Spaces:
Sleeping
Sleeping
File size: 11,870 Bytes
5b7955a | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 | """Integration tests — wire together multiple real components, mock Supabase & LLM."""
import unittest
from unittest.mock import MagicMock, patch
from rag_engine.prompts.context_template import build_query_prompt
from rag_engine.prompts.response_format import ResponseFormatter
from rag_engine.prompts.system_prompt import SYSTEM_PROMPT
from rag_engine.retrieval.context_builder import ContextBuilder
from rag_engine.retrieval.query_preprocessor import QueryPreprocessor
# ====================================================================== #
# TEST 1 — full retrieval-to-context pipeline
# ====================================================================== #
class TestFullRetrievalToContextPipeline(unittest.TestCase):
def test_full_retrieval_to_context_pipeline(self):
"""query → preprocess → filter extraction → context build"""
preprocessor = QueryPreprocessor()
context_builder = ContextBuilder()
query = "is flood damage covered"
processed = preprocessor.preprocess(query)
filters = preprocessor.extract_filters(processed, "POL-001")
mock_results = [
{
"content": "Section 3.1: Flood damage is excluded.",
"metadata": {
"section_name": "Exclusions",
"clause_type": "exclusion",
"policy_id": "POL-001",
},
"score": 0.94,
},
{
"content": "Section 2.1: Fire damage is covered up to $50,000.",
"metadata": {
"section_name": "Coverage",
"clause_type": "coverage",
"policy_id": "POL-001",
},
"score": 0.71,
},
]
context = context_builder.build(mock_results)
assert "?" in processed
assert filters["coverage_category"] == "flood"
assert filters["policy_id"] == "POL-001"
assert "Source 1" in context
assert "Exclusions" in context
assert "Section 3.1" in context
# ====================================================================== #
# TEST 2 — prompt assembly
# ====================================================================== #
class TestPromptAssembly(unittest.TestCase):
def test_prompt_assembly_with_context(self):
"""context + query → prompt is correctly assembled"""
context = "--- Source 1 ---\nSection: Exclusions\n\nFlood is excluded."
query = "Is flood covered?"
policy_id = "POL-001"
prompt = build_query_prompt(query, context, policy_id)
assert "Is flood covered?" in prompt
assert "Flood is excluded" in prompt
assert "POL-001" in prompt
assert "INSURANCE POLICY CONTEXT" in prompt
assert len(SYSTEM_PROMPT) > 200
assert "PolicyDecoder" in SYSTEM_PROMPT
# ====================================================================== #
# TEST 3 — response formatter full structure
# ====================================================================== #
class TestResponseFormatterFullStructure(unittest.TestCase):
def test_response_formatter_full_structure(self):
"""formatter returns all required fields with correct types"""
formatter = ResponseFormatter()
chunks = [
{
"content": "Flood is excluded under Section 3.1",
"metadata": {
"section_name": "Exclusions",
"clause_type": "exclusion",
"policy_id": "POL-TEST",
},
"score": 0.93,
"rerank_score": 0.97,
},
{
"content": "Fire is covered up to $50,000",
"metadata": {
"section_name": "Coverage",
"clause_type": "coverage",
"policy_id": "POL-TEST",
},
"score": 0.81,
"rerank_score": 0.85,
},
]
result = formatter.format_answer(
"Flood damage is not covered under this policy.",
"Is flood covered?",
chunks,
)
assert "answer" in result
assert "sources" in result
assert "source_count" in result
assert "policy_id" in result
assert result["answer"] == "Flood damage is not covered under this policy."
assert result["source_count"] == 2
assert result["policy_id"] == "POL-TEST"
assert len(result["sources"]) == 2
assert result["sources"][0]["section"] == "Exclusions"
assert result["sources"][0]["relevance_score"] == 0.97
assert result["sources"][1]["clause_type"] == "coverage"
# ====================================================================== #
# TEST 4 — QueryService full flow (mocked)
# ====================================================================== #
class TestQueryServiceFullFlow(unittest.TestCase):
@patch("rag_engine.embeddings.embedding_factory.get_embedder")
@patch("rag_engine.vector_store.store_factory.get_vector_store")
@patch("rag_engine.llm.llm_factory.get_llm")
@patch("sentence_transformers.CrossEncoder")
def test_query_service_full_flow_mocked(
self, mock_ce_cls, mock_get_llm, mock_get_store, mock_get_emb
):
"""QueryService end-to-end with all external calls mocked"""
# embedder
mock_embedder = MagicMock()
mock_embedder.embed_query.return_value = [0.1] * 768
mock_get_emb.return_value = mock_embedder
# vector store
mock_store = MagicMock()
mock_store.similarity_search.return_value = [
{
"content": "Flood is excluded per Section 3.1",
"metadata": {
"section_name": "Exclusions",
"clause_type": "exclusion",
"policy_id": "POL-001",
},
"score": 0.94,
}
]
mock_get_store.return_value = mock_store
# llm
mock_llm = MagicMock()
mock_llm.complete.return_value = "Flood damage is not covered."
