|
|
8c6c8e71d2
|
docs: graph memory uses jieba keyword extraction via ExtractKeywords
- Indexer uses dual recall: char-bigram TF-IDF vector search + jieba keywords
- Document layer: char-bigram TF-IDF + jieba keyword extraction
- Memory Indexer description: add jieba keyword extraction
|
2026-07-17 07:04:05 +08:00 |
|
|
|
fb2a02af22
|
docs: update vectorization algorithm - LocalWordEmbedder (jieba + TF-IDF + PMI co-occurrence)
Context layer uses LocalWordEmbedder not plain TF-IDF:
- jieba segmentation, TF-IDF weights, sliding window PMI co-occurrence
- Document/Indexer layers still use char-bigram TF-IDF (TFIDFVectorizer)
- Add 'protect last 10 events' detail to Context pruning description
|
2026-07-17 07:02:53 +08:00 |
|
|
|
dd01388e76
|
docs: fix documentation-source mismatches across all doc files
- Go version: 1.19+/1.21+ → 1.25+ (matches go.mod)
- document/doc.go → document/document.go (actual filename)
- PluginSDK: 3 channels → 4 channels (add RegisterOutputChannel)
- InjectInput/InjectInterrupt: 3 params → 4 params (add eventType)
- RegisterChannel: document agentIO.Device interface requirement
- Architecture: remove non-existent snapshot/tokenizer dirs, add skill/meta
- Architecture: fix double internal/internal/sdk, fix Jaccard threshold
- Adapter: add http_get/http_post Lua VM built-ins
- Knowledge: 4-layer → 3-layer memory, CGO for go-sqlite3 not overlayfs
- Code structure tree: fix sdk/ and lua/adapters/ path consistency
|
2026-07-16 23:17:55 +08:00 |
|
|
|
f93b5eccb6
|
docs: update for v0.7.1 - output channels, restricted API, LLM chain event
|
2026-07-16 18:54:20 +08:00 |
|
|
|
c7e22fbe30
|
refactor: move webui.listen_addr from core.daemon to webui category
|
2026-07-14 11:22:46 +08:00 |
|
|
|
28dc095dc5
|
feat: 实现多语言文档架构
- 创建 docs/zh/ 和 docs/en/ 目录结构
- 迁移文档至多语言目录，添加中英文切换链接
- 创建 README_EN.md 英文版本
- 更新 README.md 添加语言切换链接
- 删除旧文档和临时文件
|
2026-07-13 23:29:35 +08:00 |
|
