可以考虑把json替换为toon来节省token

[donnel666]

https://github.com/toon-format/toon-rust

[BigtoC]

我在这个PR #76 改了几行代码小试牛刀，确实是能节省token，虽然有将近一半的prompt只省了0.5%不到，但是能省则省，还是值得花些时间把json替代为toon。

@sopaco 我用deepwiki找到了一些需要改的地方，如果你觉得没问题，那我就继续改下去了。

🔍 Starting project preprocessing phase...
📁 Extracting project original document materials...
📁 Extracting project structure...
   💾 Cache write [structure] - Result cached
   🔭 Discovered 109 files, 45 directories
🎯 Identifying main source code files...
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: aml.rs
Estimated token usage - JSON: 664, TOON: 663, Saved: 1 (0.150602%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 655, TOON: 653, Saved: 2 (0.305344%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: contract.rs
Estimated token usage - JSON: 685, TOON: 684, Saved: 1 (0.145985%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: kyc.rs
Estimated token usage - JSON: 653, TOON: 652, Saved: 1 (0.153139%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 679, TOON: 677, Saved: 2 (0.294551%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: verify_signature.rs
Estimated token usage - JSON: 685, TOON: 683, Saved: 2 (0.291971%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: aml.rs
Estimated token usage - JSON: 663, TOON: 662, Saved: 1 (0.150830%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: kyc.rs
Estimated token usage - JSON: 680, TOON: 679, Saved: 1 (0.147059%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 688, TOON: 686, Saved: 2 (0.290698%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: hsm.rs
Estimated token usage - JSON: 674, TOON: 672, Saved: 2 (0.296736%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 669, TOON: 667, Saved: 2 (0.298954%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: lib.rs
Estimated token usage - JSON: 388, TOON: 386, Saved: 2 (0.515464%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: mod.rs
Estimated token usage - JSON: 420, TOON: 418, Saved: 2 (0.476190%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: lib.rs
Estimated token usage - JSON: 455, TOON: 453, Saved: 2 (0.439560%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: lib.rs
Estimated token usage - JSON: 413, TOON: 411, Saved: 2 (0.484262%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: mod.rs
Estimated token usage - JSON: 413, TOON: 411, Saved: 2 (0.484262%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: mod.rs
Estimated token usage - JSON: 415, TOON: 413, Saved: 2 (0.481928%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: mod.rs
Estimated token usage - JSON: 414, TOON: 412, Saved: 2 (0.483092%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: lib.rs
Estimated token usage - JSON: 479, TOON: 478, Saved: 1 (0.208768%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: mod.rs
Estimated token usage - JSON: 423, TOON: 421, Saved: 2 (0.472813%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: banner.rs
Estimated token usage - JSON: 655, TOON: 653, Saved: 2 (0.305344%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: mod.rs
Estimated token usage - JSON: 410, TOON: 408, Saved: 2 (0.487805%)
   💾 Cache write [ai_code_purpose] - Result cached
   Identified 91 main source code files
🤖 Analyzing core files using AI...
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: app.rs
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 1502, TOON: 1424, Saved: 78 (5.193076%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: aml.rs
Estimated token usage - JSON: 1505, TOON: 1388, Saved: 117 (7.774086%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 2109, TOON: 1891, Saved: 218 (10.336652%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: contract.rs
Estimated token usage - JSON: 2212, TOON: 2065, Saved: 147 (6.645570%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: main.rs
Estimated token usage - JSON: 1748, TOON: 1650, Saved: 98 (5.606407%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: app.rs
