tests/test_db.py

"""
PhD Research OS — Unit Tests (Phase 0 requirement: 20+ tests)
"""

import os
import sys
import json
import pytest

sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))

from phd_research_os.db import (
    init_db, get_db, create_claim, get_claim, update_claim, search_claims,
    create_source, get_source, create_goal, get_goals_by_priority,
    create_conflict, create_decision, create_override,
    log_api_usage, get_cost_summary, to_fixed, from_fixed
)

TEST_DB = "test_research_os.db"


@pytest.fixture(autouse=True)
def setup_teardown():
    """Create fresh DB for each test."""
    init_db(TEST_DB)
    yield
    if os.path.exists(TEST_DB):
        os.remove(TEST_DB)
    if os.path.exists(TEST_DB + "-wal"):
        os.remove(TEST_DB + "-wal")
    if os.path.exists(TEST_DB + "-shm"):
        os.remove(TEST_DB + "-shm")


def get_test_conn():
    return get_db(TEST_DB)


# ============================================================
# Fixed-Point Math Tests
# ============================================================

def test_fixed_point_conversion():
    assert to_fixed(0.85) == 850
    assert from_fixed(850) == 0.85

def test_fixed_point_precision():
    assert to_fixed(0.123) == 123
    assert to_fixed(0.999) == 999
    assert to_fixed(0.0) == 0
    assert to_fixed(1.0) == 1000

def test_fixed_point_roundtrip():
    for val in [0.0, 0.001, 0.5, 0.85, 0.999, 1.0]:
        assert abs(from_fixed(to_fixed(val)) - val) < 0.001


# ============================================================
# Claim CRUD Tests
# ============================================================

def test_create_claim_complete():
    conn = get_test_conn()
    cid = create_claim(conn, "Graphene shows mobility > 10000 cm²/Vs", "Fact", 0.92,
                       evidence_strength=0.95, study_quality_weight=1.0,
                       journal_tier_weight=1.0, completeness_penalty=1.0)
    claim = get_claim(conn, cid)
    assert claim is not None
    assert claim['status'] == 'Complete'
    assert claim['epistemic_tag'] == 'Fact'
    assert abs(claim['confidence'] - 0.92) < 0.01
    conn.close()

def test_create_claim_incomplete():
    conn = get_test_conn()
    cid = create_claim(conn, "Temperature may affect binding", "Hypothesis", 0.35,
                       missing_fields=["temperature", "pH", "ionic_strength"])
    claim = get_claim(conn, cid)
    assert claim['status'] == 'Incomplete'
    assert len(claim['missing_fields']) == 3
    conn.close()

def test_claim_id_format():
    conn = get_test_conn()
    cid = create_claim(conn, "Test", "Fact", 0.5)
    assert cid.startswith("CLM_")
    conn.close()

def test_claim_schema_version():
    conn = get_test_conn()
    cid = create_claim(conn, "Test", "Fact", 0.5)
    claim = get_claim(conn, cid)
    assert claim['schema_version'] == '1.0'
    conn.close()

def test_get_nonexistent_claim():
    conn = get_test_conn()
    assert get_claim(conn, "CLM_NONEXIST") is None
    conn.close()

def test_update_claim():
    conn = get_test_conn()
    cid = create_claim(conn, "Old text", "Fact", 0.5)
    update_claim(conn, cid, text="Updated text", confidence=0.9)
    claim = get_claim(conn, cid)
    assert claim['text'] == "Updated text"
    assert abs(claim['confidence'] - 0.9) < 0.01
    conn.close()

def test_search_claims_by_text():
    conn = get_test_conn()
    create_claim(conn, "Graphene biosensor detection", "Fact", 0.8)
    create_claim(conn, "Lithium battery cycling", "Fact", 0.7)
    results = search_claims(conn, query="graphene")
    assert len(results) == 1
    assert "graphene" in results[0]['text'].lower()
    conn.close()

def test_search_claims_by_epistemic_tag():
    conn = get_test_conn()
    create_claim(conn, "Measured value", "Fact", 0.9)
    create_claim(conn, "We hypothesize", "Hypothesis", 0.3)
    results = search_claims(conn, epistemic_tag="Hypothesis")
    assert len(results) == 1
    conn.close()

def test_search_claims_by_min_confidence():
    conn = get_test_conn()
    create_claim(conn, "High confidence", "Fact", 0.95)
    create_claim(conn, "Low confidence", "Hypothesis", 0.2)
    results = search_claims(conn, min_confidence=0.5)
    assert len(results) == 1
    assert results[0]['confidence'] >= 0.5
    conn.close()

