Travis Vasceannie
8ad47a7640
* Modernize research graph metadata for LangGraph v1 * Update src/biz_bud/core/langgraph/graph_builder.py Co-authored-by: qodo-merge-pro[bot] <151058649+qodo-merge-pro[bot]@users.noreply.github.com> --------- Co-authored-by: qodo-merge-pro[bot] <151058649+qodo-merge-pro[bot]@users.noreply.github.com>
154 lines
6.0 KiB
Python
154 lines
6.0 KiB
Python
"""Lightweight tests covering the LangGraph v1 RAG workflow configuration."""
|
|
|
|
from __future__ import annotations
|
|
|
|
import sys
|
|
from pathlib import Path
|
|
from types import ModuleType
|
|
|
|
import pytest
|
|
from langgraph.graph.state import CachePolicy, RetryPolicy
|
|
|
|
# ---------------------------------------------------------------------------
|
|
# Optional dependency shims
|
|
# ---------------------------------------------------------------------------
|
|
# The RAG graph pulls in a wide range of optional dependencies through the
|
|
# package ``biz_bud.graphs``. We only need to import the graph module itself
|
|
# to introspect configuration, so provide lightweight stubs that satisfy the
|
|
# imports without requiring the full production stack.
|
|
|
|
project_root = Path(__file__).resolve().parents[3]
|
|
|
|
if "biz_bud" not in sys.modules: # pragma: no cover - package bootstrap
|
|
biz_bud_pkg = ModuleType("biz_bud")
|
|
biz_bud_pkg.__path__ = [str(project_root / "src" / "biz_bud")]
|
|
sys.modules["biz_bud"] = biz_bud_pkg
|
|
|
|
if "biz_bud.graphs" not in sys.modules: # pragma: no cover - avoid heavy __init__
|
|
graphs_pkg = ModuleType("biz_bud.graphs")
|
|
graphs_pkg.__path__ = [str(project_root / "src" / "biz_bud" / "graphs")]
|
|
sys.modules["biz_bud.graphs"] = graphs_pkg
|
|
|
|
|
|
if "biz_bud.graphs.rag" not in sys.modules: # pragma: no cover - package shim
|
|
from importlib.machinery import ModuleSpec
|
|
|
|
rag_pkg = ModuleType("biz_bud.graphs.rag")
|
|
rag_path = str(project_root / "src" / "biz_bud" / "graphs" / "rag")
|
|
rag_pkg.__path__ = [rag_path]
|
|
rag_pkg.__spec__ = ModuleSpec("biz_bud.graphs.rag", loader=None, is_package=True)
|
|
rag_pkg.__spec__.submodule_search_locations = [rag_path]
|
|
sys.modules["biz_bud.graphs.rag"] = rag_pkg
|
|
|
|
|
|
def _passthrough(state, *_args, **_kwargs): # pragma: no cover - helper stub
|
|
return state
|
|
|
|
|
|
if "biz_bud.graphs.rag.nodes" not in sys.modules: # pragma: no cover - node stubs
|
|
rag_nodes = ModuleType("biz_bud.graphs.rag.nodes")
|
|
rag_nodes.analyze_content_for_rag_node = _passthrough
|
|
rag_nodes.check_existing_content_node = _passthrough
|
|
rag_nodes.check_r2r_duplicate_node = _passthrough
|
|
rag_nodes.decide_processing_node = _passthrough
|
|
rag_nodes.determine_processing_params_node = _passthrough
|
|
rag_nodes.upload_to_r2r_node = _passthrough
|
|
sys.modules[rag_nodes.__name__] = rag_nodes
|
|
|
|
if "biz_bud.graphs.rag.nodes.integrations" not in sys.modules: # pragma: no cover
|
|
integrations_module = ModuleType("biz_bud.graphs.rag.nodes.integrations")
|
|
integrations_module.repomix_process_node = _passthrough
|
|
sys.modules[integrations_module.__name__] = integrations_module
|
|
|
|
if "biz_bud.graphs.rag.nodes.scraping" not in sys.modules: # pragma: no cover
|
|
scraping_module = ModuleType("biz_bud.graphs.rag.nodes.scraping")
|
|
scraping_module.batch_process_urls_node = _passthrough
|
|
scraping_module.discover_urls_node = _passthrough
