9800", style: "dashed", width: 2 },
    EMPLOYS:          { strength: 0.5, color: "#607d8b", style: "solid", width: 1 },
    REPRESENTS:       { strength: 0.8, color: "#795548", style: "solid", width: 2 },
    // Adversarial
    RETALIATES:       { strength: 0.9, color: "#ff0000", style: "animated", width: 3 },
    COORDINATES_WITH: { strength: 0.8, color: "#ff4444", style: "dashed", width: 2 },
    CONSPIRES_WITH:   { strength: 1.0, color: "#cc0000", style: "animated", width: 4 },
    WIELDS:           { strength: 0.7, color: "#e74c3c", style: "solid", width: 2 },
    TARGETS:          { strength: 0.8, color: "#c0392b", style: "solid", width: 2 },
    // Authority
    CITES:            { strength: 0.5, color: "#2ecc71", style: "solid", width: 1 },
    SUPERSEDES:       { strength: 0.6, color: "#e67e22", style: "dashed", width: 2 },
    AMENDS:           { strength: 0.4, color: "#f1c40f", style: "dashed", width: 1 },
    VIOLATES:         { strength: 0.9, color: "#e74c3c", style: "animated", width: 3 },
    GOVERNS:          { strength: 0.6, color: "#1abc9c", style: "solid", width: 2 },
    // Cartel
    FORMER_PARTNER:   { strength: 1.0, color: "#ff0000", style: "solid", width: 5 },
    SPOUSE_OF:        { strength: 1.0, color: "#e91e63", style: "solid", width: 5 },
    SHARES_ADDRESS:   { strength: 0.7, color: "#ff6600", style: "dashed", width: 2 },
    EX_PARTE_WITH:    { strength: 0.9, color: "#ff0000", style: "animated", width: 4 },
  };

  const defaults = LINK_DEFAULTS[type] || { strength: 0.5, color: "#999", style: "solid", width: 1 };

  return {
    id: `link_${type.toLowerCase()}_${source}_${target}`,
    source,
    target,
    type,
    ...defaults,
    ...metadata,
    bidirectional: ["MARRIED_TO", "RELATED_TO", "LIVES_WITH", "COORDINATES_WITH",
                    "CONSPIRES_WITH", "FORMER_PARTNER", "SPOUSE_OF",
                    "SHARES_ADDRESS"].includes(type),
  };
}

---

4. Layer Architecture — The 13 Layers of THEMANBEARPIG

Each layer is a semantic grouping of nodes that share a conceptual domain. Layers are NOT z-indexes — they are filter/visibility groups with independent force simulation parameters.

4.1 Layer Definitions

/**
 * LAYER_META — Master configuration for all 13 layers.
 * Each layer defines: visual properties, force parameters, data sources, and filter rules.
 */
const LAYER_META = {
  0: {
    id: 0,
    name: "CORE",
    description: "Central actors — the gravitational center of the litigation universe",
    color_primary: "#ffffff",
    color_secondary: "#e0e0e0",
    background: "radial-gradient(circle, #1a1a2e 0%, #0d0d1a 100%)",
    node_types: ["PERSON"],
    subtypes: ["plaintiff", "adversary", "child", "judicial"],
    max_nodes: 4,
    always_visible: true,
    z_order: 100,  // Highest z-order — always on top

    // Force simulation parameters
    force: {
      charge: -800,              // Strong repulsion — core actors spread wide
      charge_min_distance: 100,
      charge_max_distance: 500,
      collision_radius: 60,
      link_distance: 200,
      link_strength: 0.8,
      center_strength: 0.3,     // Strong pull to center
      radial_radius: 0,         // No radial constraint (centered)
      radial_strength: 0,
      velocity_decay: 0.4,
      alpha_decay: 0.02,
    },

    // Data source queries
    data_sources: [
      {
        table: "evidence_quotes",
        query: "SELECT DISTINCT actor FROM evidence_quotes WHERE actor IN ('Andrew Pigors', 'Emily Watson', 'Jenny McNeill')",
        node_factory: "createPersonNode",
      },
    ],
  },

  1: {
    id: 1,
    name: "ADVERSARY",
    description: "Emily's network — family members, allies, co-conspirators",
    color_primary: "#ff6b6b",
    color_secondary: "#ee5a24",
    background: "radial-gradient(circle, #2d1f1f 0%, #1a0d0d 100%)",
    node_types: ["PERSON"],
    subtypes: ["adversary", "family_member", "witness"],
    max_nodes: 20,
    always_visible: false,
    z_order: 90,

    force: {
      charge: -400,
      charge_min_distance: 50,
      charge_max_distance: 400,
      collision_radius: 40,
      link_distance: 150,
      link_strength: 0.6,
      center_strength: 0.05,
      radial_radius: 250,       // Orbit around core at 250px
      radial_strength: 0.3,
      velocity_decay: 0.4,
      alpha_decay: 0.02,
    },

    data_sources: [
      {
        table: "impeachment_matrix",
        query: "SELECT DISTINCT target AS actor FROM impeachment_matrix WHERE target LIKE '%Watson%' OR target LIKE '%Berry%'",
        node_factory: "createPersonNode",
      },
    ],
  },

  2: {
    id: 2,
    name: "JUDICIAL",
    description: "Court actors — judges, FOC, attorneys, court staff",
    color_primary: "#8e44ad",
    color_secondary: "#9b59b6",
    background: "radial-gradient(circle, #1a1a2e 0%, #0d0d1f 100%)",
    node_types: ["PERSON", "INSTITUTION"],
    subtypes: ["judicial", "agency_official", "attorney", "court"],
    max_nodes: 30,
    always_visible: false,
    z_order: 85,

    force: {
      charge: -350,
      charge_min_distance: 40,
      charge_max_distance: 350,
      collision_radius: 35,
      link_distance: 120,
      link_strength: 0.5,
      center_strength: 0.03,
      radial_radius: 350,
      radial_strength: 0.25,
      velocity_decay: 0.4,
      alpha_decay: 0.025,
    },

    data_sources: [
      {
        table: "judicial_violations",
        query: "SELECT DISTINCT judge_name AS actor FROM judicial_violations",
        node_factory: "createPersonNode",
      },
      {
        table: "berry_mcneill_intelligence",
        query: "SELECT DISTINCT entity_name AS actor FROM berry_mcneill_intelligence",
        node_factory: "createPersonNode",
      },
    ],
  },