def test_create_50_claims():
    """Phase 0 acceptance: Can create 50 Claim Objects programmatically."""
    conn = get_test_conn()
    ids = []
    for i in range(50):
        cid = create_claim(conn, f"Claim number {i}", 
                           ["Fact", "Interpretation", "Hypothesis"][i % 3],
                           round(0.1 + (i * 0.018), 3))
        ids.append(cid)
    assert len(ids) == 50
    assert len(set(ids)) == 50  # All unique
    conn.close()

def test_claim_parameters():
    conn = get_test_conn()
    params = {"temperature_C": 25.0, "pH": 7.4, "ionic_strength_mM": 10.0}
    cid = create_claim(conn, "Test", "Fact", 0.5, parameters=params)
    claim = get_claim(conn, cid)
    assert claim['parameters'] == params
    conn.close()


# ============================================================
# Source Tests
# ============================================================

def test_create_source():
    conn = get_test_conn()
    doi = create_source(conn, "10.1234/test", "Test Paper",
                        ["Author A", "Author B"], 2024, "Nature", 1)
    src = get_source(conn, doi)
    assert src is not None
    assert src['title'] == "Test Paper"
    assert len(src['authors']) == 2
    conn.close()

def test_source_not_found():
    conn = get_test_conn()
    assert get_source(conn, "10.9999/nonexist") is None
    conn.close()


# ============================================================
# Goal Tests
# ============================================================

def test_create_goal():
    conn = get_test_conn()
    gid = create_goal(conn, "Achieve sub-fM LOD", "high")
    goals = get_goals_by_priority(conn)
    assert len(goals) == 1
    assert goals[0]['priority'] == 'high'
    assert goals[0]['status'] == 'Active'
    conn.close()

def test_goals_sorted_by_priority():
    conn = get_test_conn()
    create_goal(conn, "Low priority", "low")
    create_goal(conn, "High priority", "high")
    create_goal(conn, "Medium priority", "medium")
    goals = get_goals_by_priority(conn)
    assert goals[0]['priority'] == 'high'
    assert goals[1]['priority'] == 'medium'
    assert goals[2]['priority'] == 'low'
    conn.close()


# ============================================================
# Conflict Tests
# ============================================================

def test_create_conflict():
    conn = get_test_conn()
    cid_a = create_claim(conn, "Sensitivity increases", "Fact", 0.8)
    cid_b = create_claim(conn, "Sensitivity decreases", "Fact", 0.7)
    conf_id = create_conflict(conn, cid_a, cid_b, "value_mismatch",
                              "Different surface chemistry",
                              ["surface treatment", "buffer composition"])
    
    row = conn.execute("SELECT * FROM conflicts WHERE conflict_id = ?", (conf_id,)).fetchone()
    assert row is not None
    assert dict(row)['hypothesis_confidence'] == 'low'  # ALWAYS low
    assert dict(row)['resolution_status'] == 'Unresolved'
    conn.close()


# ============================================================
# Override Tests
# ============================================================

def test_expert_override():
    conn = get_test_conn()
    cid = create_claim(conn, "Override test", "Fact", 0.5)
    ovr_id = create_override(conn, cid, "Dr. Smith",
                             "Direct experimental evidence", 0.95)
    
    claim = get_claim(conn, cid)
    assert abs(claim['confidence'] - 0.95) < 0.01
    assert claim['expert_override'] is not None
    assert claim['expert_override']['who'] == 'Dr. Smith'
    conn.close()


# ============================================================
# Decision Tests
# ============================================================

def test_create_decision():
    conn = get_test_conn()
    gid = create_goal(conn, "Test goal", "high")
    dec_id = create_decision(conn, "experiment", "Run control experiment",
                             0.72, gid, priority="high",
                             estimated_effort="2 weeks")
    
    row = conn.execute("SELECT * FROM decisions WHERE decision_id = ?", (dec_id,)).fetchone()
    assert row is not None
    assert dict(row)['status'] == 'Proposed'
    conn.close()


# ============================================================
# API Usage Logging Tests
# ============================================================

def test_api_usage_logging():
    conn = get_test_conn()
    log_api_usage(conn, "claude-haiku", 500, 200, 0.001, "claim_extraction")
    log_api_usage(conn, "claude-haiku", 300, 150, 0.0008, "conflict_detection")
    
    summary = get_cost_summary(conn, days=1)
    assert summary['num_calls'] == 2
    conn.close()


if __name__ == "__main__":
    pytest.main([__file__, "-v"])