alys-akusa 0.1.8 → 0.1.19
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 +44 -12
- package/dist/index.cjs +2921 -221
- package/package.json +3 -2
package/README.md
CHANGED
|
@@ -1,31 +1,63 @@
|
|
|
1
1
|
# Alys CLI
|
|
2
2
|
|
|
3
|
-
Alys is the
|
|
3
|
+
Alys is the terminal-native runtime for autonomous AI data preparation.
|
|
4
4
|
|
|
5
5
|
```bash
|
|
6
6
|
npm install -g alys-akusa
|
|
7
|
-
alys
|
|
7
|
+
alys prepare ./company-docs
|
|
8
8
|
```
|
|
9
9
|
|
|
10
10
|
Or run without installing:
|
|
11
11
|
|
|
12
12
|
```bash
|
|
13
|
-
npx alys-akusa
|
|
13
|
+
npx alys-akusa prepare ./company-docs
|
|
14
|
+
npx alys-akusa audit ./knowledge-base
|
|
15
|
+
npx alys-akusa simulate-rag ./knowledge-base
|
|
16
|
+
npx alys-akusa improve ./knowledge-base
|
|
17
|
+
npx alys-akusa benchmark ./knowledge-base
|
|
18
|
+
npx alys-akusa ingest ./knowledge-base
|
|
19
|
+
npx alys-akusa finetune-ready ./support-tickets
|
|
14
20
|
```
|
|
15
21
|
|
|
16
|
-
|
|
17
|
-
|
|
18
|
-
|
|
22
|
+
Alys turns messy local knowledge into OpenAI fine-tuning JSONL, Anthropic instruction
|
|
23
|
+
records, RAG chunks, QA datasets, eval datasets, embeddings-ready corpora, manifests,
|
|
24
|
+
source lists, and AI readiness reports.
|
|
25
|
+
|
|
26
|
+
## What It Does
|
|
27
|
+
|
|
28
|
+
```text
|
|
29
|
+
messy knowledge
|
|
30
|
+
-> ingestion
|
|
31
|
+
-> parsing
|
|
32
|
+
-> chunking
|
|
33
|
+
-> deduplication
|
|
34
|
+
-> semantic grouping
|
|
35
|
+
-> grounding
|
|
36
|
+
-> evaluation
|
|
37
|
+
-> AI-ready exports
|
|
38
|
+
```
|
|
19
39
|
|
|
20
|
-
|
|
21
|
-
|
|
22
|
-
|
|
40
|
+
Local preparation writes artifacts under `~/Alys/prepared/{run_id}` and does not
|
|
41
|
+
require users to bring model API keys.
|
|
42
|
+
Retrieval simulation writes `retrieval-simulation.json` and `retrieval-simulation.md`
|
|
43
|
+
with hit rate, overlap failures, weak boundaries, unsupported-answer risk, and
|
|
44
|
+
hallucination risk.
|
|
45
|
+
Improve writes a safe optimized copy under `~/Alys/improved/{run_id}` with
|
|
46
|
+
`improved-rag-chunks.jsonl`, `cleaned-corpus.md`, and an improvement report. It
|
|
47
|
+
does not mutate original source files.
|
|
48
|
+
Benchmark writes `knowledge-benchmark.json` and `knowledge-benchmark.md` under
|
|
49
|
+
`~/Alys/benchmarks/{run_id}` to compare original vs improved retrieval behavior,
|
|
50
|
+
answer support, grounding, hallucination risk, and regressions.
|
|
23
51
|
|
|
24
52
|
## Local Development
|
|
25
53
|
|
|
26
54
|
```bash
|
|
27
55
|
pnpm install
|
|
28
|
-
pnpm alys
|
|
29
|
-
pnpm alys --
|
|
30
|
-
pnpm alys --
|
|
56
|
+
pnpm alys -- audit ./knowledge-base --yes
|
|
57
|
+
pnpm alys -- simulate-rag ./knowledge-base --queries 12 --top-k 5 --yes
|
|
58
|
+
pnpm alys -- improve ./knowledge-base --yes
|
|
59
|
+
pnpm alys -- benchmark ./knowledge-base --queries 16 --top-k 5 --yes
|
|
60
|
+
pnpm alys -- prepare ./docs --profile all --yes
|
|
61
|
+
pnpm alys -- ingest ./knowledge-base --profile rag,embeddings,eval --yes
|
|
62
|
+
pnpm alys -- finetune-ready ./tickets --profile openai,anthropic,qa --yes
|
|
31
63
|
```
|