opencode-skills-collection 4.0.69 → 4.0.70

package/bundled-skills/fact-check-x-unified/tests/multi_platform_test.py

@@ -0,0 +1,369 @@
+#!/usr/bin/env python3
+from __future__ import annotations
+
+import json
+import os
+import subprocess
+import sys
+import tempfile
+from pathlib import Path
+
+
+ROOT = Path(__file__).resolve().parents[1]
+SKILLS = ROOT.parent
+if (ROOT / "modules").exists():
+    SKILLS = ROOT / "modules"
+COMPARE = SKILLS / "fact-check-x-knowledge-compare" / "tests" / "fixtures"
+AUTHORITY = SKILLS / "fact-check-x-authoritative-verify" / "tests" / "fixtures"
+
+
+def command(*arguments: str) -> list[str]:
+    return [sys.executable, str(ROOT / "scripts" / "fact_check_x.py"), *arguments]
+
+
+def run(arguments: list[str], environment: dict[str, str] | None = None) -> dict:
+    process = subprocess.run(
+        arguments,
+        text=True,
+        capture_output=True,
+        check=False,
+        env=environment,
+    )
+    if process.returncode:
+        raise AssertionError(process.stdout or process.stderr)
+    payload = json.loads([line for line in process.stdout.splitlines() if line.strip()][-1])
+    checkpoint = payload.get("checkpoint") or {}
+    acknowledgement = checkpoint.get("acknowledgement") or {}
+    if checkpoint.get("status") == "awaiting_user" and acknowledgement:
+        acknowledged = subprocess.run(
+            [sys.executable, str(ROOT / "scripts" / "fact_check_x.py"),
+             "acknowledge-stage", "--run-dir", str(Path(checkpoint["path"]).parent),
+             "--stage", checkpoint["stage"], "--token", acknowledgement["token"],
+             "--decision", "continue"],
+            text=True, capture_output=True, check=False, env=environment,
+        )
+        if acknowledged.returncode:
+            raise AssertionError(acknowledged.stdout or acknowledged.stderr)
+    return payload
+
+
+def run_failed(arguments: list[str]) -> dict:
+    process = subprocess.run(
+        arguments,
+        text=True,
+        capture_output=True,
+        check=False,
+    )
+    if process.returncode == 0:
+        raise AssertionError("命令应以非零状态拒绝无平台输入")
+    return json.loads([line for line in process.stdout.splitlines() if line.strip()][-1])
+
+
+def write_json(path: Path, value: dict) -> None:
+    path.write_text(json.dumps(value, ensure_ascii=False), encoding="utf-8")
+
+
+def complete_pipeline(
+    root: Path,
+    case_name: str,
+    results: dict,
+    analysis: dict,
+    assessment: dict,
+) -> Path:
+    run_dir = root / f"{case_name}-run"
+    results_path = root / f"{case_name}-results.json"
+    analysis_path = root / f"{case_name}-analysis.json"
+    write_json(results_path, results)
+    write_json(analysis_path, analysis)
+
+    prepared = run(
+        command(
+            "prepare-comparison",
+            "--results",
+            str(results_path),
+            "--run-dir",
+            str(run_dir),
+        )
+    )
+    assert len(prepared["platforms"]) == len(results["platforms"])
+    completed = run(
+        command(
+            "complete-comparison",
+            "--results",
+            str(results_path),
+            "--analysis",
+            str(analysis_path),
+            "--run-dir",
+            str(run_dir),
+        )
+    )
+    assert completed["stage"] == "comparison_completed"
+    assert completed["artifacts"]["comparisonAnalysis"] == str(
+        (run_dir / "comparison-analysis.json").resolve()
+    )
+    assert (run_dir / "comparison-analysis.json").is_file()
+
+    keyless = {
+        **os.environ,
+        "TRUSTED_SEARCH_KEY": "",
+        "FACT_CHECK_X_TRUSTED_SEARCH_KEY_FILE": str(
+            run_dir / "missing-trusted-search-key"
+        ),
+    }
+    run(command("prepare-authority", "--run-dir", str(run_dir)), keyless)
+    run(
+        command(
+            "search-authority",
+            "--run-dir",
+            str(run_dir),
+            "--max-workers",
+            "12",
+        ),
+        keyless,
+    )
+    assessments = run_dir / "authority" / "assessments"
+    assessments.mkdir(parents=True)
+    write_json(assessments / "K1.json", assessment)
+    run(command("finalize-authority", "--run-dir", str(run_dir)), keyless)
