Add unit tests for compendium and research modules; remove deprecated test file

Author: btfranklin

tests/compendium/__init__.py

@@ -1,4 +1,5 @@
-import json
+from __future__ import annotations
+
 import xml.etree.ElementTree as ET
 
 from compendiumscribe.compendium import (
@@ -9,10 +10,6 @@
     Section,
     etree_to_string,
 )
-from compendiumscribe.research import (
-    ResearchConfig,
-    build_compendium,
-)
 
 
 def test_compendium_to_xml_contains_expected_structure():
@@ -33,13 +30,12 @@ def test_compendium_to_xml_contains_expected_structure():
                     Insight(
                         title="Coherence is fragile",
                         evidence=(
-                            "Most systems retain coherence "
-                            "for microseconds before error "
-                            "correction overwhelms throughput."
+                            "Most systems retain coherence for microseconds"
+                            " before error correction overwhelms throughput."
                         ),
                         implications=(
-                            "Large-scale machines require aggressive "
-                            "error mitigation."
+                            "Large-scale machines require aggressive error"
+                            " mitigation."
                         ),
                         citation_refs=["C1"],
                     )
@@ -54,8 +50,8 @@ def test_compendium_to_xml_contains_expected_structure():
                 publisher="ACM",
                 published_at="2023-04-01",
                 summary=(
-                    "Overview of leading quantum error correction "
-                    "strategies."
+                    "Overview of leading quantum error correction"
+                    " strategies."
                 ),
             )
         ],
@@ -196,12 +192,11 @@ def test_compendium_from_payload_normalizes_fields():
                     {
                         "title": "Private investment surged",
                         "evidence": (
-                            "Funding grew 45% year over year "
-                            "according to PitchBook."
+                            "Funding grew 45% year over year according to"
+                            " PitchBook."
                         ),
                         "implications": (
-                            "Competition for talent is "
-                            "increasing."
+                            "Competition for talent is increasing."
                         ),
                         "citations": ["C1"],
                     }
@@ -215,10 +210,7 @@ def test_compendium_from_payload_normalizes_fields():
                 "url": "https://example.com/pitchbook",
                 "publisher": "PitchBook",
                 "published_at": "2024-01-15",
-                "summary": (
-                    "Investment trends across quantum "
-                    "startups."
-                ),
+                "summary": "Investment trends across quantum startups.",
             }
         ],
         "open_questions": ["How will regulation shape deployment?"],
@@ -244,213 +236,6 @@ def test_compendium_from_payload_normalizes_fields():
     assert compendium.trace[0].event_id == "ws_1"
 
