acsets-relational-thinking

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

SKILL: ACSets Relational Thinking

SKILL: ACSets 关系型思维

Version: 2.0.0 Trit: 0 (ERGODIC) Domain: database, category-theory, rewriting Source: Topos Institute RelationalThinking Course + AlgebraicJulia

版本: 2.0.0 三元态: 0（遍历性）领域: 数据库、范畴论、重写来源: Topos Institute 关系型思维课程 + AlgebraicJulia

Overview

概述

ACSets (Attributed C-Sets) are functors X: C → Set where C is a small category (schema). This skill integrates:

RelationalThinking Course - Topos Institute's pedagogical approach
DPO Rewriting - Double Pushout graph transformation
Self-Play Loop - Query → Execute → Evaluate → Refine
GF(3) Conservation - Triadic skill composition

ACSets（属性C集合）是函子X: C → Set，其中C是一个小范畴（模式）。本技能整合了以下内容：

关系型思维课程 - Topos Institute的教学方法
DPO重写 - 双推出图变换
自循环机制 - 查询→执行→评估→优化
GF(3)守恒 - 三元技能组合

Core Concept: C-Set as Functor

核心概念：作为函子的C集合

Schema (Category C)          Instance (Functor X: C → Set)
┌─────────────────┐          ┌─────────────────────────────┐
│  E ──src──→ V   │    X     │  X(E) = {e1, e2, e3}        │
│    ←──tgt──     │   ───→   │  X(V) = {v1, v2}            │
└─────────────────┘          │  X(src): e1↦v1, e2↦v1, e3↦v2│
                             │  X(tgt): e1↦v2, e2↦v2, e3↦v1│
                             └─────────────────────────────┘

Schema (Category C)          Instance (Functor X: C → Set)
┌─────────────────┐          ┌─────────────────────────────┐
│  E ──src──→ V   │    X     │  X(E) = {e1, e2, e3}        │
│    ←──tgt──     │   ───→   │  X(V) = {v1, v2}            │
└─────────────────┘          │  X(src): e1↦v1, e2↦v1, e3↦v2│
                             │  X(tgt): e1↦v2, e2↦v2, e3↦v1│
                             └─────────────────────────────┘

Schema Definition

模式定义

Basic Graph Schema

基础图模式

julia

using Catlab.CategoricalAlgebra

@present SchGraph(FreeSchema) begin
  V::Ob                    # Vertices (object)
  E::Ob                    # Edges (object)
  src::Hom(E, V)           # Source morphism
  tgt::Hom(E, V)           # Target morphism
end

@acset_type Graph(SchGraph, index=[:src, :tgt])

julia

using Catlab.CategoricalAlgebra

@present SchGraph(FreeSchema) begin
  V::Ob                    # Vertices (object)
  E::Ob                    # Edges (object)
  src::Hom(E, V)           # Source morphism
  tgt::Hom(E, V)           # Target morphism
end

@acset_type Graph(SchGraph, index=[:src, :tgt])

Entity-Subtype Hierarchy (from RelationalThinking Ch8)

实体-子类型层级（来自《关系型思维》第8章）

julia

@present SchKitchen(FreeSchema) begin
  Entity::Ob
  
  Food::Ob
  food_in_on::Hom(Food, Entity)
  food_is_entity::Hom(Food, Entity)
  
  Kitchenware::Ob
  ware_in_on::Hom(Kitchenware, Entity)
  ware_is_entity::Hom(Kitchenware, Entity)
  
  BreadLoaf::Ob
  bread_loaf_is_food::Hom(BreadLoaf, Food)
  Knife::Ob
  knife_is_ware::Hom(Knife, Kitchenware)
end

@acset_type Kitchen(SchKitchen)

julia

@present SchKitchen(FreeSchema) begin
  Entity::Ob
  
  Food::Ob
  food_in_on::Hom(Food, Entity)
  food_is_entity::Hom(Food, Entity)
  
  Kitchenware::Ob
  ware_in_on::Hom(Kitchenware, Entity)
  ware_is_entity::Hom(Kitchenware, Entity)
  
  BreadLoaf::Ob
  bread_loaf_is_food::Hom(BreadLoaf, Food)
  Knife::Ob
  knife_is_ware::Hom(Knife, Kitchenware)
end

@acset_type Kitchen(SchKitchen)

DPO Rewriting (Double Pushout)

DPO重写（双推出）

The Pattern: L ← K → R

模式：L ← K → R

     L ←──l── K ──r──→ R
     │        │        │
match│        │        │
     ↓        ↓        ↓
     G ←───── D ──────→ H
       pushout      pushout
       complement

L = Find (what to match)
K = Keep (overlap preserved)
R = Replace (new structure)
G = Host graph
H = Result graph

     L ←──l── K ──r──→ R
     │        │        │
match│        │        │
     ↓        ↓        ↓
     G ←───── D ──────→ H
       pushout      pushout
       complement