Estimated token usage - JSON: 891, TOON: 864, Saved: 27 (3.030303%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: kyc.rs
Estimated token usage - JSON: 1967, TOON: 1849, Saved: 118 (5.998983%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: routers.rs
Estimated token usage - JSON: 1303, TOON: 1223, Saved: 80 (6.139678%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 2044, TOON: 1851, Saved: 193 (9.442270%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 2459, TOON: 2043, Saved: 416 (16.917446%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: kyc.rs
Estimated token usage - JSON: 2170, TOON: 1903, Saved: 267 (12.304147%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: kyc.rs
Estimated token usage - JSON: 1996, TOON: 1822, Saved: 174 (8.717435%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 1425, TOON: 1339, Saved: 86 (6.035088%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: settings.rs
Estimated token usage - JSON: 2477, TOON: 2335, Saved: 142 (5.732741%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: aml.rs
Estimated token usage - JSON: 1596, TOON: 1602, Saved: 0 (0.000000%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 7189, TOON: 6699, Saved: 490 (6.815969%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: routers.rs
Estimated token usage - JSON: 1539, TOON: 1409, Saved: 130 (8.447044%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 2149, TOON: 2013, Saved: 136 (6.328525%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 1437, TOON: 1338, Saved: 99 (6.889353%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 2908, TOON: 2726, Saved: 182 (6.258597%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: aml.rs
Estimated token usage - JSON: 1976, TOON: 1760, Saved: 216 (10.931174%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: kyc.rs
Estimated token usage - JSON: 3125, TOON: 2970, Saved: 155 (4.960000%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 2618, TOON: 2466, Saved: 152 (5.805959%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: hsm.rs
Estimated token usage - JSON: 3033, TOON: 2578, Saved: 455 (15.001649%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 3554, TOON: 3323, Saved: 231 (6.499719%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: primary_sale.rs
Estimated token usage - JSON: 1730, TOON: 1556, Saved: 174 (10.057803%)
Estimated token usage - JSON: 1225, TOON: 1169, Saved: 56 (4.571429%)
   💾 Cache write [ai_code_insight] - Result cached
   💾 Cache write [ai_code_insight] - Result cached
Estimated token usage - JSON: 851, TOON: 849, Saved: 2 (0.235018%)
   💾 Cache write [ai_code_insight] - Result cached
✓ Concurrent code analysis completed, successfully analyzed 91 files
🔗 Analyzing component relationships...
   ⌛ Cache miss [ai_relationships_insights] - AI inference required
   🤖 Performing AI analysis: Dependency Relationship Analysis
Estimated token usage - JSON: 8802, TOON: 8251, Saved: 551 (6.259941%)
   💾 Cache write [ai_relationships_insights] - Result cached
✅ Project preprocessing completed, took 230.65 seconds
=== Preprocessing completed, results stored to Memory (Duration: 230.66s) ===
🚀 Starting Litho Studies Research investigation pipeline...
🤖 Executing System Context Research Report agent analysis...
   ⌛ Cache miss [studies_research/System Context Research Report] - AI inference required
   🤖 Performing AI analysis: System Context Research Report
Estimated token usage - JSON: 14950, TOON: 14866, Saved: 84 (0.561873%)
   💾 Cache write [studies_research/System Context Research Report] - Result cached
✅ Sub-Agent [System Context Research Report] execution completed
✓ System Context Research Report analysis completed
🤖 Executing Domain Modules Research Report agent analysis...
   ⌛ Cache miss [studies_research/Domain Modules Research Report] - AI inference required
   🤖 Performing AI analysis: Domain Modules Research Report
Estimated token usage - JSON: 22188, TOON: 21669, Saved: 519 (2.339102%)
   💾 Cache write [studies_research/Domain Modules Research Report] - Result cached