  3: {
    id: 3,
    name: "EVIDENCE",
    description: "Document and recording nodes — the proof layer",
    color_primary: "#3498db",
    color_secondary: "#2980b9",
    background: "radial-gradient(circle, #0d1f2d 0%, #061018 100%)",
    node_types: ["EVIDENCE"],
    subtypes: ["document", "recording", "police_report", "court_order",
               "communication", "photo", "video", "medical_record", "financial_record"],
    max_nodes: 500,
    always_visible: false,
    z_order: 70,

    force: {
      charge: -100,             // Light repulsion — many nodes, keep compact
      charge_min_distance: 20,
      charge_max_distance: 200,
      collision_radius: 15,
      link_distance: 80,
      link_strength: 0.3,
      center_strength: 0.01,
      radial_radius: 500,
      radial_strength: 0.15,
      velocity_decay: 0.5,
      alpha_decay: 0.03,
    },

    data_sources: [
      {
        table: "evidence_quotes",
        query: `SELECT source_file, category, lane,
                       COUNT(*) as quote_count,
                       MAX(CAST(relevance_score AS REAL)) as max_relevance
                FROM evidence_quotes
                WHERE is_duplicate = 0
                GROUP BY source_file
                ORDER BY quote_count DESC
                LIMIT 500`,
        node_factory: "createEvidenceNode",
      },
    ],
  },

  4: {
    id: 4,
    name: "AUTHORITY",
    description: "Legal citations and rules — the weapons of law",
    color_primary: "#2ecc71",
    color_secondary: "#27ae60",
    background: "radial-gradient(circle, #0d2d1f 0%, #061810 100%)",
    node_types: ["AUTHORITY"],
    subtypes: ["statute", "court_rule", "case_law", "constitutional", "federal_statute"],
    max_nodes: 200,
    always_visible: false,
    z_order: 65,

    force: {
      charge: -150,
      charge_min_distance: 25,
      charge_max_distance: 250,
      collision_radius: 20,
      link_distance: 100,
      link_strength: 0.4,
      center_strength: 0.02,
      radial_radius: 550,
      radial_strength: 0.2,
      velocity_decay: 0.45,
      alpha_decay: 0.025,
    },

    data_sources: [
      {
        table: "authority_chains_v2",
        query: `SELECT primary_citation, supporting_citation, relationship,
                       source_type, lane, COUNT(*) as chain_count
                FROM authority_chains_v2
                GROUP BY primary_citation
                ORDER BY chain_count DESC
                LIMIT 200`,
        node_factory: "createAuthorityNode",
      },
    ],
  },

  5: {
    id: 5,
    name: "FILING",
    description: "Filing pipeline F1-F10 — the actionable output layer",
    color_primary: "#e67e22",
    color_secondary: "#d35400",
    background: "radial-gradient(circle, #2d1f0d 0%, #180d06 100%)",
    node_types: ["FILING"],
    subtypes: ["motion", "brief", "complaint", "petition", "affidavit",
               "exhibit", "proposed_order", "certificate_of_service"],
    max_nodes: 50,
    always_visible: false,
    z_order: 75,

    force: {
      charge: -250,
      charge_min_distance: 30,
      charge_max_distance: 300,
      collision_radius: 30,
      link_distance: 130,
      link_strength: 0.5,
      center_strength: 0.02,
      radial_radius: 400,
      radial_strength: 0.2,
      velocity_decay: 0.4,
      alpha_decay: 0.02,
    },

    data_sources: [
      {
        table: "filing_readiness",
        query: "SELECT * FROM filing_readiness ORDER BY confidence DESC",
        node_factory: "createFilingNode",
      },
    ],
  },

  6: {
    id: 6,
    name: "TIMELINE",
    description: "Chronological event nodes — the temporal spine",
    color_primary: "#f1c40f",
    color_secondary: "#f39c12",
    background: "radial-gradient(circle, #2d2d0d 0%, #181806 100%)",
    node_types: ["EVENT"],
    subtypes: ["hearing", "order_entered", "violation", "incident",
               "contact", "withholding", "arrest", "filing_event"],
    max_nodes: 300,
    always_visible: false,
    z_order: 60,

    force: {
      charge: -80,
      charge_min_distance: 15,
      charge_max_distance: 180,
      collision_radius: 12,
      link_distance: 60,
      link_strength: 0.2,
      center_strength: 0.01,
      radial_radius: 600,
      radial_strength: 0.15,
      velocity_decay: 0.5,
      alpha_decay: 0.03,
      // Timeline uses x-axis for temporal ordering
      temporal_force: true,
      temporal_scale: "linear",   // linear|log
      temporal_strength: 0.8,
    },

    data_sources: [
      {
        table: "timeline_events",
        query: `SELECT event_date, event_text, category, lane, actor
                FROM timeline_events
                WHERE event_date IS NOT NULL
                ORDER BY event_date DESC
                LIMIT 300`,
        node_factory: "createEventNode",
      },
    ],
  },

  7: {
    id: 7,
    name: "WEAPON",
    description: "Weaponized legal instruments — false allegations, PPO abuse, contempt",
    color_primary: "#e74c3c",
    color_secondary: "#c0392b",
    background: "radial-gradient(circle, #2d0d0d 0%, #180606 100%)",
    node_types: ["WEAPON"],
    subtypes: ["false_allegation", "ppo_weaponization", "contempt_abuse",
               "evidence_suppression", "ex_parte_abuse", "medication_coercion",
               "incarceration_weapon"],
    max_nodes: 50,
    always_visible: false,
    z_order: 80,

    force: {
      charge: -300,
      charge_min_distance: 40,
      charge_max_distance: 350,
      collision_radius: 35,
      link_distance: 140,
      link_strength: 0.6,
      center_strength: 0.02,
      radial_radius: 300,
      radial_strength: 0.25,
      velocity_decay: 0.35,
      alpha_decay: 0.02,
    },

    data_sources: [
      {
        table: "impeachment_matrix",
        query: `SELECT category, evidence_summary, impeachment_value,
                       cross_exam_question, filing_relevance
                FROM impeachment_matrix
                WHERE impeachment_value >= 7
                ORDER BY impeachment_value DESC
                LIMIT 50`,
        node_factory: "createWeaponNode",
      },
    ],
  },

  8: {
    id: 8,
    name: "CARTEL",
    description: "Judicial cartel connections — McNeill/Hoopes/Ladas-Hoopes triangle",
    color_primary: "#ff0000",
    color_secondary: "#cc0000",
    background: "radial-gradient(circle, #2d0000 0%, #1a0000 100%)",
    node_types: ["PERSON", "INSTITUTION"],
    subtypes: ["judicial", "agency_official", "court", "law_firm"],
    max_nodes: 15,
    always_visible: false,
    z_order: 95,

    force: {
      charge: -500,
      charge_min_distance: 60,
      charge_max_distance: 400,
      collision_radius: 50,
      link_distance: 180,
      link_strength: 0.9,       // Tight coupling — cartel sticks together
      center_strength: 0.05,
      radial_radius: 200,
      radial_strength: 0.3,
      velocity_decay: 0.3,
      alpha_decay: 0.015,
    },

    data_sources: [
      {
        table: "berry_mcneill_intelligence",
        query: "SELECT * FROM berry_mcneill_intelligence ORDER BY created_at DESC",
        node_factory: "createCartelNode",
      },
    ],
  },

