Files
ThothII/harness/README.md
T
marcopan 0a9ebddaec feat(harness): Onda 0b — setup workspace per-cliente + build LSH (D14a validato)
Repo workspace per-cliente tht-workspace-psd/ (git separato): 35 evidence Thoth
frontmatturate (da etl/docs/evidence/, non tutto etl/docs), psd.yaml con path
assoluti ancorati al repo, THT_DOCS_ROOT -> radice repo cliente. README documeta
il deployment shape (2 checkout + .env).

LSH build sul DWH reale (REST, VPN): 75737 valori / 491 colonne eligible -> indice
175M nel repo cliente. tht lsh query 'ablazione' -> 8 colonne (D14a non-collapsing
validato end-to-end: procedure_type, descrizione_procedura, intervento, ...).

Correzioni al piano eseguite durante l'implementazione:
- aggiunto 'tht schema introspect' (prerequisito di lsh build, omesso nel piano)
- uso -c (il cmd ha --config, non --workspace residuo)
- path assoluti nel psd.yaml (paths sono relativi alla CWD, non al file YAML)
- evidence corretta: solo etl/docs/evidence/ (35), non tutto etl/docs (895)

Test L2 riallineato: WORKSPACE -> psd.yaml nel repo cliente (non tht-test.yaml),
index_dir -> paths.indexes/'lsh', nome -> db_schema (non letterale 'datawarehouse').
Bug latente del test (path mismatch) mai emerso prima: ora PASS invece di SKIP.

Verifica: pytest 165 passed, 5 deselected; pytest -m l2 test_value_grounding_real PASS.
2026-06-27 15:31:03 +02:00

116 lines
4.4 KiB
Markdown

# ThothII harness — `tht`
The self-contained Pi layer of ThothII: a deterministic Python CLI (`tht`) + a Pi gate
extension (`.pi/extensions/tht-gate.js`) that runs the 8-phase NL→SQL workflow,
emitting and consuming widget-descriptor JSON. Derived from the reference implementation as a validated
starting point, **adapted** (not assumed reliable) to the ThothII contract.
## Install
```bash
cd harness
python -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"
```
The `tht` command is now on PATH. Node ≥ 20 is needed for the gate JS tests
(`npm test`).
## Configure
### `.env` (gitignored — secrets live ONLY here)
```bash
cp .env.example .env
# fill in: THT_PROFILE, THT_DB_*, THT_DWH_API_KEY, THT_VEC_API_KEY,
# THT_VEC_WRITE_API_KEY, THT_SSL_CA, THT_OLLAMA_URL, ...
```
Keys are never logged; URLs are fine. Rotate any key that appeared in chat.
### `workspaces/<name>.yaml`
A workspace wires the relational DWH + the pgvector (dual-key) + embeddings + evidence.
See `workspaces/tht.example.yaml`. `${THT_*}}` tokens expand from `.env`.
### Per-customer workspace repo (deployment shape)
ThothII is generic; **evidence content and LSH indexes are per-customer** and live in a
**separate repo**, not copied into `harness/`. The deployment is two checkouts + one `.env`:
```
ThothII/ generic harness (this repo)
└── harness/ tht CLI + .env (THT_* secrets, points at the customer repo)
tht-workspace-<customer>/ per-customer workspace repo (separate git repo)
├── <customer>.yaml workspace YAML (paths ABSOLUTE, anchored here)
├── evidence/ curated evidence markdown (frontmatter-structured, by domain)
├── artifacts/ derived artifacts (gitignored; regenerated by tht)
├── indexes/ LSH index (gitignored; regenerated by `tht lsh build`)
└── sessions/ per-session artifacts (gitignored)
```
`THT_DOCS_ROOT` in `.env` points at the **root** of the customer repo
(`evidence.source_root` + `evidence.evidence_dir` resolve the curated folder).
All `tht` commands take the customer workspace via `-c`:
```bash
tht schema introspect -c /path/to/tht-workspace-<customer>/<customer>.yaml
tht lsh build -c /path/to/tht-workspace-<customer>/<customer>.yaml
tht session new -c /path/to/tht-workspace-<customer>/<customer>.yaml
```
## The workflow
`workflow.yaml` is the **single source of workflow truth** (spec F2). Eight phases
(F1 chiarimento → F8 datamart). Edit it to change the workflow; `phase.py`, the gate,
and the skill all read from it — no mirrored constants to drift.
```bash
tht phase meta --json # the gate reads workflow facts from here
tht phase meta # human-readable
```
## Run
The harness runs inside Pi (`pi --mode rpc`, cwd = `harness/`):
```bash
pi --mode rpc
# in Pi: /nuova-domanda "<your question in natural language>"
```
The gate (`tht-gate.js`) presents reviewer widgets (widget-descriptor), the reviewer
decides, and `tht` persists decisions to the append-only ledger (`review_decisions.jsonl`).
See `docs/testing.md` for what each interaction level validates.
## Test
```bash
pytest # L0 (testcontainers, real Postgres) + L1 (pure logic + gate builders)
npm test # gate widget-builder golden + fuzzy tests (JS)
pytest -m l2 # L2: real GLM 5.2 + remote DWH (pre-release; needs .env + VPN + CA bundle)
```
L0 needs Docker (present on the dev machine). L2 is manual, non-deterministic, and
skips cleanly without `.env`. See `docs/testing.md` for the honest split of what each
level covers and — crucially — does **not** cover.
## Layout
```
tht/ Python package (CLI + workflow + phase + decisions + db/rest/mschema/
vectorstore/evidence/search/session + memory)
.pi/ Pi project (settings, prompts, themes, extensions/tht-gate.js + gate/)
workflow.yaml single source of workflow truth (F2)
workspaces/ workspace YAML definitions (D3)
scripts/ reader/writer RPC SQL for pgvector (D11)
tests/ L0 (testcontainers), L1 (logic + builders), L2 (real model + DB)
docs/ testing guide + workflow editing
```
## Reference
- Architecture spec: `docs/superpowers/specs/2026-06-25-thothii-architecture-design.md`
- Implementation plan: `docs/superpowers/plans/2026-06-25-harness-implementation.md`