rei-lang 0.4.0 → 0.5.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +241 -47
- package/bin/rei.js +1 -1
- package/dist/index.d.mts +562 -1
- package/dist/index.d.ts +562 -1
- package/dist/index.js +3592 -986
- package/dist/index.js.map +1 -1
- package/dist/index.mjs +3589 -987
- package/dist/index.mjs.map +1 -1
- package/package.json +7 -2
- package/theory/theories-67.ts +1696 -1696
package/README.md
CHANGED
|
@@ -1,8 +1,8 @@
|
|
|
1
1
|
# Rei (0₀式) — D-FUMT Computational Language
|
|
2
2
|
|
|
3
3
|
[](https://www.npmjs.com/package/rei-lang)
|
|
4
|
-
[](./LICENSE)
|
|
5
|
+
[]()
|
|
6
6
|
|
|
7
7
|
**Rei** (0₀式 / れいしき) is a mathematical computation language based on **D-FUMT** (Dimensional Fujimoto Universal Mathematical Theory). Its center-periphery patterns as language primitives achieve an **average 74% code reduction** over equivalent implementations in general-purpose languages.
|
|
8
8
|
|
|
@@ -10,6 +10,16 @@
|
|
|
10
10
|
|
|
11
11
|
---
|
|
12
12
|
|
|
13
|
+
## What's New in v0.5
|
|
14
|
+
|
|
15
|
+
v0.5 introduces a **self-organizing agent runtime** — entities perceive, decide, and act autonomously, coordinated by a conflict-resolving mediator.
|
|
16
|
+
|
|
17
|
+
- **EventBus** — Type-safe event-driven architecture with flow momentum tracking
|
|
18
|
+
- **Entity Agent** — Six-attribute autonomous agents (perceive → decide → act cycle)
|
|
19
|
+
- **Mediator** — Concurrent execution engine with conflict detection and resolution strategies
|
|
20
|
+
|
|
21
|
+
---
|
|
22
|
+
|
|
13
23
|
## Install
|
|
14
24
|
|
|
15
25
|
```bash
|
|
@@ -51,8 +61,8 @@ npx rei
|
|
|
51
61
|
|
|
52
62
|
```
|
|
53
63
|
╔══════════════════════════════════════════╗
|
|
54
|
-
║ Rei (0₀式) REPL v0.
|
|
55
|
-
║ D-FUMT Computational Language
|
|
64
|
+
║ Rei (0₀式) REPL v0.5.0 ║
|
|
65
|
+
║ D-FUMT Computational Language ║
|
|
56
66
|
╚══════════════════════════════════════════╝
|
|
57
67
|
|
|
58
68
|
零 > 𝕄{5; 1, 2, 3, 4} |> compute :weighted
|
|
@@ -71,13 +81,6 @@ compress karma(i, e, r)
|
|
|
71
81
|
npx rei program.rei
|
|
72
82
|
```
|
|
73
83
|
|
|
74
|
-
### Inline Evaluation
|
|
75
|
-
|
|
76
|
-
```bash
|
|
77
|
-
npx rei -e "2 + 3 * 4"
|
|
78
|
-
# → 14
|
|
79
|
-
```
|
|
80
|
-
|
|
81
84
|
---
|
|
82
85
|
|
|
83
86
|
## Language Features
|
|
@@ -146,17 +149,9 @@ g |> forward // solid → omega (Ω)
|
|
|
146
149
|
g.omega // 1
|
|
147
150
|
```
|
|
148
151
|
|
|
149
|
-
Or from the primordial dot (・):
|
|
150
|
-
|
|
151
|
-
```rei
|
|
152
|
-
let g = ・
|
|
153
|
-
g |> forward // dot
|
|
154
|
-
g |> forward // line
|
|
155
|
-
```
|
|
156
|
-
|
|
157
152
|
### Four-Valued Logic (四価0π)
|
|
158
153
|
|
|
159
|
-
Beyond true/false — Theory #21:
|
|
154
|
+
Beyond true/false — based on D-FUMT Theory #21:
|
|
160
155
|
|
|
161
156
|
```rei
|
|
162
157
|
⊤ // true
|
|
@@ -169,11 +164,177 @@ Beyond true/false — Theory #21:
|
|
|
169
164
|
⊥ ∨ ⊤ // ⊤
|
|
170
165
|
```
|
|
171
166
|
|
|
172
|
-
|
|
167
|
+
---
|
|
168
|
+
|
|
169
|
+
## v0.5 Agent Runtime
|
|
170
|
+
|
|
171
|
+
### EventBus (イベントバス)
|
|
172
|
+
|
|
173
|
+
Type-safe event system with flow momentum tracking. Events follow a `category:action` pattern (e.g., `entity:fuse`, `agent:act`, `space:diffuse`).