+    delivered = run(
+        command(
+            "deliver",
+            "--results",
+            str(results_path),
+            "--run-dir",
+            str(run_dir),
+        ),
+        keyless,
+    )
+    assert delivered["status"] == "completed"
+    pipeline = json.loads((run_dir / "pipeline.json").read_text(encoding="utf-8"))
+    assert pipeline["artifacts"]["comparisonAnalysis"] == str(
+        (run_dir / "comparison-analysis.json").resolve()
+    )
+    assert all(
+        (run_dir / name).is_file()
+        for name in (
+            "01-capture-report.html",
+            "02-comparison-report.html",
+            "03-authority-report.html",
+            "04-final-report.html",
+            "05-complete-report-package.zip",
+        )
+    )
+    comparison_deliverable = (run_dir / "02-comparison-report.html").read_text(
+        encoding="utf-8"
+    )
+    assert all(
+        f'href="{target}"' in comparison_deliverable
+        for target in (
+            "01-capture-report.html",
+            "03-authority-report.html",
+            "04-final-report.html",
+        )
+    )
+    return run_dir
+
+
+def main() -> int:
+    with tempfile.TemporaryDirectory(prefix="fact-check-x-dynamic-platform-") as temp:
+        root = Path(temp)
+        results_fixture = json.loads(
+            (COMPARE / "results.json").read_text(encoding="utf-8")
+        )
+        analysis_fixture = json.loads(
+            (COMPARE / "comparison-analysis.json").read_text(encoding="utf-8")
+        )
+        assessment_fixture = json.loads(
+            (AUTHORITY / "K1-assessment.json").read_text(encoding="utf-8")
+        )
+
+        empty_results = json.loads(
+            json.dumps(results_fixture, ensure_ascii=False)
+        )
+        empty_results["platforms"] = []
+        empty_results_path = root / "empty-results.json"
+        write_json(empty_results_path, empty_results)
+        rejected = run_failed(
+            command(
+                "prepare-comparison",
+                "--results",
+                str(empty_results_path),
+                "--run-dir",
+                str(root / "empty-run"),
+            )
+        )
+        assert "至少需要 1 个采集平台" in rejected["error"]
+
+        single_results = json.loads(
+            json.dumps(results_fixture, ensure_ascii=False)
+        )
+        single_results["platforms"] = single_results["platforms"][:1]
+        single_analysis = json.loads(
+            json.dumps(analysis_fixture, ensure_ascii=False)
+        )
+        single_point = single_analysis["knowledgePoints"][0]
+        single_point["claims"] = {
+            "dknowc-chat": single_point["claims"]["dknowc-chat"]
+        }
+        single_point["comparison"] = {
+            "status": "single",
+            "summary": "深知晓单平台知识点结构化结果",
+        }
+        single_analysis["synthesisDraft"] = {
+            "status": "unverified",
+            "answer": "深知晓给出的综合草案为每人每月最高提取1400元，尚未经过后续权威核验。",
+            "basisKnowledgePointIds": ["K1"],
+        }
+        single_assessment = json.loads(
+            json.dumps(assessment_fixture, ensure_ascii=False)
+        )
+        single_assessment["verdicts"] = {
+            "dknowc-chat": single_assessment["verdicts"]["dknowc-chat"]
+        }
+        single_run = complete_pipeline(
+            root,
+            "single-platform",
+            single_results,
+            single_analysis,
+            single_assessment,
+        )
+        single_comparison = json.loads(
+            (single_run / "comparison.json").read_text(encoding="utf-8")
+        )
+        assert [item["platform"] for item in single_comparison["platforms"]] == [
+            "dknowc-chat"
+        ]
+        assert (
+            single_comparison["knowledgePoints"][0]["comparison"]["status"]
+            == "single"
+        )
+        single_verification = json.loads(
+            (single_run / "verification.json").read_text(encoding="utf-8")
+        )
+        assert [item["platform"] for item in single_verification["platforms"]] == [
+            "dknowc-chat"
+        ]
+        assert set(
+            single_verification["knowledgePoints"][0]["authority"]["verdicts"]
+        ) == {"dknowc-chat"}