  9: {
    id: 9,
    name: "IMPEACHMENT",
    description: "Credibility attack chains — contradictions, lies, perjury",
    color_primary: "#e91e63",
    color_secondary: "#c2185b",
    background: "radial-gradient(circle, #2d0d1a 0%, #18060d 100%)",
    node_types: ["EVIDENCE", "WEAPON"],
    subtypes: ["contradiction", "impeachment_chain", "credibility_attack"],
    max_nodes: 100,
    always_visible: false,
    z_order: 78,

    force: {
      charge: -200,
      charge_min_distance: 25,
      charge_max_distance: 250,
      collision_radius: 22,
      link_distance: 100,
      link_strength: 0.5,
      center_strength: 0.02,
      radial_radius: 450,
      radial_strength: 0.2,
      velocity_decay: 0.45,
      alpha_decay: 0.025,
    },

    data_sources: [
      {
        table: "contradiction_map",
        query: `SELECT claim_id, source_a, source_b, contradiction_text, severity, lane
                FROM contradiction_map
                ORDER BY severity DESC
                LIMIT 100`,
        node_factory: "createImpeachmentNode",
      },
    ],
  },

  10: {
    id: 10,
    name: "CLAIM",
    description: "Causes of action — the 27 COAs across all lanes",
    color_primary: "#00bcd4",
    color_secondary: "#0097a7",
    background: "radial-gradient(circle, #0d2d2d 0%, #061818 100%)",
    node_types: ["CLAIM"],
    subtypes: ["cause_of_action", "defense", "counterclaim", "affirmative_defense"],
    max_nodes: 40,
    always_visible: false,
    z_order: 72,

    force: {
      charge: -250,
      charge_min_distance: 30,
      charge_max_distance: 300,
      collision_radius: 28,
      link_distance: 120,
      link_strength: 0.4,
      center_strength: 0.02,
      radial_radius: 500,
      radial_strength: 0.2,
      velocity_decay: 0.4,
      alpha_decay: 0.02,
    },

    data_sources: [
      {
        table: "claims",
        query: "SELECT * FROM claims WHERE status = 'active' ORDER BY strength_score DESC",
        node_factory: "createClaimNode",
      },
    ],
  },

  11: {
    id: 11,
    name: "EMERGENCE",
    description: "Auto-discovered cross-layer connections — DBSCAN clusters, novelty signals",
    color_primary: "#00e676",
    color_secondary: "#00c853",
    background: "radial-gradient(circle, #0d2d0d 0%, #061806 100%)",
    node_types: ["EVIDENCE", "PERSON", "EVENT"],
    subtypes: ["emergence_cluster", "novelty_signal", "gap_indicator"],
    max_nodes: 50,
    always_visible: false,
    z_order: 55,

    force: {
      charge: -120,
      charge_min_distance: 20,
      charge_max_distance: 200,
      collision_radius: 18,
      link_distance: 90,
      link_strength: 0.3,
      center_strength: 0.01,
      radial_radius: 650,
      radial_strength: 0.1,
      velocity_decay: 0.5,
      alpha_decay: 0.03,
    },

    data_sources: [],  // Computed from cross-layer analysis, not direct DB query
  },

  12: {
    id: 12,
    name: "PREDICTION",
    description: "Adversary behavior forecasts — escalation, retaliation, timing",
    color_primary: "#ff9100",
    color_secondary: "#ff6d00",
    background: "radial-gradient(circle, #2d1a0d 0%, #180d06 100%)",
    node_types: ["EVENT"],
    subtypes: ["prediction", "forecast", "counter_strategy"],
    max_nodes: 20,
    always_visible: false,
    z_order: 50,

    force: {
      charge: -200,
      charge_min_distance: 30,
      charge_max_distance: 250,
      collision_radius: 25,
      link_distance: 110,
      link_strength: 0.4,
      center_strength: 0.01,
      radial_radius: 700,
      radial_strength: 0.1,
      velocity_decay: 0.45,
      alpha_decay: 0.025,
    },

    data_sources: [],  // Computed from temporal pattern analysis
  },
};

4.2 Layer Interaction Rules

/**
 * Cross-layer link rules — which layers can connect to which.
 * This prevents visual spaghetti by restricting link routing.
 */
const LAYER_CONNECTIONS = {
  //        0  1  2  3  4  5  6  7  8  9  10 11 12
  /* 0 */  [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1],  // CORE connects to all
  /* 1 */  [1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 1],  // ADVERSARY
  /* 2 */  [1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 0, 0],  // JUDICIAL
  /* 3 */  [1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0],  // EVIDENCE
  /* 4 */  [1, 0, 1, 1, 1, 1, 0, 0, 0, 0, 1, 0, 0],  // AUTHORITY
  /* 5 */  [1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 0, 0],  // FILING
  /* 6 */  [1, 1, 1, 1, 0, 1, 1, 1, 0, 0, 0, 1, 1],  // TIMELINE
  /* 7 */  [1, 1, 1, 1, 0, 0, 1, 1, 1, 1, 1, 0, 0],  // WEAPON
  /* 8 */  [1, 1, 1, 0, 0, 0, 0, 1, 1, 0, 0, 1, 0],  // CARTEL
  /* 9 */  [1, 1, 0, 1, 0, 0, 0, 1, 0, 1, 0, 1, 0],  // IMPEACHMENT
  /*10 */  [1, 0, 1, 1, 1, 1, 0, 1, 0, 0, 1, 0, 0],  // CLAIM
  /*11 */  [1, 1, 0, 1, 0, 0, 1, 0, 1, 1, 0, 1, 1],  // EMERGENCE
  /*12 */  [1, 1, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 1],  // PREDICTION
};

/**
 * Check if a link between two layers is permitted.
 */
function canConnect(sourceLayer, targetLayer) {
  return LAYER_CONNECTIONS[sourceLayer][targetLayer] === 1;
}

---

5. Graph Construction from Database

THEMANBEARPIG builds its graph entirely from litigation_context.db. No hardcoded data beyond the CORE_PERSONS registry. This section defines the complete data pipeline.

5.1 Master Build Pipeline

"""
THEMANBEARPIG Graph Builder — Master Pipeline
Reads from litigation_context.db, produces JSON graph for D3.js.