[sopaco]

我在这个PR #76 改了几行代码小试牛刀，确实是能节省token，虽然有将近一半的prompt只省了0.5%不到，但是能省则省，还是值得花些时间把json替代为toon。

@sopaco 我用deepwiki找到了一些需要改的地方，如果你觉得没问题，那我就继续改下去了。

🔍 Starting project preprocessing phase...
📁 Extracting project original document materials...
📁 Extracting project structure...
   💾 Cache write [structure] - Result cached
   🔭 Discovered 109 files, 45 directories
🎯 Identifying main source code files...
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: aml.rs
Estimated token usage - JSON: 664, TOON: 663, Saved: 1 (0.150602%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 655, TOON: 653, Saved: 2 (0.305344%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: contract.rs
Estimated token usage - JSON: 685, TOON: 684, Saved: 1 (0.145985%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: kyc.rs
Estimated token usage - JSON: 653, TOON: 652, Saved: 1 (0.153139%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 679, TOON: 677, Saved: 2 (0.294551%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: verify_signature.rs
Estimated token usage - JSON: 685, TOON: 683, Saved: 2 (0.291971%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: aml.rs
Estimated token usage - JSON: 663, TOON: 662, Saved: 1 (0.150830%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: kyc.rs
Estimated token usage - JSON: 680, TOON: 679, Saved: 1 (0.147059%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 688, TOON: 686, Saved: 2 (0.290698%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: hsm.rs
Estimated token usage - JSON: 674, TOON: 672, Saved: 2 (0.296736%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 669, TOON: 667, Saved: 2 (0.298954%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: lib.rs
Estimated token usage - JSON: 388, TOON: 386, Saved: 2 (0.515464%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: mod.rs
Estimated token usage - JSON: 420, TOON: 418, Saved: 2 (0.476190%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: lib.rs
Estimated token usage - JSON: 455, TOON: 453, Saved: 2 (0.439560%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: lib.rs
Estimated token usage - JSON: 413, TOON: 411, Saved: 2 (0.484262%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: mod.rs
Estimated token usage - JSON: 413, TOON: 411, Saved: 2 (0.484262%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: mod.rs
Estimated token usage - JSON: 415, TOON: 413, Saved: 2 (0.481928%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: mod.rs
Estimated token usage - JSON: 414, TOON: 412, Saved: 2 (0.483092%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: lib.rs
Estimated token usage - JSON: 479, TOON: 478, Saved: 1 (0.208768%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: mod.rs
Estimated token usage - JSON: 423, TOON: 421, Saved: 2 (0.472813%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: banner.rs
Estimated token usage - JSON: 655, TOON: 653, Saved: 2 (0.305344%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: mod.rs
Estimated token usage - JSON: 410, TOON: 408, Saved: 2 (0.487805%)
   💾 Cache write [ai_code_purpose] - Result cached
   Identified 91 main source code files
🤖 Analyzing core files using AI...
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: app.rs
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 1502, TOON: 1424, Saved: 78 (5.193076%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: aml.rs
Estimated token usage - JSON: 1505, TOON: 1388, Saved: 117 (7.774086%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 2109, TOON: 1891, Saved: 218 (10.336652%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: contract.rs
Estimated token usage - JSON: 2212, TOON: 2065, Saved: 147 (6.645570%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: main.rs
Estimated token usage - JSON: 1748, TOON: 1650, Saved: 98 (5.606407%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: app.rs
Estimated token usage - JSON: 891, TOON: 864, Saved: 27 (3.030303%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: kyc.rs
Estimated token usage - JSON: 1967, TOON: 1849, Saved: 118 (5.998983%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: routers.rs
Estimated token usage - JSON: 1303, TOON: 1223, Saved: 80 (6.139678%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 2044, TOON: 1851, Saved: 193 (9.442270%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 2459, TOON: 2043, Saved: 416 (16.917446%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: kyc.rs
Estimated token usage - JSON: 2170, TOON: 1903, Saved: 267 (12.304147%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: kyc.rs
Estimated token usage - JSON: 1996, TOON: 1822, Saved: 174 (8.717435%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 1425, TOON: 1339, Saved: 86 (6.035088%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: settings.rs
Estimated token usage - JSON: 2477, TOON: 2335, Saved: 142 (5.732741%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: aml.rs
Estimated token usage - JSON: 1596, TOON: 1602, Saved: 0 (0.000000%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 7189, TOON: 6699, Saved: 490 (6.815969%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: routers.rs
Estimated token usage - JSON: 1539, TOON: 1409, Saved: 130 (8.447044%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 2149, TOON: 2013, Saved: 136 (6.328525%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 1437, TOON: 1338, Saved: 99 (6.889353%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 2908, TOON: 2726, Saved: 182 (6.258597%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: aml.rs
Estimated token usage - JSON: 1976, TOON: 1760, Saved: 216 (10.931174%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: kyc.rs
Estimated token usage - JSON: 3125, TOON: 2970, Saved: 155 (4.960000%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 2618, TOON: 2466, Saved: 152 (5.805959%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: hsm.rs
Estimated token usage - JSON: 3033, TOON: 2578, Saved: 455 (15.001649%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 3554, TOON: 3323, Saved: 231 (6.499719%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: primary_sale.rs
Estimated token usage - JSON: 1730, TOON: 1556, Saved: 174 (10.057803%)
Estimated token usage - JSON: 1225, TOON: 1169, Saved: 56 (4.571429%)
   💾 Cache write [ai_code_insight] - Result cached
   💾 Cache write [ai_code_insight] - Result cached
Estimated token usage - JSON: 851, TOON: 849, Saved: 2 (0.235018%)
   💾 Cache write [ai_code_insight] - Result cached
✓ Concurrent code analysis completed, successfully analyzed 91 files
🔗 Analyzing component relationships...
   ⌛ Cache miss [ai_relationships_insights] - AI inference required
   🤖 Performing AI analysis: Dependency Relationship Analysis
Estimated token usage - JSON: 8802, TOON: 8251, Saved: 551 (6.259941%)
   💾 Cache write [ai_relationships_insights] - Result cached
✅ Project preprocessing completed, took 230.65 seconds
=== Preprocessing completed, results stored to Memory (Duration: 230.66s) ===
🚀 Starting Litho Studies Research investigation pipeline...
🤖 Executing System Context Research Report agent analysis...
   ⌛ Cache miss [studies_research/System Context Research Report] - AI inference required
   🤖 Performing AI analysis: System Context Research Report
Estimated token usage - JSON: 14950, TOON: 14866, Saved: 84 (0.561873%)
   💾 Cache write [studies_research/System Context Research Report] - Result cached
✅ Sub-Agent [System Context Research Report] execution completed
✓ System Context Research Report analysis completed
🤖 Executing Domain Modules Research Report agent analysis...
   ⌛ Cache miss [studies_research/Domain Modules Research Report] - AI inference required
   🤖 Performing AI analysis: Domain Modules Research Report
Estimated token usage - JSON: 22188, TOON: 21669, Saved: 519 (2.339102%)
   💾 Cache write [studies_research/Domain Modules Research Report] - Result cached

