Spaces:
Running
Running
File size: 9,242 Bytes
ac68ae4 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 | import os
ENV_DEFAULTS = {
"BRAINPAT_TOKEN": "test-token",
"MODELS_MODE": "local",
"PIPELINE_MODE": "accurate",
"EMBEDDINGS_LOCAL_MODEL": "local-model",
"EMBEDDINGS_SMALL_MODEL": "small-model",
"EMBEDDING_NODES_DIMENSION": "3",
"EMBEDDING_TRIPLETS_DIMENSION": "3",
"EMBEDDING_OBSERVATIONS_DIMENSION": "3",
"EMBEDDING_DATA_DIMENSION": "3",
"EMBEDDING_RELATIONSHIPS_DIMENSION": "3",
"REDIS_HOST": "localhost",
"REDIS_PORT": "6379",
"NEO4J_HOST": "localhost",
"NEO4J_PORT": "7687",
"NEO4J_USERNAME": "neo4j",
"NEO4J_PASSWORD": "password",
"MILVUS_HOST": "localhost",
"MILVUS_PORT": "19530",
"MONGO_CONNECTION_STRING": "mongodb://localhost:27017",
"CELERY_WORKER_CONCURRENCY": "1",
"OLLAMA_HOST": "localhost",
"OLLAMA_PORT": "11434",
"OLLAMA_LLM_SMALL_MODEL": "small",
"OLLAMA_LLM_LARGE_MODEL": "large",
}
for key, value in ENV_DEFAULTS.items():
os.environ.setdefault(key, value)
from unittest.mock import MagicMock, patch
from src.core.agents.architect_agent import ArchitectAgent
from src.core.agents.scout_agent import ScoutEntity
from src.core.saving.architect_batch import (
BatchEndpoint,
BatchExtractResponse,
BatchRelationship,
)
from src.core.saving.ingest_cost import IngestCostLedger
def test_run_batch_extract_queues_valid_without_escalate():
agent = ArchitectAgent(
llm_adapter=MagicMock(),
cache_adapter=MagicMock(),
kg=MagicMock(),
vector_store=MagicMock(),
embeddings=MagicMock(),
ingestion_manager=MagicMock(),
)
source = "Alice hired Bob as a contractor."
entities = [
ScoutEntity(uuid="u1", name="Alice", type="PERSON"),
ScoutEntity(uuid="u2", name="Hiring", type="EVENT"),
ScoutEntity(uuid="u3", name="Bob", type="PERSON"),
]
payload = BatchExtractResponse(
relationships=[
BatchRelationship(
tail=BatchEndpoint(uuid="u1", name="Alice", type="PERSON"),
tip=BatchEndpoint(uuid="u2", name="Hiring", type="EVENT"),
name="MADE",
source_span="Alice hired Bob as a contractor",
),
BatchRelationship(
tail=BatchEndpoint(uuid="u2", name="Hiring", type="EVENT"),
tip=BatchEndpoint(uuid="u3", name="Bob", type="PERSON"),
name="TARGETED",
source_span="Alice hired Bob as a contractor",
),
]
)
with patch.object(agent, "_get_agent"), patch.object(
agent, "run_tooler"
) as tooler:
agent.agent = MagicMock()
agent.agent.invoke.return_value = {
"structured_response": payload,
"messages": [],
}
ledger = IngestCostLedger()
rels = agent.run_batch_extract(
source,
entities,
cost_ledger=ledger,
escalate=True,
timeout=5,
)
tooler.assert_not_called()
assert len(rels) == 2
assert ledger.architect_units == 1
assert ledger.architect_escalations == 0
assert ledger.architect_schema_calls == 1
def test_run_batch_extract_repairs_then_escalates_on_failure():
agent = ArchitectAgent(
llm_adapter=MagicMock(),
cache_adapter=MagicMock(),
kg=MagicMock(),
vector_store=MagicMock(),
embeddings=MagicMock(),
ingestion_manager=MagicMock(),
)
source = "Alice met Bob."