mock_get_llm.return_value = mock_llm
# reranker — mock CrossEncoder so it doesn't download a model
mock_ce = MagicMock()
mock_ce.predict.return_value = [0.97]
mock_ce_cls.return_value = mock_ce
from rag_engine.services.query_service import QueryService
service = QueryService()
result = service.query("Is flood covered?", "POL-001")
mock_embedder.embed_query.assert_called_once()
mock_store.similarity_search.assert_called_once()
mock_llm.complete.assert_called_once()
assert result["answer"] == "Flood damage is not covered."
assert result["policy_id"] == "POL-001"
assert result["source_count"] == 1
# ====================================================================== #
# TEST 5 — IngestionService skip existing
# ====================================================================== #
class TestIngestionServiceSkipExisting(unittest.TestCase):
@patch("rag_engine.ingestion.pipeline.IngestionPipeline")
@patch("rag_engine.embeddings.embedding_factory.get_embedder")
@patch("rag_engine.vector_store.store_factory.get_vector_store")
def test_ingestion_service_skip_existing(
self, mock_get_store, mock_get_emb, mock_pipeline_cls
):
"""IngestionService skips if policy already exists"""
mock_store = MagicMock()
mock_store.policy_exists.return_value = True
mock_store.get_policy_chunk_count.return_value = 42
mock_get_store.return_value = mock_store
from rag_engine.services.ingestion_service import IngestionService
service = IngestionService()
result = service.ingest("fake.pdf", "POL-EXISTS", overwrite=False)
assert result["status"] == "skipped"
assert result["chunks"] == 42
assert result["reason"] == "already_exists"
mock_store.policy_exists.assert_called_once_with("POL-EXISTS")
# ====================================================================== #
# TEST 6 — IngestionService overwrite
# ====================================================================== #
class TestIngestionServiceOverwrite(unittest.TestCase):
@patch("rag_engine.ingestion.pipeline.IngestionPipeline")
@patch("rag_engine.embeddings.embedding_factory.get_embedder")
@patch("rag_engine.vector_store.store_factory.get_vector_store")
def test_ingestion_service_overwrite(
self, mock_get_store, mock_get_emb, mock_pipeline_cls
):
"""IngestionService calls delete when overwrite=True"""
mock_store = MagicMock()
mock_store.policy_exists.return_value = True
mock_get_store.return_value = mock_store
mock_embedder = MagicMock()
mock_embedder.embed_documents.return_value = [[0.1] * 768, [0.2] * 768]
mock_get_emb.return_value = mock_embedder
mock_pipeline = MagicMock()
mock_pipeline.run.return_value = [
("chunk text 1", MagicMock(model_dump=lambda: {"policy_id": "POL-001"})),
("chunk text 2", MagicMock(model_dump=lambda: {"policy_id": "POL-001"})),
]
mock_pipeline_cls.return_value = mock_pipeline
from rag_engine.services.ingestion_service import IngestionService
service = IngestionService()
result = service.ingest("fake.pdf", "POL-001", overwrite=True)
mock_store.delete_policy.assert_called_once_with("POL-001")
mock_store.add_chunks.assert_called_once()
assert result["status"] == "success"
assert result["chunks"] == 2
# ====================================================================== #
# TEST 7 — ResponseFormatter highlight fields
# ====================================================================== #
class TestResponseFormatterHighlightFields(unittest.TestCase):
def test_sources_contain_page_number_and_highlight_text(self):
"""Sources must include page_number and highlight_text for PDF highlighting."""
formatter = ResponseFormatter()
mock_chunk = {
"content": "The insured shall pay a deductible of $500.",
"metadata": {
"section_name": "Deductibles",
"clause_type": "deductible",
"policy_id": "POL-TEST",
"page_number": 5,
},
"score": 0.90,
}
result = formatter.format_answer("Test answer", "test question", [mock_chunk])
assert result["sources"][0]["page_number"] == 5
assert result["sources"][0]["highlight_text"] == "The insured shall pay a deductible of $500."
assert result["sources"][0]["snippet"] is not None
# ====================================================================== #
# TEST 8 — DocumentCleaner.extract_page_map
# ====================================================================== #
class TestCleanerExtractPageMap(unittest.TestCase):
def test_cleaner_extract_page_map(self):
"""extract_page_map should parse PAGE_START markers into a char-offset dict."""
from rag_engine.ingestion.cleaner import DocumentCleaner
raw_markdown = (
"---PAGE_START:1---\n"
"some text on page one\n"
"\n"
"---PAGE_START:3---\n"
"more text on page three\n"
)
cleaner = DocumentCleaner()
page_map = cleaner.extract_page_map(raw_markdown)
assert len(page_map) > 0
page_values = set(page_map.values())
assert 1 in page_values
assert 3 in page_values
if __name__ == "__main__":
unittest.main()
|