Usage:
    python mbp_graph_builder.py --output build/mbp_graph.json
"""

import sqlite3
import json
import hashlib
from pathlib import Path
from datetime import date

DB_PATH = Path(r"C:\Users\andre\LitigationOS\litigation_context.db")

def connect():
    """Open WAL connection with optimal PRAGMAs."""
    conn = sqlite3.connect(str(DB_PATH))
    conn.row_factory = sqlite3.Row
    conn.execute("PRAGMA busy_timeout = 60000")
    conn.execute("PRAGMA journal_mode = WAL")
    conn.execute("PRAGMA cache_size = -32000")
    conn.execute("PRAGMA temp_store = MEMORY")
    conn.execute("PRAGMA synchronous = NORMAL")
    return conn


def stable_id(*parts):
    """Generate a deterministic node ID from parts."""
    raw = "_".join(str(p).lower().replace(" ", "_") for p in parts)
    return raw[:80]  # Cap length for readability


def build_person_nodes(conn):
    """Build PERSON nodes from multiple source tables."""
    persons = {}

    # 1. Core persons (always present)
    for p in CORE_PERSONS:
        persons[p["id"]] = p

    # 2. From evidence_quotes — extract unique actors
    rows = conn.execute("""
        SELECT actor, lane, COUNT(*) as cnt
        FROM evidence_quotes
        WHERE actor IS NOT NULL AND actor != ''
        GROUP BY actor, lane
        ORDER BY cnt DESC
        LIMIT 100
    """).fetchall()

    for r in rows:
        pid = stable_id("person", "witness", r["actor"])
        if pid not in persons:
            persons[pid] = {
                "id": pid,
                "type": "PERSON",
                "subtype": "witness",
                "label": r["actor"],
                "short_label": r["actor"].split()[-1] if r["actor"] else "Unknown",
                "alignment": "unknown",
                "threat_level": 3,
                "evidence_count": r["cnt"],
                "layer": 1 if r["lane"] in ("A", "D") else 2,
                "case_lanes": [r["lane"]] if r["lane"] else [],
            }
        else:
            persons[pid]["evidence_count"] = persons[pid].get("evidence_count", 0) + r["cnt"]

    # 3. From impeachment_matrix — adversary targets
    rows = conn.execute("""
        SELECT target, COUNT(*) as cnt, MAX(impeachment_value) as max_imp
        FROM impeachment_matrix
        WHERE target IS NOT NULL
        GROUP BY target
        ORDER BY cnt DESC
    """).fetchall()

    for r in rows:
        pid = stable_id("person", "adversary", r["target"])
        if pid in persons:
            persons[pid]["impeachment_count"] = r["cnt"]
            persons[pid]["threat_level"] = min(10, max(
                persons[pid].get("threat_level", 0),
                int(r["max_imp"] or 0)
            ))
        # New persons from impeachment are already captured as CORE_PERSONS

    # 4. From judicial_violations — judges
    rows = conn.execute("""
        SELECT judge_name, COUNT(*) as cnt
        FROM judicial_violations
        WHERE judge_name IS NOT NULL
        GROUP BY judge_name
    """).fetchall()

    for r in rows:
        pid = stable_id("person", "judicial", r["judge_name"])
        if pid in persons:
            persons[pid]["violation_count"] = r["cnt"]

    return list(persons.values())


def build_evidence_nodes(conn):
    """Build EVIDENCE nodes from evidence_quotes grouped by source_file."""
    rows = conn.execute("""
        SELECT source_file, category, lane,
               COUNT(*) as quote_count,
               MAX(CAST(relevance_score AS REAL)) as max_relevance,
               MIN(event_date) as earliest_date
        FROM evidence_quotes
        WHERE source_file IS NOT NULL AND is_duplicate = 0
        GROUP BY source_file
        HAVING quote_count >= 2
        ORDER BY quote_count DESC
        LIMIT 500
    """).fetchall()

    nodes = []
    for r in rows:
        eid = stable_id("evidence", r["category"] or "doc", r["source_file"])
        fname = Path(r["source_file"]).name if r["source_file"] else "Unknown"
        nodes.append({
            "id": eid,
            "type": "EVIDENCE",
            "subtype": classify_evidence_subtype(fname),
            "label": fname[:50],
            "short_label": fname[:20],
            "source_file": r["source_file"],
            "category": r["category"],
            "lane": r["lane"],
            "relevance_score": r["max_relevance"] or 0.5,
            "quote_count": r["quote_count"],
            "date": r["earliest_date"],
            "layer": 3,
        })
    return nodes


def classify_evidence_subtype(filename):
    """Classify evidence subtype from filename."""
    fn = filename.lower()
    if any(x in fn for x in ["nspd", "police", "incident"]):
        return "police_report"
    if any(x in fn for x in [".mp3", ".wav", ".m4a", "audio", "recording"]):
        return "recording"
    if any(x in fn for x in [".mp4", ".mov", ".avi", "video"]):
        return "video"
    if any(x in fn for x in [".jpg", ".png", ".jpeg", "photo", "screenshot"]):
        return "photo"
    if any(x in fn for x in ["order", "judgment", "ruling"]):
        return "court_order"
    if any(x in fn for x in ["appclose", "email", "message", "text"]):
        return "communication"
    if any(x in fn for x in ["medical", "health", "eval"]):
        return "medical_record"
    return "document"


def build_authority_nodes(conn):
    """Build AUTHORITY nodes from authority_chains_v2."""
    rows = conn.execute("""
        SELECT primary_citation, source_type, lane,
               COUNT(*) as chain_count
        FROM authority_chains_v2
        WHERE primary_citation IS NOT NULL
        GROUP BY primary_citation
        ORDER BY chain_count DESC
        LIMIT 200
    """).fetchall()

    nodes = []
    for r in rows:
        aid = stable_id("authority", r["source_type"] or "rule", r["primary_citation"])
        citation = r["primary_citation"]
        nodes.append({
            "id": aid,
            "type": "AUTHORITY",
            "subtype": classify_authority_subtype(citation),
            "label": citation[:50],
            "short_label": citation[:20],
            "citation": citation,
            "source_type": r["source_type"],
            "lane": r["lane"],
            "chain_count": r["chain_count"],
            "layer": 4,
        })
    return nodes


def classify_authority_subtype(citation):
    """Classify authority subtype from citation text."""
    c = citation.upper()
    if "MCR" in c:
        return "court_rule"
    if "MCL" in c:
        return "statute"
    if "USC" in c or "FRCP" in c:
        return "federal_statute"
    if "CONST" in c or "AMEND" in c:
        return "constitutional"
    if " V " in c or " V. " in c:
        return "case_law"
    return "statute"


def build_event_nodes(conn):
    """Build EVENT nodes from timeline_events."""
    rows = conn.execute("""
        SELECT event_date, event_text, category, lane, actor
        FROM timeline_events
        WHERE event_date IS NOT NULL AND event_text IS NOT NULL
        ORDER BY event_date DESC
        LIMIT 300
    """).fetchall()

    nodes = []
    for r in rows:
        eid = stable_id("event", r["category"] or "generic",
                        r["event_date"], r["event_text"][:30])
        nodes.append({
            "id": eid,
            "type": "EVENT",
            "subtype": classify_event_subtype(r["category"]),
            "label": (r["event_text"] or "")[:60],
            "short_label": (r["event_text"] or "")[:25],
            "date": r["event_date"],
            "category": r["category"],
            "lane": r["lane"],
            "actor": r["actor"],
            "layer": 6,
        })
    return nodes