|
|
174
|
+
|
|
175
|
+
```typescript
|
|
176
|
+
import { rei, ReiEventBus } from 'rei-lang';
|
|
177
|
+
|
|
178
|
+
// Access via Evaluator
|
|
179
|
+
const ev = rei.evaluator();
|
|
180
|
+
const bus = ev.eventBus;
|
|
181
|
+
|
|
182
|
+
// Subscribe to events
|
|
183
|
+
bus.on('entity:fuse', (event) => {
|
|
184
|
+
console.log('Fusion occurred:', event.data);
|
|
185
|
+
});
|
|
186
|
+
|
|
187
|
+
// Subscribe with filter
|
|
188
|
+
bus.subscribe(
|
|
189
|
+
(e) => e.category === 'agent',
|
|
190
|
+
(e) => console.log(`Agent ${e.data.agentId}: ${e.action}`)
|
|
191
|
+
);
|
|
192
|
+
```
|
|
193
|
+
|
|
194
|
+
In Rei syntax with pipe commands:
|
|
195
|
+
|
|
196
|
+
```rei
|
|
197
|
+
// Emit a custom event
|
|
198
|
+
"mySource" |> emit("entity:transform", "data")
|
|
199
|
+
|
|
200
|
+
// Subscribe to events
|
|
201
|
+
"entity:*" |> subscribe
|
|
202
|
+
|
|
203
|
+
// Check flow momentum state
|
|
204
|
+
0 |> flow_momentum // → { state: "expanding", rate: 12.5, ... }
|
|
205
|
+
```
|
|
206
|
+
|
|
207
|
+
### Entity Agent (自律エンティティ)
|
|
208
|
+
|
|
209
|
+
Entities in Rei are autonomous agents with the six D-FUMT attributes (場・流れ・記憶・層・関係・意志), executing a perceive → decide → act lifecycle.
|
|
173
210
|
|
|
174
211
|
```rei
|
|
175
|
-
|
|
176
|
-
let
|
|
212
|
+
// Spawn an agent from a value
|
|
213
|
+
let a = 𝕄{10; 1, 2, 3} |> spawn
|
|
214
|
+
|
|
215
|
+
// Agent lifecycle
|
|
216
|
+
a |> perceive // observe environment → Perception
|
|
217
|
+
a |> decide // choose action → Decision
|
|
218
|
+
a |> act // execute decision → ActionResult
|
|
219
|
+
|
|
220
|
+
// Agent introspection
|
|
221
|
+
a |> agent_sigma // σ metadata: state, step, memory, bindings
|
|
222
|
+
a |> 自律σ // 日本語版
|
|
223
|
+
|
|
224
|
+
// Agent behaviors: reactive / proactive / cooperative / competitive / contemplative
|
|
225
|
+
```
|
|
226
|
+
|
|
227
|
+
```typescript
|
|
228
|
+
// Programmatic API
|
|
229
|
+
import { ReiAgent, AgentRegistry } from 'rei-lang';
|
|
230
|
+
|
|
231
|
+
const registry = new AgentRegistry();
|
|
232
|
+
const agent = registry.spawn(42, {
|
|
233
|
+
behavior: 'cooperative',
|
|
234
|
+
autonomyLevel: 0.8,
|
|
235
|
+
});
|
|
236
|
+
|
|
237
|
+
const perception = agent.perceive(registry);
|
|
238
|
+
const decision = agent.decide(perception);
|
|
239
|
+
const result = agent.act(registry, decision);
|
|
240
|
+
```
|
|
241
|
+
|
|
242
|
+
### Mediator (調停エンジン)
|
|
243
|
+
|
|
244
|
+
The Mediator coordinates multiple agents running concurrently, detecting and resolving conflicts with configurable strategies.