L = 查找（要匹配的内容）
K = 保留（保留重叠部分）
R = 替换（新结构）
G = 宿主图
H = 结果图

Rule Definition

规则定义

julia

using AlgebraicRewriting

julia

using AlgebraicRewriting

Slice bread rule: adds BreadSlice when knife + loaf present

slice_bread = @migration(SchKitchen, begin L => @join begin loaf::BreadLoaf knife::Knife end R => @join begin loaf::BreadLoaf slice::BreadSlice food_in_on(bread_slice_is_food(slice)) == food_in_on(bread_loaf_is_food(loaf)) knife::Knife end K => @join begin loaf::BreadLoaf knife::Knife end end)

rule = make_rule(slice_bread, yKitchen)

undefined

rule = make_rule(slice_bread, yKitchen)

undefined

Apply Rewrite

应用重写

julia

matches = get_matches(rule, state)
new_state = rewrite_match(rule, matches[1])

julia

matches = get_matches(rule, state)
new_state = rewrite_match(rule, matches[1])

Self-Play Loop

自循环机制

The ACSet self-refinement monad:

Query₀ → Execute → Evaluate → Mine Patterns → Refine → Query₁ → ...

ACSet自优化单子：

Query₀ → Execute → Evaluate → Mine Patterns → Refine → Query₁ → ...

Implementation

实现

julia

struct SelfRefinementLoop
  schema::Presentation
  state::ACSet
  patterns::Vector{Pattern}
  generation::Int
end

function step!(loop::SelfRefinementLoop, rule::Rule)
  # Find matches
  matches = get_matches(rule, loop.state)
  
  # Evaluate each match
  evaluations = [evaluate_match(m, loop.patterns) for m in matches]
  
  # Mine new patterns from successful evaluations
  new_patterns = mine_patterns(evaluations)
  append!(loop.patterns, new_patterns)
  
  # Apply best match
  best = argmax(e -> e.score, evaluations)
  loop.state = rewrite_match(rule, matches[best.index])
  loop.generation += 1
  
  loop
end

julia

struct SelfRefinementLoop
  schema::Presentation
  state::ACSet
  patterns::Vector{Pattern}
  generation::Int
end

function step!(loop::SelfRefinementLoop, rule::Rule)
  # Find matches
  matches = get_matches(rule, loop.state)
  
  # Evaluate each match
  evaluations = [evaluate_match(m, loop.patterns) for m in matches]
  
  # Mine new patterns from successful evaluations
  new_patterns = mine_patterns(evaluations)
  append!(loop.patterns, new_patterns)
  
  # Apply best match
  best = argmax(e -> e.score, evaluations)
  loop.state = rewrite_match(rule, matches[best.index])
  loop.generation += 1
  
  loop
end

Convergence Criterion

收敛准则

julia

function converged(loop::SelfRefinementLoop; threshold=0.95)
  recent = loop.patterns[end-10:end]
  stability = std([p.score for p in recent])
  stability < (1 - threshold)
end

julia

function converged(loop::SelfRefinementLoop; threshold=0.95)
  recent = loop.patterns[end-10:end]
  stability = std([p.score for p in recent])
  stability < (1 - threshold)
end

GF(3) Integration

GF(3)集成

Trit Assignment

三元态分配

julia

function acset_to_trits(g::Graph, seed::UInt64)
  rng = SplitMix64(seed)
  trits = Int[]
  for e in parts(g, :E)
    h = next_u64!(rng)
    hue = (h >> 16 & 0xffff) / 65535.0 * 360
    trit = hue < 60 || hue >= 300 ? 1 :
           hue < 180 ? 0 : -1
    push!(trits, trit)
  end
  trits
end

julia

function acset_to_trits(g::Graph, seed::UInt64)
  rng = SplitMix64(seed)
  trits = Int[]
  for e in parts(g, :E)
    h = next_u64!(rng)
    hue = (h >> 16 & 0xffff) / 65535.0 * 360
    trit = hue < 60 || hue >= 300 ? 1 :
           hue < 180 ? 0 : -1
    push!(trits, trit)
  end
  trits
end

Conservation check

gf3_conserved(trits) = sum(trits) % 3 == 0

undefined

gf3_conserved(trits) = sum(trits) % 3 == 0

undefined

Synergistic Triads Containing ACSets

包含ACSets的协同三元组

clj-kondo-3color (-1) ⊗ acsets (0) ⊗ rama-gay-clojure (+1) = 0 ✓
three-match (-1) ⊗ acsets (0) ⊗ gay-mcp (+1) = 0 ✓
slime-lisp (-1) ⊗ acsets (0) ⊗ cider-clojure (+1) = 0 ✓
hatchery-papers (-1) ⊗ acsets (0) ⊗ frontend-design (+1) = 0 ✓

clj-kondo-3color (-1) ⊗ acsets (0) ⊗ rama-gay-clojure (+1) = 0 ✓
three-match (-1) ⊗ acsets (0) ⊗ gay-mcp (+1) = 0 ✓
slime-lisp (-1) ⊗ acsets (0) ⊗ cider-clojure (+1) = 0 ✓
hatchery-papers (-1) ⊗ acsets (0) ⊗ frontend-design (+1) = 0 ✓