def classify_event_subtype(category):
    """Map timeline category to event subtype."""
    mapping = {
        "hearing": "hearing",
        "order": "order_entered",
        "violation": "violation",
        "incident": "incident",
        "contact": "contact",
        "withholding": "withholding",
        "arrest": "arrest",
        "filing": "filing_event",
    }
    return mapping.get((category or "").lower(), "incident")


def build_weapon_nodes(conn):
    """Build WEAPON nodes from impeachment_matrix (high-severity)."""
    rows = conn.execute("""
        SELECT category, evidence_summary, impeachment_value,
               cross_exam_question, filing_relevance, event_date
        FROM impeachment_matrix
        WHERE impeachment_value >= 7
        ORDER BY impeachment_value DESC
        LIMIT 50
    """).fetchall()

    nodes = []
    for r in rows:
        wid = stable_id("weapon", r["category"] or "unknown",
                        (r["evidence_summary"] or "")[:30])
        nodes.append({
            "id": wid,
            "type": "WEAPON",
            "subtype": classify_weapon_subtype(r["category"]),
            "label": (r["evidence_summary"] or "")[:60],
            "short_label": (r["category"] or "Unknown")[:25],
            "category": r["category"],
            "severity": r["impeachment_value"] or 5,
            "cross_exam_question": r["cross_exam_question"],
            "filing_relevance": r["filing_relevance"],
            "date": r["event_date"],
            "layer": 7,
        })
    return nodes


def classify_weapon_subtype(category):
    """Map impeachment category to weapon subtype."""
    cat = (category or "").lower()
    if "false" in cat or "allegation" in cat:
        return "false_allegation"
    if "ppo" in cat or "protection" in cat:
        return "ppo_weaponization"
    if "contempt" in cat or "jail" in cat:
        return "contempt_abuse"
    if "evidence" in cat or "suppress" in cat or "exclud" in cat:
        return "evidence_suppression"
    if "ex parte" in cat:
        return "ex_parte_abuse"
    if "medic" in cat or "prescription" in cat:
        return "medication_coercion"
    return "false_allegation"


def build_links(conn, nodes):
    """Build links from relationships in the database."""
    links = []
    node_ids = {n["id"] for n in nodes}

    # 1. CONTRADICTS links from contradiction_map
    rows = conn.execute("""
        SELECT source_a, source_b, contradiction_text, severity, lane
        FROM contradiction_map
        WHERE source_a IS NOT NULL AND source_b IS NOT NULL
        LIMIT 200
    """).fetchall()

    for r in rows:
        sid = stable_id("evidence", "doc", r["source_a"])
        tid = stable_id("evidence", "doc", r["source_b"])
        if sid in node_ids and tid in node_ids:
            links.append({
                "source": sid,
                "target": tid,
                "type": "CONTRADICTS",
                "strength": min(1.0, (r["severity"] or 5) / 10.0),
                "label": (r["contradiction_text"] or "")[:40],
                "style": "animated",
                "color": "#e74c3c",
                "width": max(1, (r["severity"] or 5) // 2),
            })

    # 2. VIOLATES links from judicial_violations
    rows = conn.execute("""
        SELECT judge_name, violation_type, COUNT(*) as cnt
        FROM judicial_violations
        WHERE judge_name IS NOT NULL
        GROUP BY judge_name, violation_type
        LIMIT 100
    """).fetchall()

    for r in rows:
        sid = stable_id("person", "judicial", r["judge_name"])
        tid = stable_id("weapon", r["violation_type"] or "violation",
                        r["judge_name"][:20])
        if sid in node_ids:
            links.append({
                "source": sid,
                "target": tid,
                "type": "VIOLATES",
                "strength": min(1.0, r["cnt"] / 100.0),
                "style": "animated",
                "color": "#e74c3c",
                "width": min(5, max(1, r["cnt"] // 20)),
            })

    # 3. CITES links from authority_chains_v2
    rows = conn.execute("""
        SELECT primary_citation, supporting_citation, relationship, lane
        FROM authority_chains_v2
        WHERE primary_citation IS NOT NULL AND supporting_citation IS NOT NULL
        GROUP BY primary_citation, supporting_citation
        LIMIT 300
    """).fetchall()

    for r in rows:
        sid = stable_id("authority", "rule", r["primary_citation"])
        tid = stable_id("authority", "rule", r["supporting_citation"])
        if sid in node_ids and tid in node_ids and sid != tid:
            links.append({
                "source": sid,
                "target": tid,
                "type": "CITES",
                "strength": 0.5,
                "style": "solid",
                "color": "#2ecc71",
                "width": 1,
            })

    # 4. CARTEL links (hardcoded — these are verified facts)
    cartel_links = [
        ("person_judicial_mcneill", "person_family_cavan_berry", "SPOUSE_OF"),
        ("person_judicial_mcneill", "person_judicial_hoopes", "FORMER_PARTNER"),
        ("person_judicial_mcneill", "person_judicial_ladas_hoopes", "FORMER_PARTNER"),
        ("person_judicial_hoopes", "person_judicial_ladas_hoopes", "MARRIED_TO"),
        ("person_family_cavan_berry", "person_agency_rusco", "SHARES_ADDRESS"),
        ("person_family_ronald_berry", "person_family_cavan_berry", "RELATED_TO"),
        ("person_family_ronald_berry", "person_adversary_emily_watson", "LIVES_WITH"),
        ("person_adversary_emily_watson", "person_family_albert_watson", "RELATED_TO"),
        ("person_adversary_emily_watson", "person_family_lori_watson", "RELATED_TO"),
    ]

    for src, tgt, ltype in cartel_links:
        if src in node_ids and tgt in node_ids:
            links.append({
                "source": src,
                "target": tgt,
                "type": ltype,
                **LINK_DEFAULTS.get(ltype, {}),
            })

    return links


def build_graph():
    """Master graph builder — produces complete D3.js-ready JSON."""
    conn = connect()

    # Build all node types
    persons = build_person_nodes(conn)
    evidence = build_evidence_nodes(conn)
    authorities = build_authority_nodes(conn)
    events = build_event_nodes(conn)
    weapons = build_weapon_nodes(conn)

    all_nodes = persons + evidence + authorities + events + weapons

    # Deduplicate by ID
    seen = set()
    unique_nodes = []
    for n in all_nodes:
        if n["id"] not in seen:
            seen.add(n["id"])
            unique_nodes.append(n)

    # Build links
    links = build_links(conn, unique_nodes)

    # Compute visual properties
    for node in unique_nodes:
        compute_node_visuals(node)

    # Compute separation counter dynamically
    sep_date = date(2025, 7, 29)
    sep_days = (date.today() - sep_date).days

    graph = {
        "metadata": {
            "generated": date.today().isoformat(),
            "node_count": len(unique_nodes),
            "link_count": len(links),
            "layer_count": 13,
            "separation_days": sep_days,
            "version": "1.0.0",
        },
        "nodes": unique_nodes,
        "links": links,
        "layers": LAYER_META,
    }

    conn.close()
    return graph

---

6. Node Sizing and Coloring Rules

Node visual properties are COMPUTED, never hardcoded. The computation depends on the node type and its associated metrics.