|
|
245
|
+
|
|
246
|
+
```rei
|
|
247
|
+
// Concurrent execution — all agents perceive → decide → resolve conflicts → act
|
|
248
|
+
0 |> mediate // 1 round of concurrent execution
|
|
249
|
+
0 |> mediate(5) // 5 rounds with convergence detection
|
|
250
|
+
0 |> mediate(10, 0.8) // max 10 rounds, convergence threshold 0.8
|
|
251
|
+
|
|
252
|
+
// Strategies: priority / cooperative / sequential / cancel_both / mediator
|
|
253
|
+
0 |> mediate_strategy("cooperative") // 協調 — merge conflicting intentions
|
|
254
|
+
0 |> 調停戦略("協調") // 日本語版
|
|
255
|
+
|
|
256
|
+
// Inter-agent messaging
|
|
257
|
+
"agent_a" |> mediate_message("agent_b", "hello") // point-to-point
|
|
258
|
+
"agent_a" |> mediate_broadcast("共有データ") // broadcast to all
|
|
259
|
+
|
|
260
|
+
// Mediator σ — statistics and convergence history
|
|
261
|
+
0 |> mediator_sigma // → { totalRounds, conflicts, convergenceHistory, ... }
|
|
262
|
+
0 |> 調停σ // 日本語版
|
|
263
|
+
```
|
|
264
|
+
|
|
265
|
+
Conflict types detected automatically: **target contention** (same target), **resource conflict** (shared resource), **mutual fuse** (A↔B fusion), **contradictory** (fuse vs separate).
|
|
266
|
+
|
|
267
|
+
Resolution strategies:
|
|
268
|
+
|
|
269
|
+
| Strategy | Japanese | Behavior |
|
|
270
|
+
|----------|----------|----------|
|
|
271
|
+
| `priority` | 優先 | Highest confidence × priority wins |
|
|
272
|
+
| `cooperative` | 協調 | Merge intentions into compromise |
|
|
273
|
+
| `sequential` | 順次 | Execute in priority order |
|
|
274
|
+
| `cancel_both` | 両方取消 | Cancel conflicting actions |
|
|
275
|
+
| `mediator` | 調停者 | Mediator's own judgment (中道) |
|
|
276
|
+
|
|
277
|
+
```typescript
|
|
278
|
+
// Programmatic API
|
|
279
|
+
import { ReiMediator, AgentRegistry } from 'rei-lang';
|
|
280
|
+
|
|
281
|
+
const registry = new AgentRegistry();
|
|
282
|
+
const mediator = new ReiMediator(registry);
|
|
283
|
+
|
|
284
|
+
// Spawn agents
|
|
285
|
+
registry.spawn(10, { behavior: 'cooperative' });
|
|
286
|
+
registry.spawn(20, { behavior: 'competitive' });
|
|
287
|
+
|
|
288
|
+
// Run concurrent rounds with convergence detection
|
|
289
|
+
const result = mediator.run({
|
|
290
|
+
maxRounds: 10,
|
|
291
|
+
convergenceThreshold: 0.8,
|
|
292
|
+
});
|
|
293
|
+
console.log(result.converged, result.totalRounds);
|
|
294
|
+
```
|
|
295
|
+
|
|
296
|
+
---
|
|
297
|
+
|
|
298
|
+
## Six Attributes (六属性)
|
|
299
|
+
|
|
300
|
+
Every entity in Rei carries six attributes from D-FUMT theory:
|
|
301
|
+
|
|
302
|
+
| Attribute | Japanese | Role |
|
|
303
|
+
|-----------|----------|------|
|
|
304
|
+
| Field (場) | ば | Spatial context and neighbors |
|
|
305
|
+
| Flow (流れ) | ながれ | Temporal momentum and EventBus connection |
|
|
306
|
+
| Memory (記憶) | きおく | History of observations and actions |
|
|
307
|
+
| Layer (層) | そう | Hierarchical depth and nesting |
|
|
308
|
+
| Relation (関係) | かんけい | Bindings between entities |
|
|
309
|
+
| Will (意志) | いし | Intention-driven computation strategy |
|
|
310
|
+
|
|
311
|
+
```rei
|
|
312
|
+
// Relation binding
|
|
313
|
+
let x = 42