Orthogonal Bundles

正交束

ACSets participates in the STRUCTURAL bundle:

Direction	Skills	Behavior
STRUCTURAL	clj-kondo(-1) ⊗ acsets(0) ⊗ rama-gay(+1)	Schema validation → transport → generation
TEMPORAL	three-match(-1) ⊗ unworld(0) ⊗ gay-mcp(+1)	Reduction → derivation → coloring
STRATEGIC	proofgeneral(-1) ⊗ glass-bead(0) ⊗ rubato(+1)	Verification → hopping → composition

ACSets属于STRUCTURAL（结构型）束：

方向	技能	行为
STRUCTURAL	clj-kondo(-1) ⊗ acsets(0) ⊗ rama-gay(+1)	模式验证 → 传输 → 生成
TEMPORAL	three-match(-1) ⊗ unworld(0) ⊗ gay-mcp(+1)	约简 → 推导 → 着色
STRATEGIC	proofgeneral(-1) ⊗ glass-bead(0) ⊗ rubato(+1)	验证 → 跳转 → 组合

Visual Conventions (from RelationalThinking)

视觉约定（来自《关系型思维》）

Concept	Visual
Schema object	Gray circle
Schema morphism	Colored arrow (cyan=src, blue=tgt)
Instance element	Filled shape
Morphism mapping	Slot/containment
Pushout	Merged regions with distinct colors
DPO rule	Thought bubble (L,K,R)

概念	视觉表现
模式对象	灰色圆形
模式态射	彩色箭头（青色=src，蓝色=tgt）
实例元素	填充形状
态射映射	插槽/包含关系
推出	不同颜色的合并区域
DPO规则	思维气泡（L,K,R）

Commands

命令

bash

undefined

bash

undefined

Schema operations

just acset-schema FILE # Display schema diagram just acset-instance FILE # Display instance elements just acset-morphism F G # Show homomorphism F → G

DPO rewriting

just acset-rule RULE STATE # Apply rewrite rule just acset-matches RULE STATE # Find all matches just acset-chain RULES STATE # Chain multiple rewrites

Self-play

just acset-selfplay SCHEMA # Run self-refinement loop just acset-patterns STATE # Mine patterns from state just acset-converge SCHEMA # Run until convergence

GF(3) operations

just acset-trits STATE SEED # Color state with seed just acset-gf3 STATE # Check GF(3) conservation just acset-triads # Show synergistic triads

---

just acset-trits STATE SEED # Color state with seed just acset-gf3 STATE # Check GF(3) conservation just acset-triads # Show synergistic triads

---

References

参考资料

Topos Institute

AlgebraicJulia

Papers

Related Skills

—

```
rama-gay-clojure
```
(+1) - Scalable backends with color tracing
```
clj-kondo-3color
```
(-1) - Schema validation/linting
```
glass-bead-game
```
(0) - World hopping across schemas
```
unworld
```
(0) - Derivational chains for state evolution
```
discohy-streams
```
(0) - DisCoPy categorical color streams

Skill Name: acsets-relational-thinking Type: Category-Theoretic Database / DPO Rewriting Trit: 0 (ERGODIC) GF(3): Conserved via triadic composition

—

acsets-relational-thinking

Original

Translation

SKILL: ACSets Relational Thinking

SKILL: ACSets 关系型思维

Overview

概述

Core Concept: C-Set as Functor

核心概念：作为函子的C集合

Schema Definition

模式定义

Basic Graph Schema

基础图模式

Entity-Subtype Hierarchy (from RelationalThinking Ch8)

实体-子类型层级（来自《关系型思维》第8章）

DPO Rewriting (Double Pushout)

DPO重写（双推出）

The Pattern: L ← K → R

模式：L ← K → R

Rule Definition

规则定义

Slice bread rule: adds BreadSlice when knife + loaf present

Slice bread rule: adds BreadSlice when knife + loaf present

Apply Rewrite

应用重写

Self-Play Loop

自循环机制

Implementation

实现

Convergence Criterion

收敛准则

GF(3) Integration

GF(3)集成

Trit Assignment

三元态分配

Conservation check

Conservation check

Synergistic Triads Containing ACSets

包含ACSets的协同三元组

Orthogonal Bundles

正交束

Visual Conventions (from RelationalThinking)

视觉约定（来自《关系型思维》）

Commands

命令

Schema operations

Schema operations

DPO rewriting

DPO rewriting

Self-play

Self-play

GF(3) operations

GF(3) operations

References

参考资料

Topos Institute

Topos Institute

AlgebraicJulia

AlgebraicJulia

Papers

相关技能

Related Skills