|
|
scraping_module.route_url_node = _passthrough
|
|
scraping_module.scrape_status_summary_node = _passthrough
|
|
sys.modules[scraping_module.__name__] = scraping_module
|
|
|
|
if "biz_bud.nodes" not in sys.modules: # pragma: no cover - minimal node facade
|
|
nodes_module = ModuleType("biz_bud.nodes")
|
|
nodes_module.finalize_status_node = _passthrough
|
|
nodes_module.preserve_url_fields_node = _passthrough
|
|
sys.modules[nodes_module.__name__] = nodes_module
|
|
|
|
|
|
if "docling" not in sys.modules: # pragma: no cover - optional dependency shim
|
|
docling_module = ModuleType("docling")
|
|
converter_module = ModuleType("docling.document_converter")
|
|
|
|
class _StubDocumentConverter: # pragma: no cover - minimal convert stub
|
|
async def convert(self, *_args, **_kwargs):
|
|
return None
|
|
|
|
converter_module.DocumentConverter = _StubDocumentConverter
|
|
docling_module.document_converter = converter_module
|
|
sys.modules["docling"] = docling_module
|
|
sys.modules["docling.document_converter"] = converter_module
|
|
|
|
|
|
graph_module_name = "biz_bud.graphs.rag.graph"
|
|
graph_module_path = project_root / "src" / "biz_bud" / "graphs" / "rag" / "graph.py"
|
|
|
|
if graph_module_name not in sys.modules: # pragma: no cover - manual module load
|
|
import importlib.util
|
|
|
|
spec = importlib.util.spec_from_file_location(graph_module_name, graph_module_path)
|
|
if spec and spec.loader:
|
|
module = importlib.util.module_from_spec(spec)
|
|
sys.modules[graph_module_name] = module
|
|
spec.loader.exec_module(module)
|
|
|
|
|
|
from biz_bud.graphs.rag.graph import ( # noqa: E402 - imported after shims
|
|
URLToRAGGraphContext,
|
|
URLToRAGGraphInput,
|
|
URLToRAGGraphOutput,
|
|
create_url_to_r2r_graph,
|
|
)
|
|
|
|
|
|
@pytest.fixture(scope="module")
|
|
def compiled_rag_graph():
|
|
"""Compile the RAG graph once for metadata inspection."""
|
|
|
|
return create_url_to_r2r_graph()
|
|
|
|
|
|
def test_rag_graph_declares_schemas(compiled_rag_graph):
|
|
"""The graph should expose LangGraph v1 schemas for type safety."""
|
|
|
|
assert compiled_rag_graph.context_schema is URLToRAGGraphContext
|
|
assert compiled_rag_graph.InputType is URLToRAGGraphInput
|
|
if compiled_rag_graph.output_schema is not None:
|
|
assert compiled_rag_graph.output_schema.__name__ == "url_to_r2r_graph_output"
|
|
|
|
|
|
def test_rag_graph_metadata_and_entry_point(compiled_rag_graph):
|
|
"""Verify graph-level metadata aligns with the registry definition."""
|
|
|
|
assert compiled_rag_graph.name == "url_to_r2r_graph"
|
|
builder = compiled_rag_graph.builder
|
|
assert ("__start__", "route_url") in builder.edges
|
|
|
|
|
|
def test_rag_nodes_use_langgraph_policies(compiled_rag_graph):
|
|
"""Key nodes should leverage retry, cache, and deferral policies."""
|
|
|
|
builder = compiled_rag_graph.builder
|
|
route_node = builder.nodes["route_url"]
|
|
assert isinstance(route_node.metadata, dict)
|
|
assert route_node.metadata["category"] == "routing"
|
|
assert isinstance(route_node.cache_policy, CachePolicy)
|
|
|
|
scrape_node = builder.nodes["scrape_url"]
|
|
assert isinstance(scrape_node.retry_policy, RetryPolicy)
|
|
assert scrape_node.defer is True
|
|
|
|
upload_node = builder.nodes["r2r_upload"]
|
|
assert isinstance(upload_node.retry_policy, RetryPolicy)
|
|
assert upload_node.defer is True
|
|
|