很好的尝试。不过PR代码中的Input和Output数据内的JSON并没有真正转为TOON结构，只是把最外层包装换了，实际效果应该不止0.5%。如果要修改的化需要修改或替换现在基于rig extractor的JSON生成，rig框架的extractor内部包含了对输出JSON格式的prompt注入和JSONSchema定义

可以先改造Ollama Extractor（https://github.com/sopaco/deepwiki-rs/pull/60），之前有人提交了这个代码，是针对ollama做的优化，这个是脱离rig框架实现的

[donnel666]

我在这个PR #76 改了几行代码小试牛刀，确实是能节省token，虽然有将近一半的prompt只省了0.5%不到，但是能省则省，还是值得花些时间把json替代为toon。
@sopaco 我用deepwiki找到了一些需要改的地方，如果你觉得没问题，那我就继续改下去了。

🔍 Starting project preprocessing phase...
📁 Extracting project original document materials...
📁 Extracting project structure...
   💾 Cache write [structure] - Result cached
   🔭 Discovered 109 files, 45 directories
🎯 Identifying main source code files...
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: aml.rs
Estimated token usage - JSON: 664, TOON: 663, Saved: 1 (0.150602%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 655, TOON: 653, Saved: 2 (0.305344%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: contract.rs
Estimated token usage - JSON: 685, TOON: 684, Saved: 1 (0.145985%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: kyc.rs
Estimated token usage - JSON: 653, TOON: 652, Saved: 1 (0.153139%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 679, TOON: 677, Saved: 2 (0.294551%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: verify_signature.rs
Estimated token usage - JSON: 685, TOON: 683, Saved: 2 (0.291971%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: aml.rs
Estimated token usage - JSON: 663, TOON: 662, Saved: 1 (0.150830%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: kyc.rs
Estimated token usage - JSON: 680, TOON: 679, Saved: 1 (0.147059%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 688, TOON: 686, Saved: 2 (0.290698%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: hsm.rs
Estimated token usage - JSON: 674, TOON: 672, Saved: 2 (0.296736%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 669, TOON: 667, Saved: 2 (0.298954%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: lib.rs
Estimated token usage - JSON: 388, TOON: 386, Saved: 2 (0.515464%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: mod.rs
Estimated token usage - JSON: 420, TOON: 418, Saved: 2 (0.476190%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: lib.rs
Estimated token usage - JSON: 455, TOON: 453, Saved: 2 (0.439560%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: lib.rs
Estimated token usage - JSON: 413, TOON: 411, Saved: 2 (0.484262%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: mod.rs
Estimated token usage - JSON: 413, TOON: 411, Saved: 2 (0.484262%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: mod.rs
Estimated token usage - JSON: 415, TOON: 413, Saved: 2 (0.481928%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: mod.rs
Estimated token usage - JSON: 414, TOON: 412, Saved: 2 (0.483092%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: lib.rs
Estimated token usage - JSON: 479, TOON: 478, Saved: 1 (0.208768%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: mod.rs
Estimated token usage - JSON: 423, TOON: 421, Saved: 2 (0.472813%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: banner.rs
Estimated token usage - JSON: 655, TOON: 653, Saved: 2 (0.305344%)
   💾 Cache write [ai_code_purpose] - Result cached
   ⌛ Cache miss [ai_code_purpose] - AI inference required
   🤖 Performing AI analysis: mod.rs
Estimated token usage - JSON: 410, TOON: 408, Saved: 2 (0.487805%)
   💾 Cache write [ai_code_purpose] - Result cached
   Identified 91 main source code files
🤖 Analyzing core files using AI...
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: app.rs
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 1502, TOON: 1424, Saved: 78 (5.193076%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: aml.rs
Estimated token usage - JSON: 1505, TOON: 1388, Saved: 117 (7.774086%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 2109, TOON: 1891, Saved: 218 (10.336652%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: contract.rs
Estimated token usage - JSON: 2212, TOON: 2065, Saved: 147 (6.645570%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: main.rs
Estimated token usage - JSON: 1748, TOON: 1650, Saved: 98 (5.606407%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: app.rs
Estimated token usage - JSON: 891, TOON: 864, Saved: 27 (3.030303%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: kyc.rs