6.1 Sizing Rules

/**
 * Compute node radius based on type-specific metrics.
 * Returns radius in pixels (min 8, max 60).
 */
function computeRadius(node) {
  const MIN_RADIUS = 8;
  const MAX_RADIUS = 60;

  let raw;
  switch (node.type) {
    case "PERSON":
      // Threat level (0-10) is the primary driver
      // Evidence count adds secondary scaling
      raw = 15 + (node.threat_level || 0) * 3
            + Math.log2(1 + (node.evidence_count || 0)) * 2;
      break;

    case "EVIDENCE":
      // Relevance score (0-1) and quote count
      raw = 10 + (node.relevance_score || 0.5) * 15
            + Math.log2(1 + (node.quote_count || 0)) * 3;
      // Smoking guns get 50% boost
      if (node.is_smoking_gun) raw *= 1.5;
      break;

    case "AUTHORITY":
      // Chain count = how many things cite this authority
      raw = 10 + Math.log2(1 + (node.chain_count || 0)) * 5;
      break;

    case "FILING":
      // Readiness score (0-1) determines size
      raw = 15 + (node.readiness_score || 0) * 25;
      break;

    case "EVENT":
      // Significance (1-10)
      raw = 8 + (node.significance || 5) * 3;
      break;

    case "WEAPON":
      // Severity (1-10) with emphasis on high-severity weapons
      raw = 12 + (node.severity || 5) * 4;
      break;

    case "INSTITUTION":
      raw = 20;  // Fixed size for institutions
      break;

    case "CLAIM":
      // Strength score (0-1)
      raw = 12 + (node.strength_score || 0.5) * 30;
      break;

    default:
      raw = 12;
  }

  return Math.max(MIN_RADIUS, Math.min(MAX_RADIUS, raw));
}

6.2 Coloring Rules

/**
 * Color palette by node type and classification.
 */
const COLOR_PALETTE = {
  PERSON: {
    allied:    "#4fc3f7",    // Bright blue — friendly
    adverse:   "#ef5350",    // Red — adversary
    neutral:   "#bdbdbd",    // Gray — neutral
    protected: "#ffd54f",    // Gold — child (special)
    unknown:   "#78909c",    // Blue-gray
  },
  EVIDENCE: {
    police_report:   "#42a5f5",
    recording:       "#ab47bc",
    court_order:     "#5c6bc0",
    communication:   "#26a69a",
    photo:           "#66bb6a",
    video:           "#ab47bc",
    medical_record:  "#ec407a",
    financial_record:"#ffa726",
    document:        "#78909c",
  },
  AUTHORITY: {
    statute:          "#66bb6a",
    court_rule:       "#4db6ac",
    case_law:         "#81c784",
    constitutional:   "#aed581",
    federal_statute:  "#4dd0e1",
  },
  FILING: {
    DRAFT:          "#78909c",
    QA_REVIEW:      "#ffa726",
    SERVICE_READY:  "#66bb6a",
    FILED:          "#42a5f5",
    DOCKETED:       "#7e57c2",
    MONITORING:     "#26a69a",
  },
  EVENT: {
    favorable: "#66bb6a",
    adverse:   "#ef5350",
    neutral:   "#bdbdbd",
    mixed:     "#ffa726",
  },
  WEAPON: {
    low:    "#ff8a65",       // 1-3 severity
    medium: "#ef5350",       // 4-6 severity
    high:   "#d32f2f",       // 7-8 severity
    critical:"#b71c1c",      // 9-10 severity
  },
  INSTITUTION: {
    court:                 "#5c6bc0",
    agency:                "#7e57c2",
    law_firm:              "#8d6e63",
    law_enforcement_agency:"#546e7a",
  },
  CLAIM: {
    active:   "#4dd0e1",
    pending:  "#ffa726",
    dismissed:"#78909c",
  },
};


/**
 * Compute node color from type + classification.
 */
function computeColor(node) {
  const palette = COLOR_PALETTE[node.type];
  if (!palette) return "#78909c";

  switch (node.type) {
    case "PERSON":
      return palette[node.alignment] || palette.unknown;
    case "EVIDENCE":
      return palette[node.subtype] || palette.document;
    case "AUTHORITY":
      return palette[node.subtype] || palette.statute;
    case "FILING":
      return palette[node.status] || palette.DRAFT;
    case "EVENT":
      return palette[node.outcome] || palette.neutral;
    case "WEAPON": {
      const sev = node.severity || 5;
      if (sev >= 9) return palette.critical;
      if (sev >= 7) return palette.high;
      if (sev >= 4) return palette.medium;
      return palette.low;
    }
    case "INSTITUTION":
      return palette[node.subtype] || palette.court;
    case "CLAIM":
      return palette[node.status] || palette.active;
    default:
      return "#78909c";
  }
}


/**
 * Compute all visual properties for a node.
 */
function computeNodeVisuals(node) {
  node.radius = computeRadius(node);
  node.color = computeColor(node);

  // Glow effect for high-threat persons and smoking-gun evidence
  node.glow = (node.type === "PERSON" && (node.threat_level || 0) >= 8)
           || (node.type === "EVIDENCE" && node.is_smoking_gun)
           || (node.type === "WEAPON" && (node.severity || 0) >= 9);

  // Pulse animation for urgent items
  node.pulse = (node.type === "FILING" && node.urgency === "CRITICAL")
            || (node.type === "WEAPON" && (node.severity || 0) >= 8);

  // Opacity based on relevance (evidence nodes only)
  if (node.type === "EVIDENCE") {
    node.opacity = 0.4 + (node.relevance_score || 0.5) * 0.6;
  } else {
    node.opacity = 1.0;
  }
}

---

7. Performance Constraints and Optimization

THEMANBEARPIG must render 2,500+ nodes at 30+ FPS on modest hardware (AMD Ryzen 3 3200G, 24GB RAM, Vega 8 integrated GPU). This demands aggressive optimization at every layer.