|
|
314
|
+
let y = 100
|
|
315
|
+
x |> bind(y, "collaborator") // create a relation
|
|
316
|
+
|
|
317
|
+
// Will-driven computation
|
|
318
|
+
let m = 𝕄{5; 1, 2, 3}
|
|
319
|
+
m |> intend("maximize") // set intention
|
|
320
|
+
m |> will_compute // compute guided by will
|
|
321
|
+
```
|
|
322
|
+
|
|
323
|
+
---
|
|
324
|
+
|
|
325
|
+
## RCT Compression (Rei Compression Theory)
|
|
326
|
+
|
|
327
|
+
D-FUMT Theory #67 — generative compression outperforming gzip on structured data:
|
|
328
|
+
|
|
329
|
+
```rei
|
|
330
|
+
// Core compression (Direction 1-2)
|
|
331
|
+
let data = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]
|
|
332
|
+
data |> compress // → compact θ representation
|
|
333
|
+
data |> compress |> decompress // → lossless round-trip
|
|
334
|
+
|
|
335
|
+
// Semantic compression (Direction 3)
|
|
336
|
+
"let x = 𝕄{5; 1, 2, 3}" |> semantic_compress // → θ parameters
|
|
337
|
+
"let x = 𝕄{5; 1, 2, 3}" |> 意味圧縮 // 日本語版
|
|
177
338
|
```
|
|
178
339
|
|
|
179
340
|
---
|
|
@@ -199,39 +360,66 @@ Evaluate a string of Rei code. State persists across calls.
|
|
|
199
360
|
|
|
200
361
|
Clear all variable and function bindings.
|
|
201
362
|
|
|
202
|
-
### `rei.
|
|
203
|
-
|
|
204
|
-
Parse code and return the AST without evaluating.
|
|
363
|
+
### `rei.evaluator(): Evaluator`
|
|
205
364
|
|
|
206
|
-
|
|
207
|
-
|
|
208
|
-
Tokenize code and return the token stream.
|
|
365
|
+
Access the underlying Evaluator instance (for EventBus, AgentRegistry, etc.).
|
|
209
366
|
|
|
210
367
|
### Classes
|
|
211
368
|
|
|
212
|
-
|
|
213
|
-
|
|
214
|
-
|
|
215
|
-
|
|
369
|
+
| Class | Description |
|
|
370
|
+
|-------|-------------|
|
|
371
|
+
| `Lexer` | Tokenizer |
|
|
372
|
+
| `Parser` | Recursive descent parser |
|
|
373
|
+
| `Evaluator` | AST evaluator with environment/scope chain |
|
|
374
|
+
| `ReiEventBus` | Type-safe event system with flow momentum |
|
|
375
|
+
| `ReiAgent` | Autonomous entity with six attributes |
|
|
376
|
+
| `AgentRegistry` | Agent lifecycle management |
|
|
377
|
+
| `ReiMediator` | Concurrent execution engine with conflict resolution |
|
|
216
378
|
|
|
217
379
|
### Types
|
|
218
380
|
|
|
219
|
-
|
|
220
|
-
|
|
221
|
-
|
|
222
|
-
|
|
223
|
-
|
|
381
|
+
| Type | Shape |
|
|
382
|
+
|------|-------|
|
|
383
|
+
| `MultiDimNumber` | `{ center, neighbors, mode, weights? }` |
|
|
384
|
+
| `ReiExtended` | `{ base, order, subscripts, valStar() }` |
|
|
385
|
+
| `GenesisState` | `{ state, omega, history }` |
|
|
386
|
+
| `ReiFunction` | `{ name, params, body, closure }` |
|
|
387
|
+
| `Quad` | `{ value: 'top' \| 'bottom' \| 'topPi' \| 'bottomPi' }` |
|
|
388
|
+
| `ReiEvent` | `{ type, category, action, timestamp, data, source?, depth }` |
|
|
389
|
+
| `AgentSigma` | `{ state, kind, behavior, step, perception, decisions, memory }` |
|
|
390
|
+
| `MediatorSigma` | `{ totalRounds, conflicts, convergenceHistory, agentCount }` |
|
|
224
391
|
|
|
225
392
|
---
|
|
226
393
|
|
|
227
|
-
##
|
|
394
|
+
## Bilingual Pipe Commands (日本語対応)
|
|
395
|
+
|
|
396
|
+