Estimated token usage - JSON: 1967, TOON: 1849, Saved: 118 (5.998983%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: routers.rs
Estimated token usage - JSON: 1303, TOON: 1223, Saved: 80 (6.139678%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 2044, TOON: 1851, Saved: 193 (9.442270%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 2459, TOON: 2043, Saved: 416 (16.917446%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: kyc.rs
Estimated token usage - JSON: 2170, TOON: 1903, Saved: 267 (12.304147%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: kyc.rs
Estimated token usage - JSON: 1996, TOON: 1822, Saved: 174 (8.717435%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 1425, TOON: 1339, Saved: 86 (6.035088%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: settings.rs
Estimated token usage - JSON: 2477, TOON: 2335, Saved: 142 (5.732741%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: aml.rs
Estimated token usage - JSON: 1596, TOON: 1602, Saved: 0 (0.000000%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 7189, TOON: 6699, Saved: 490 (6.815969%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: routers.rs
Estimated token usage - JSON: 1539, TOON: 1409, Saved: 130 (8.447044%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 2149, TOON: 2013, Saved: 136 (6.328525%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 1437, TOON: 1338, Saved: 99 (6.889353%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 2908, TOON: 2726, Saved: 182 (6.258597%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: aml.rs
Estimated token usage - JSON: 1976, TOON: 1760, Saved: 216 (10.931174%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: kyc.rs
Estimated token usage - JSON: 3125, TOON: 2970, Saved: 155 (4.960000%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: webhook.rs
Estimated token usage - JSON: 2618, TOON: 2466, Saved: 152 (5.805959%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: hsm.rs
Estimated token usage - JSON: 3033, TOON: 2578, Saved: 455 (15.001649%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: wallet.rs
Estimated token usage - JSON: 3554, TOON: 3323, Saved: 231 (6.499719%)
   💾 Cache write [ai_code_insight] - Result cached
   ⌛ Cache miss [ai_code_insight] - AI inference required
   🤖 Performing AI analysis: primary_sale.rs
Estimated token usage - JSON: 1730, TOON: 1556, Saved: 174 (10.057803%)
Estimated token usage - JSON: 1225, TOON: 1169, Saved: 56 (4.571429%)
   💾 Cache write [ai_code_insight] - Result cached
   💾 Cache write [ai_code_insight] - Result cached
Estimated token usage - JSON: 851, TOON: 849, Saved: 2 (0.235018%)
   💾 Cache write [ai_code_insight] - Result cached
✓ Concurrent code analysis completed, successfully analyzed 91 files
🔗 Analyzing component relationships...
   ⌛ Cache miss [ai_relationships_insights] - AI inference required
   🤖 Performing AI analysis: Dependency Relationship Analysis
Estimated token usage - JSON: 8802, TOON: 8251, Saved: 551 (6.259941%)
   💾 Cache write [ai_relationships_insights] - Result cached
✅ Project preprocessing completed, took 230.65 seconds
=== Preprocessing completed, results stored to Memory (Duration: 230.66s) ===
🚀 Starting Litho Studies Research investigation pipeline...
🤖 Executing System Context Research Report agent analysis...
   ⌛ Cache miss [studies_research/System Context Research Report] - AI inference required
   🤖 Performing AI analysis: System Context Research Report
Estimated token usage - JSON: 14950, TOON: 14866, Saved: 84 (0.561873%)
   💾 Cache write [studies_research/System Context Research Report] - Result cached
✅ Sub-Agent [System Context Research Report] execution completed
✓ System Context Research Report analysis completed
🤖 Executing Domain Modules Research Report agent analysis...
   ⌛ Cache miss [studies_research/Domain Modules Research Report] - AI inference required
   🤖 Performing AI analysis: Domain Modules Research Report
Estimated token usage - JSON: 22188, TOON: 21669, Saved: 519 (2.339102%)
   💾 Cache write [studies_research/Domain Modules Research Report] - Result cached

很好的尝试。不过PR代码中的Input和Output数据内的JSON并没有真正转为TOON结构，只是把最外层包装换了，实际效果应该不止0.5%。如果要修改的化需要修改或替换现在基于rig extractor的JSON生成，rig框架的extractor内部包含了对输出JSON格式的prompt注入和JSONSchema定义

可以先改造Ollama Extractor（https://github.com/sopaco/deepwiki-rs/pull/60），之前有人提交了这个代码，是针对ollama做的优化，这个是脱离rig框架实现的

我还在研究这个项目，我对rust不是很熟悉。