7.1 Level of Detail (LOD) Rendering

/**
 * LOD system — reduce rendering complexity based on zoom level.
 *
 * Zoom Level  | Node Rendering        | Link Rendering      | Labels
 * ------------|----------------------|--------------------|---------
 * < 0.3       | Circles only (no icon)| Hidden             | Hidden
 * 0.3 - 0.6   | Circles + color       | Major links only   | Core only
 * 0.6 - 1.0   | Full icons            | All visible links  | Layer names
 * 1.0 - 2.0   | Icons + metrics       | All + animated     | Node labels
 * > 2.0       | Full detail + tooltips| All + labels       | Full text
 */
const LOD_THRESHOLDS = {
  MINIMAL:      0.3,    // Below this: dots only
  LOW:          0.6,    // Circles with color
  MEDIUM:       1.0,    // Full node rendering
  HIGH:         2.0,    // Detailed rendering
  ULTRA:        3.0,    // Maximum detail
};

function getLODLevel(zoomScale) {
  if (zoomScale < LOD_THRESHOLDS.MINIMAL) return "MINIMAL";
  if (zoomScale < LOD_THRESHOLDS.LOW) return "LOW";
  if (zoomScale < LOD_THRESHOLDS.MEDIUM) return "MEDIUM";
  if (zoomScale < LOD_THRESHOLDS.HIGH) return "HIGH";
  return "ULTRA";
}

function shouldRenderLabel(node, lod) {
  switch (lod) {
    case "MINIMAL": return false;
    case "LOW":     return node.layer === 0;  // Core actors only
    case "MEDIUM":  return node.type === "PERSON" || node.type === "FILING";
    case "HIGH":    return true;
    case "ULTRA":   return true;
    default:        return false;
  }
}

function shouldRenderIcon(node, lod) {
  switch (lod) {
    case "MINIMAL": return false;
    case "LOW":     return false;
    case "MEDIUM":  return true;
    case "HIGH":    return true;
    case "ULTRA":   return true;
    default:        return false;
  }
}

7.2 Viewport Culling with Quadtree

/**
 * Quadtree-based viewport culling.
 * Only render nodes within the visible viewport + margin.
 * D3's built-in d3.quadtree handles spatial indexing.
 */
function getVisibleNodes(nodes, viewport, margin = 100) {
  const { x, y, width, height } = viewport;
  const left   = x - margin;
  const right  = x + width + margin;
  const top    = y - margin;
  const bottom = y + height + margin;

  return nodes.filter(n =>
    n.x >= left && n.x <= right &&
    n.y >= top  && n.y <= bottom
  );
}

/**
 * For link culling: only render links where at least one endpoint is visible.
 */
function getVisibleLinks(links, visibleNodeIds) {
  const idSet = new Set(visibleNodeIds);
  return links.filter(l =>
    idSet.has(typeof l.source === "object" ? l.source.id : l.source) ||
    idSet.has(typeof l.target === "object" ? l.target.id : l.target)
  );
}

7.3 Canvas vs SVG Decision Matrix

NODE COUNT     RENDERER        REASON
< 500          SVG             DOM events, CSS styling, accessibility
500 - 2000     Canvas          Performance, still interactive via hit detection
2000 - 5000    Canvas + WebGL  GPU-accelerated rendering
> 5000         WebGL (PixiJS)  Full GPU, instanced rendering

/**
 * Renderer selection based on node count.
 */
function selectRenderer(nodeCount) {
  if (nodeCount < 500)  return "svg";
  if (nodeCount < 2000) return "canvas";
  if (nodeCount < 5000) return "canvas-webgl";
  return "webgl";
}

7.4 Force Simulation Performance

/**
 * Force simulation optimization for 2500+ nodes.
 *
 * Key optimizations:
 * 1. Barnes-Hut approximation (theta = 0.9) for charge forces
 * 2. Warm start — don't reset alpha on data updates
 * 3. Adaptive alpha decay — faster decay when layout stabilizes
 * 4. Worker thread for simulation (Web Worker)
 */
const SIMULATION_CONFIG = {
  // Barnes-Hut theta — higher = faster but less accurate
  // 0.9 is good for 2500+ nodes
  theta: 0.9,

  // Alpha parameters
  alpha_start: 1.0,
  alpha_min: 0.001,
  alpha_decay: 0.0228,        // Default D3 value
  alpha_target: 0,

  // Velocity decay — how quickly nodes slow down
  velocity_decay: 0.4,

  // Iteration budget per frame
  iterations_per_tick: 1,      // Keep at 1 for 60fps

  // Web Worker offloading
  use_worker: true,            // Offload simulation to worker thread
  worker_tick_rate: 16,        // ~60fps in worker

  // Warm restart on data change
  warm_restart_alpha: 0.3,    // Don't fully restart — just nudge
};

7.5 Memory Budget

COMPONENT          ESTIMATED SIZE     NOTES
2500 nodes         ~2.5 MB            ~1KB per node (properties + D3 internals)
5000 links         ~2.0 MB            ~400 bytes per link
Quadtree           ~1.0 MB            Spatial index
Canvas buffer      ~8.0 MB            1920x1080 @ 32bpp double-buffered
D3 simulation      ~3.0 MB            Force calculation state
Total              ~16.5 MB           Well within 24GB system RAM

---

8. Data Refresh Protocol

The graph should support incremental updates without full rebuild.

8.1 Incremental Update Strategy

/**
 * Incremental graph update — merge new data without resetting simulation.
 *
 * 1. Query DB for rows modified since last_updated timestamp
 * 2. Create/update nodes for changed rows
 * 3. Create/update links for changed relationships
 * 4. Remove nodes for deleted rows
 * 5. Warm-restart simulation (alpha = 0.3, not 1.0)
 */
function incrementalUpdate(graph, newNodes, newLinks) {
  const existingIds = new Set(graph.nodes.map(n => n.id));

  // Add new nodes
  for (const node of newNodes) {
    if (!existingIds.has(node.id)) {
      computeNodeVisuals(node);
      graph.nodes.push(node);
    } else {
      // Update existing node properties (preserve position)
      const existing = graph.nodes.find(n => n.id === node.id);
      Object.assign(existing, node, { x: existing.x, y: existing.y });
      computeNodeVisuals(existing);
    }
  }

  // Merge new links
  const existingLinkIds = new Set(graph.links.map(l =>
    `${l.source?.id || l.source}_${l.target?.id || l.target}_${l.type}`
  ));

  for (const link of newLinks) {
    const linkId = `${link.source}_${link.target}_${link.type}`;
    if (!existingLinkIds.has(linkId)) {
      graph.links.push(link);
    }
  }

  return graph;
}

8.2 Cache Strategy

/**
 * Graph cache with timestamp-based invalidation.
 * Full rebuild: every 5 minutes or on manual trigger.
 * Incremental: every 30 seconds when graph is visible.
 */
const CACHE_CONFIG = {
  full_rebuild_interval: 300000,    // 5 minutes in ms
  incremental_interval: 30000,      // 30 seconds
  storage_key: "mbp_graph_cache",
  max_cache_age: 3600000,           // 1 hour — force full rebuild
};