All pipe commands have Japanese aliases:
|
|
397
|
+
|
|
398
|
+
| English | 日本語 | Description |
|
|
399
|
+
|---------|--------|-------------|
|
|
400
|
+
| `spawn` | `生成` | Create agent from value |
|
|
401
|
+
| `perceive` | `知覚` | Agent observes environment |
|
|
402
|
+
| `decide` | `判断` | Agent chooses action |
|
|
403
|
+
| `act` | `行動` | Agent executes decision |
|
|
404
|
+
| `agent_sigma` | `自律σ` | Agent metadata |
|
|
405
|
+
| `mediate` | `調停` | Concurrent round execution |
|
|
406
|
+
| `mediate_run` | `調停実行` | Multi-round execution |
|
|
407
|
+
| `mediator_sigma` | `調停σ` | Mediator statistics |
|
|
408
|
+
| `mediate_strategy` | `調停戦略` | Set conflict strategy |
|
|
409
|
+
| `mediate_message` | `調停通信` | Point-to-point message |
|
|
410
|
+
| `mediate_broadcast` | `調停放送` | Broadcast to all agents |
|
|
411
|
+
| `emit` | `発火` | Emit event |
|
|
412
|
+
| `subscribe` | `購読` | Subscribe to events |
|
|
413
|
+
| `bind` | `結合` | Create relation binding |
|
|
414
|
+
| `intend` | `意図` | Set entity intention |
|
|
415
|
+
| `compress` | `圧縮` | RCT compression |
|
|
416
|
+
| `decompress` | `復元` | RCT decompression |
|
|
228
417
|
|
|
229
|
-
|
|
418
|
+
---
|
|
419
|
+
|
|
420
|
+
## BNF Specification
|
|
230
421
|
|
|
231
|
-
|
|
232
|
-
- 45 keywords, 10 operators, 9 types
|
|
233
|
-
- Full backward compatibility with v0.1
|
|
234
|
-
- Complete operator precedence table
|
|
422
|
+
The complete BNF v0.3 specification is available in [`spec/`](./spec/).
|
|
235
423
|
|
|
236
424
|
---
|
|
237
425
|
|
|
@@ -239,10 +427,15 @@ Key features integrated from 21 D-FUMT theories:
|
|
|
239
427
|
|
|
240
428
|
Rei is grounded in **D-FUMT** (Dimensional Fujimoto Universal Mathematical Theory), a framework of 66 interconnected theories spanning pure mathematics to philosophy and AI consciousness. The language's core innovation — **center-periphery patterns as language primitives** — derives from D-FUMT's multi-dimensional number system theory.
|
|
241
429
|
|
|
430
|
+
See [`theory/`](./theory/) for the complete theoretical documentation.
|
|
431
|
+
|
|
242
432
|
---
|
|
243
|
-
|
|
433
|
+
|
|
434
|
+
## ☮️ Peace Use Clause / 平和利用条項
|
|
435
|
+
|
|
244
436
|
Rei is licensed under Apache 2.0 with an additional Peace Use Clause.
|
|
245
437
|
Rei は Apache 2.0 ライセンスに加え、平和利用条項が適用されます。
|
|
438
|
+
|
|
246
439
|
Rooted in the Buddhist concept of "Kū" (空, Emptiness) and D-FUMT's consciousness mathematics, Rei is designed exclusively for the peaceful advancement of humanity. This software may not be used for:
|
|
247
440
|
仏教の「空」の概念と D-FUMT の意識数学に基づき、Rei は人類の平和的発展のためにのみ設計されています。以下の目的での使用は禁止されています:
|
|
248
441
|
|
|
@@ -252,6 +445,7 @@ Rooted in the Buddhist concept of "Kū" (空, Emptiness) and D-FUMT's consciousn
|
|
|
252
445
|
|
|
253
446
|
✅ Education, research, humanitarian aid, and creative endeavors are encouraged.
|
|
254
447
|
✅ 教育・研究・人道支援・創造的活動での使用を推奨します。
|
|
255
|
-
See PEACE_USE_CLAUSE.md for full details.
|
|
256
448
|
|
|
257
|
-
|
|
449
|
+
See [PEACE_USE_CLAUSE.md](./PEACE_USE_CLAUSE.md) for full details.
|
|
450
|
+
|
|
451
|
+
Apache 2.0 © Nobuki Fujimoto
|