entities = [
ScoutEntity(uuid="u1", name="Alice", type="PERSON"),
ScoutEntity(uuid="u2", name="Bob", type="PERSON"),
]
bad = BatchExtractResponse(
relationships=[
BatchRelationship(
tail=BatchEndpoint(uuid="u1", name="PERSON", type="PERSON"),
tip=BatchEndpoint(uuid="u2", name="Bob", type="PERSON"),
name="MET",
source_span="Alice met Bob",
)
]
)
with patch.object(agent, "_get_agent"), patch.object(
agent, "run_tooler", return_value=[]
) as tooler:
agent.agent = MagicMock()
agent.agent.invoke.return_value = {
"structured_response": bad,
"messages": [],
}
ledger = IngestCostLedger()
agent.run_batch_extract(
source,
entities,
cost_ledger=ledger,
escalate=True,
timeout=5,
max_retries=1,
)
assert agent.agent.invoke.call_count == 2
tooler.assert_called_once()
assert ledger.architect_escalations == 1
assert ledger.escalate_rate == 1.0
assert ledger.architect_repair_calls == 1
def test_run_batch_extract_keeps_partial_without_escalate():
agent = ArchitectAgent(
llm_adapter=MagicMock(),
cache_adapter=MagicMock(),
kg=MagicMock(),
vector_store=MagicMock(),
embeddings=MagicMock(),
ingestion_manager=MagicMock(),
)
source = "Alice hired Bob as a contractor."
entities = [
ScoutEntity(uuid="u1", name="Alice", type="PERSON"),
ScoutEntity(uuid="u2", name="Hiring", type="EVENT"),
ScoutEntity(uuid="u3", name="Bob", type="PERSON"),
]
mixed = BatchExtractResponse(
relationships=[
BatchRelationship(
tail=BatchEndpoint(uuid="u1", name="Alice", type="PERSON"),
tip=BatchEndpoint(uuid="u2", name="Hiring", type="EVENT"),
name="MADE",
source_span="Alice hired Bob as a contractor",
),
BatchRelationship(
tail=BatchEndpoint(uuid="u2", name="Hiring", type="EVENT"),
tip=BatchEndpoint(uuid="u3", name="Bob", type="PERSON"),
name="TARGETED",
source_span="totally fabricated zebra claim",
),
BatchRelationship(
tail=BatchEndpoint(uuid="u1", name="PERSON", type="PERSON"),
tip=BatchEndpoint(uuid="u3", name="Bob", type="PERSON"),
name="MET",
source_span="Alice hired Bob",
),
]
)
with patch.object(agent, "_get_agent"), patch.object(
agent, "run_tooler"
) as tooler:
agent.agent = MagicMock()
agent.agent.invoke.return_value = {
"structured_response": mixed,
"messages": [],
}
ledger = IngestCostLedger()
rels = agent.run_batch_extract(
source,
entities,
cost_ledger=ledger,
escalate=True,
timeout=5,
max_retries=1,
)
assert agent.agent.invoke.call_count == 2
tooler.assert_not_called()
assert len(rels) == 1
assert rels[0].name == "MADE"
assert ledger.architect_escalations == 0
assert ledger.escalate_rate == 0.0
assert ledger.architect_repair_calls == 1
def test_run_batch_extract_prefers_primary_when_repair_worse():
agent = ArchitectAgent(
llm_adapter=MagicMock(),
cache_adapter=MagicMock(),
kg=MagicMock(),
vector_store=MagicMock(),
embeddings=MagicMock(),
ingestion_manager=MagicMock(),
)
source = "Alice hired Bob as a contractor."
entities = [
ScoutEntity(uuid="u1", name="Alice", type="PERSON"),
ScoutEntity(uuid="u2", name="Hiring", type="EVENT"),
ScoutEntity(uuid="u3", name="Bob", type="PERSON"),
]
primary = BatchExtractResponse(
relationships=[
BatchRelationship(
tail=BatchEndpoint(uuid="u1", name="Alice", type="PERSON"),
tip=BatchEndpoint(uuid="u2", name="Hiring", type="EVENT"),
name="MADE",
source_span="Alice hired Bob as a contractor",
),
BatchRelationship(
tail=BatchEndpoint(uuid="u1", name="PERSON", type="PERSON"),
tip=BatchEndpoint(uuid="u3", name="Bob", type="PERSON"),
name="MET",
source_span="Alice hired Bob",
),
]
)
worse_repair = BatchExtractResponse(
relationships=[
BatchRelationship(
tail=BatchEndpoint(uuid="u1", name="PERSON", type="PERSON"),
tip=BatchEndpoint(uuid="u3", name="Bob", type="PERSON"),
name="MET",
source_span="Alice hired Bob",
)
]
)
with patch.object(agent, "_get_agent"), patch.object(
agent, "run_tooler"
) as tooler:
agent.agent = MagicMock()
agent.agent.invoke.side_effect = [
{"structured_response": primary, "messages": []},
{"structured_response": worse_repair, "messages": []},
]
ledger = IngestCostLedger()
rels = agent.run_batch_extract(
source,
entities,
cost_ledger=ledger,
escalate=True,
timeout=5,
max_retries=1,
)
tooler.assert_not_called()
assert len(rels) == 1
assert rels[0].name == "MADE"
assert ledger.architect_escalations == 0
|