---

9. Interaction Model

9.1 Node Interactions

| Action | Result | |--------|--------| | Click | Select node → show detail panel with all properties | | Double-click | Expand node → show connected nodes (lazy load) | | Right-click | Context menu: "Show evidence", "Build impeachment", "Trace authority" | | Hover | Tooltip with key metrics (type, threat, evidence count) | | Drag | Move node, fix position (shift+drag to unfix) |

9.2 Layer Interactions

| Action | Result | |--------|--------| | Layer toggle | Show/hide entire layer (checkbox in HUD) | | Layer solo | Show only this layer + CORE (double-click layer name) | | Layer focus | Zoom to fit all nodes in this layer | | Layer opacity | Slider to fade non-focused layers |

9.3 Search

/**
 * Fuse.js fuzzy search across all nodes.
 * Returns ranked results with highlighting.
 */
const SEARCH_CONFIG = {
  keys: [
    { name: "label", weight: 0.4 },
    { name: "short_label", weight: 0.2 },
    { name: "citation", weight: 0.2 },
    { name: "key_quote", weight: 0.1 },
    { name: "id", weight: 0.1 },
  ],
  threshold: 0.3,             // Fuzzy match threshold
  includeMatches: true,
  minMatchCharLength: 2,
};

---

10. Export Formats

THEMANBEARPIG must support multiple export formats for different consumers:

| Format | Consumer | Content | |--------|----------|---------| | mbp_graph.json | D3.js renderer | Full graph with all properties | | neo4j_nodes.csv + neo4j_edges.csv | Neo4j import | Bloom-ready graph data | | gephi_gexf.xml | Gephi | GEXF format for static analysis | | mermaid.md | Documentation | Simplified Mermaid flowchart | | adjacency.json | NetworkX | Adjacency list for Python analysis |

---

11. Quality Gates

Graph Build PASS Gate

□ All CORE_PERSONS present (4 minimum)
□ Node count ≥ 100 (graph is not trivially empty)
□ Link count ≥ node_count * 0.5 (sufficient connectivity)
□ All 13 layers have ≥ 1 node each (or documented reason for empty)
□ No orphan nodes (every node has ≥ 1 link)
□ No duplicate node IDs
□ No self-referential links (source ≠ target)
□ Child node labeled "L.D.W." only (MCR 8.119(H))
□ No hallucinated persons (Jane Berry, Patricia Berry)
□ Separation counter computed dynamically
□ All node radii within [8, 60] range
□ All link widths within [1, 5] range
□ CARTEL links present (McNeill↔Hoopes↔Ladas-Hoopes triangle)
□ Frame rate ≥ 30 FPS at 2500 nodes
□ JSON output < 50 MB

Anti-Hallucination Checks (Graph-Specific)

BANNED_NODE_LABELS = [
    "Jane Berry", "Patricia Berry", "P35878",
    "Ron Berry, Esq", "Amy McNeill",
    "Emily Ann Watson", "Emily M. Watson",
    "Lincoln David Watson",  # Child's full name
]

def validate_graph(graph):
    """Validate graph against quality gates."""
    errors = []
    for node in graph["nodes"]:
        label = node.get("label", "")
        for banned in BANNED_NODE_LABELS:
            if banned.lower() in label.lower():
                errors.append(f"HALLUCINATION: Node '{node['id']}' contains banned label '{banned}'")

        # Child name check
        if node.get("subtype") == "child" and node.get("label") != "L.D.W.":
            errors.append(f"MCR 8.119(H) VIOLATION: Child node '{node['id']}' uses label '{node['label']}' instead of 'L.D.W.'")

    return errors

---

12. File Manifest

| File | Purpose | Size Target | |------|---------|-------------| | mbp_graph_builder.py | Master graph construction from DB | 800-1200 lines | | mbp_graph.json | Built graph output (D3.js ready) | 5-50 MB | | mbp_layer_config.js | LAYER_META + force simulation params | 300-500 lines | | mbp_node_types.js | Node type definitions + factories | 400-600 lines | | mbp_link_types.js | Link type definitions + factories | 200-400 lines | | mbp_visuals.js | Sizing, coloring, LOD rules | 300-500 lines | | mbp_performance.js | Culling, caching, worker offloading | 200-400 lines | | mbp_export.py | Export to Neo4j CSV, GEXF, Mermaid | 300-500 lines |

---

13. Cross-Skill Dependencies

This GENESIS skill is the foundation. Other MBP skills depend on it:

| Dependent Skill | Uses From GENESIS | |----------------|-------------------| | MBP-DATAWEAVE | Node schemas, link schemas, LAYER_META, DB query templates | | MBP-FORGE-RENDERER | LOD thresholds, color palette, sizing rules, renderer selection | | MBP-FORGE-PHYSICS | LAYER_META force parameters, simulation config | | MBP-FORGE-EFFECTS | Glow/pulse flags, color palette, opacity rules | | MBP-FORGE-DEPLOY | Graph JSON schema, file manifest, export formats | | MBP-COMBAT-ADVERSARY | PERSON node schema, threat_level, alignment | | MBP-COMBAT-WEAPONS | WEAPON node schema, weapon subtypes, severity | | MBP-COMBAT-JUDICIAL | CARTEL links, judicial node schema, violation_count | | MBP-COMBAT-EVIDENCE | EVIDENCE node schema, relevance_score, smoking_gun | | MBP-COMBAT-AUTHORITY | AUTHORITY node schema, chain_count, subtype classification | | MBP-COMBAT-IMPEACHMENT | IMPEACHMENT node schema, contradiction links | | MBP-INTERFACE-CONTROLS | Search config, interaction model, node click handlers | | MBP-INTERFACE-TIMELINE | EVENT node schema, temporal_force config | | MBP-INTERFACE-NARRATIVE | Layer descriptions, node labels, link labels | | MBP-INTERFACE-HUD | Layer toggle config, metadata display | | MBP-INTEGRATION-ENGINES | DB query templates, data_sources config | | MBP-INTEGRATION-FILING | FILING node schema, readiness_score, lane_id | | MBP-INTEGRATION-BRAINS | Graph JSON schema for brain feeding | | MBP-EMERGENCE-CONVERGENCE | EMERGENCE layer, cross-layer connection matrix | | MBP-EMERGENCE-PREDICTION | PREDICTION layer, temporal_force config | | MBP-EMERGENCE-SELFEVOLVE | LAYER_META structure for parameter learning | | MBP-TRANSCENDENCE-SONIC | threat_level → pitch mapping, severity → volume | | MBP-TRANSCENDENCE-DIMENSIONAL | Node position schema (x,y → x,y,z promotion) | | OMEGA-MBP-INDEX | Complete type registry, layer manifest, quality gates |

---

END OF SINGULARITY-MBP-GENESIS v1.0.0 — The Architectural DNA of THEMANBEARPIG 13 Layers · 20+ Node Types · 17+ Link Types · 2,500+ Nodes · Production-Ready