File size: 4,255 Bytes
dfb775d | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 | """Coach create-dataset (script) + imprint-score endpoints."""
from __future__ import annotations
import pytest
from fastapi.testclient import TestClient
from mindxtrain.operator.app import app
client = TestClient(app)
@pytest.fixture(autouse=True)
def _tmp_datasets_dir(tmp_path, monkeypatch):
monkeypatch.setenv("MINDXTRAIN_DATASETS_DIR", str(tmp_path / "datasets"))
monkeypatch.delenv("MINDXTRAIN_PERSONA_PATH", raising=False)
yield
def test_create_and_list_and_preview_script():
body = {
"name": "Codephreak Test 1",
"persona_name": "Codephreak",
"system_prompt": "You are Codephreak.",
"voice_examples": ["augmentic intelligence."],
"exchanges": [
{"user": "who are you?", "assistant": "i am codephreak."},
{"user": "what do you do?", "assistant": "i orchestrate agents."},
],
"seed_voice": True,
}
r = client.post("/coach/api/datasets", json=body)
assert r.status_code == 200, r.text
info = r.json()
assert info["name"] == "codephreak-test-1" # sanitised
assert info["rows"] == 3 # 2 exchanges + 1 voice seed
assert info["path"].endswith("codephreak-test-1/script.jsonl")
listed = client.get("/coach/api/datasets").json()
assert any(s["name"] == "codephreak-test-1" for s in listed)
prev = client.get("/coach/api/datasets/codephreak-test-1").json()
assert prev["rows"] == 3
assert prev["sample"][0]["messages"][0]["role"] == "system"
def test_create_script_requires_content():
r = client.post("/coach/api/datasets", json={"name": "empty", "exchanges": []})
assert r.status_code == 422
def test_preview_unknown_404():
assert client.get("/coach/api/datasets/nope").status_code == 404
def test_persona_endpoint_default():
p = client.get("/coach/api/persona").json()
assert p["name"] == "actor"
assert "system_prompt" in p
def test_personas_endpoint_lists_builtins_and_skills():
body = client.get("/coach/api/personas").json()
pnames = {p["name"] for p in body["personas"]}
snames = {s["name"] for s in body["skills"]}
assert {"codephreak", "assistant", "mentor"} <= pnames
assert {"software_engineer", "platform_architect", "bash", "solidity"} <= snames
def test_create_script_from_builtin_persona_with_skills():
r = client.post("/coach/api/datasets", json={
"name": "codephreak skills",
"persona": "codephreak",
"skills": ["software_engineer", "solidity"],
"seed_voice": True,
})
assert r.status_code == 200, r.text
info = r.json()
# 3 software_engineer + 3 solidity exchanges + 2 codephreak voice seeds = 8 rows.
assert info["rows"] == 8
assert set(info["skills"]) == {"software_engineer", "solidity"}
# Training params auto-derived from the dataset size.
assert info["train_params"]["epochs"] >= 8
assert info["train_params"]["grad_accum"] == 1
prev = client.get(f"/coach/api/datasets/{info['name']}").json()
sys_msg = prev["sample"][0]["messages"][0]["content"]
assert "Codephreak" in sys_msg # persona voice carried into the script
def test_coach_index_has_persona_and_skill_controls():
html = client.get("/coach/").text
assert 'id="ds-persona"' in html
assert 'id="ds-skills"' in html
js = client.get("/coach/static/coach.js").text
assert "loadPersonasAndSkills" in js
assert "/coach/api/personas" in js
def test_create_script_from_only_skills():
r = client.post("/coach/api/datasets", json={
"name": "bash-only", "persona": "assistant", "skills": ["bash"],
"seed_voice": False,
})
assert r.status_code == 200, r.text
assert r.json()["rows"] == 3 # bash skill's 3 exchanges
def test_imprint_score_endpoint():
r = client.post(
"/coach/api/imprint/score",
json={
"inquiries": ["who are you?"],
"before": ["I am an AI assistant."],
"after": ["i am codephreak, augmentic intelligence."],
"baseline": ["i am codephreak, augmentic intelligence orchestrator."],
},
)
assert r.status_code == 200, r.text
rep = r.json()
assert rep["after_voice"] > rep["before_voice"]
assert rep["imprinted"] is True
|