 
-class FakeResponse:
-    def __init__(
-        self,
-        *,
-        output_text=None,
-        output=None,
-        status="completed",
-        response_id="resp_1",
-    ):
-        self.output_text = output_text
-        self.output = output or []
-        self.status = status
-        self.id = response_id
-
-
-class FakeResponsesAPI:
-    def __init__(self, plan_json: str, research_json: str):
-        self.plan_json = plan_json
-        self.research_json = research_json
-        self.calls: list[dict[str, str]] = []
-
-    def create(self, **kwargs):
-        model = kwargs.get("model")
-        self.calls.append(
-            {"model": model, "input": kwargs.get("input", "")}
-        )
-
-        if model in {"gpt-4.1", "gpt-4.1-mini"}:
-            return FakeResponse(
-                output_text=self.plan_json,
-                response_id="plan_1",
-            )
-
-        if model == "o3-deep-research":
-            output = [
-                {
-                    "type": "web_search_call",
-                    "id": "ws_1",
-                    "status": "completed",
-                    "action": {
-                        "type": "search",
-                        "query": "quantum computing breakthroughs",
-                    },
-                },
-                {
-                    "type": "message",
-                    "content": [
-                        {
-                            "type": "output_text",
-                            "text": self.research_json,
-                        }
-                    ],
-                },
-            ]
-            return FakeResponse(
-                output=output,
-                status="completed",
-                response_id="research_1",
-            )
-
-        raise AssertionError(f"Unexpected model request: {model}")
-
-    def retrieve(self, response_id: str):
-        # pragma: no cover - not exercised in this test
-        raise AssertionError(
-            f"retrieve called unexpectedly for {response_id}"
-        )
-
-
-class FakeOpenAI:
-    def __init__(self, plan_json: str, research_json: str):
-        self.responses = FakeResponsesAPI(plan_json, research_json)
-
-
-def test_build_compendium_with_stub_client():
-    plan = {
-        "primary_objective": "Build a comprehensive compendium",
-        "audience": "Strategic leadership teams",
-        "key_sections": [
-            {"title": "Context", "focus": "Historical milestones"},
-            {"title": "Applications", "focus": "Practical deployments"},
-        ],
-        "research_questions": [
-            "What breakthroughs unlocked current capabilities?",
-            "Who are the leading vendors?",
-        ],
-        "methodology_preferences": [
-            "Verify each statistic using at least two sources",
-            "Prioritize materials from 2022 onward",
-        ],
-    }
-
-    research_payload = {
-        "topic_overview": (
-            "Quantum computing is transitioning from lab prototypes to early "
-            "commercial pilots."
-        ),
-        "methodology": [
-            "Surveyed public filings and analyst coverage",
-            "Aggregated investment data across multiple trackers",
-        ],
-        "sections": [
-            {
-                "id": "S1",
-                "title": "Technological Foundations",
-                "summary": (
-                    "Hardware approaches and error correction challenges"
-                ),
-                "key_terms": ["superconducting qubits"],
-                "guiding_questions": [
-                    "Which modalities show the most promise?"
-                ],
-                "insights": [
-                    {
-                        "title": (
-                            "Superconducting qubits dominate near-term "
-                            "roadmaps"
-                        ),
-                        "evidence": (
-                            "IBM and Google published roadmaps targeting "
-                            ">1000 qubits with heavy error mitigation by 2025."
-                        ),
-                        "implications": (
-                            "Vendor lock-in may increase as proprietary "
-                            "control stacks mature."
-                        ),
-                        "citations": ["C1", "C2"],
-                    }
-                ],
-            }
-        ],
-        "citations": [
-            {
-                "id": "C1",
-                "title": "IBM Quantum Roadmap",
-                "url": "https://example.com/ibm-roadmap",
-                "publisher": "IBM",
-                "published_at": "2023-12-01",
-                "summary": "Targets for qubit scaling and error mitigation.",
-            },
-            {
-                "id": "C2",
-                "title": "Google Quantum AI Progress Update",
-                "url": "https://example.com/google-qa",
-                "publisher": "Google",
-                "published_at": "2024-02-10",
-                "summary": "Highlights on achieving reduced error rates.",
-            },
-        ],
-        "open_questions": [
-            "How will supply chains support dilution refrigerators at scale?"
-        ],
-    }
-
-    client = FakeOpenAI(json.dumps(plan), json.dumps(research_payload))
-    config = ResearchConfig(background=False, stream_progress=False)
-
-    compendium = build_compendium(
-        "Quantum Computing",
-        client=client,
-        config=config,
-    )
-
-    assert compendium.overview.startswith(
-        "Quantum computing is transitioning"
-    )
-    assert compendium.sections[0].insights[0].citation_refs == ["C1", "C2"]
-    assert (
-        compendium.citations[1].title
-        == "Google Quantum AI Progress Update"
-    )
-    assert compendium.trace[0].event_type == "web_search_call"
-    assert len(client.responses.calls) == 2
-    assert "Quantum Computing" in client.responses.calls[1]["input"]
-
-
-def test_build_compendium_emits_progress_updates():
-    plan = {"primary_objective": "Capture topic"}
-    research_payload = {
-        "topic_overview": "Overview",
-        "methodology": [],
-        "sections": [],
-        "citations": [],
-        "open_questions": [],
-    }
-
-    client = FakeOpenAI(json.dumps(plan), json.dumps(research_payload))
-    captured: list = []
-
-    def capture_progress(update):
-        captured.append(update)
-
-    config = ResearchConfig(
-        background=False,
-        progress_callback=capture_progress,
-        stream_progress=False,
-    )
-
-    build_compendium("Test Topic", client=client, config=config)
-
-    assert captured, "Expected progress callback to receive updates"
-    phases = {event.phase for event in captured}
-    assert "planning" in phases
-    assert "deep_research" in phases
-    assert any(event.status == "completed" for event in captured)
-
-
 def test_compendium_from_payload_generates_event_id_when_missing():
     payload = {
         "topic_overview": "Overview",

tests/test_xml_conversion.py tests/compendium/test_compendium.pytests/test_xml_conversion.py renamed to tests/compendium/test_compendium.py

@@ -0,0 +1,11 @@
+from __future__ import annotations
+
+from compendiumscribe.research.config import ResearchConfig
+
+
+def test_research_config_uses_env_override(monkeypatch):
+    monkeypatch.setenv("RESEARCH_MODEL", "custom-deep-model")
+
+    config = ResearchConfig()
+
+    assert config.deep_research_model == "custom-deep-model"