+        for report_name in ("02-comparison-report.html", "04-final-report.html"):
+            report = (single_run / report_name).read_text(encoding="utf-8")
+            assert "深知晓" in report
+            assert "豆包" not in report
+        authority_report = (single_run / "03-authority-report.html").read_text(
+            encoding="utf-8"
+        )
+        assert "权威核验后的最终答案" in authority_report
+        assert "深知晓" not in authority_report and "豆包" not in authority_report
+        single_comparison_html = (
+            single_run / "02-comparison-report.html"
+        ).read_text(encoding="utf-8")
+        assert single_comparison_html.count('class="answer-panel"') == 1
+        assert "各方答案聚合（问题：" in single_comparison_html
+        assert "逐知识点结构化结果" in single_comparison_html
+        assert "<th>差异</th>" not in single_comparison_html
+        assert "跨平台" not in single_comparison_html
+
+        results = json.loads(json.dumps(results_fixture, ensure_ascii=False))
+        deepseek = json.loads(json.dumps(results["platforms"][1], ensure_ascii=False))
+        deepseek.update(
+            {
+                "platform": "deepseek",
+                "label": "DeepSeek",
+                "url": "https://chat.deepseek.com/",
+                "answerMarkdown": "每人每月最高提取 2000 元。",
+            }
+        )
+        results["platforms"].append(deepseek)
+
+        analysis = json.loads(
+            json.dumps(analysis_fixture, ensure_ascii=False)
+        )
+        analysis["knowledgePoints"][0]["claims"]["deepseek"] = json.loads(
+            json.dumps(
+                analysis["knowledgePoints"][0]["claims"]["doubao"],
+                ensure_ascii=False,
+            )
+        )
+        assessment = json.loads(
+            json.dumps(assessment_fixture, ensure_ascii=False)
+        )
+        assessment["verdicts"]["deepseek"] = json.loads(
+            json.dumps(assessment["verdicts"]["doubao"], ensure_ascii=False)
+        )
+        run_dir = complete_pipeline(
+            root,
+            "three-platform",
+            results,
+            analysis,
+            assessment,
+        )
+        comparison = json.loads(
+            (run_dir / "comparison.json").read_text(encoding="utf-8")
+        )
+        assert [item["platform"] for item in comparison["platforms"]] == [
+            "dknowc-chat",
+            "doubao",
+            "deepseek",
+        ]
+        comparison_html = (run_dir / "comparison.html").read_text(encoding="utf-8")
+        assert comparison_html.count('class="answer-panel"') == 3
+        assert "DeepSeek" in comparison_html
+
+        verification = json.loads(
+            (run_dir / "verification.json").read_text(encoding="utf-8")
+        )
+        assert [item["platform"] for item in verification["platforms"]] == [
+            "dknowc-chat",
+            "doubao",
+            "deepseek",
+        ]
+        final_html = (run_dir / "report.html").read_text(encoding="utf-8")
+        authority_html = (run_dir / "03-authority-report.html").read_text(
+            encoding="utf-8"
+        )
+        capture_html = (run_dir / "01-capture-report.html").read_text(
+            encoding="utf-8"
+        )
+        assert "DeepSeek" in final_html
+        assert "权威核验后的最终答案" in authority_html
+        assert "DeepSeek" not in authority_html
+        assert (
+            'class="shell platform-layout-compact platform-count-3" '
+            'style="--platform-count:3"'
+        ) in capture_html
+        assert "repeat(var(--platform-count), minmax(0, 1fr))" in capture_html
+        assert "repeat(3, 500px)" not in capture_html
+        assert (
+            '<main class="multi-platform platform-layout-compact '
+            'platform-count-3" style="--platform-count:3">'
+        ) in comparison_html
+        assert (
+            '<body class="platform-layout-compact platform-count-3" '
+            'style="--platform-count:3;'
+        ) in final_html
+        assert 'class="locked-platforms"' in final_html
+        assert ".locked-platforms{display:grid" in final_html
+        assert "② 逐知识点核验明细" in final_html
+        assert final_html.count("② 逐知识点核验明细") == 1
+        assert 'data-fcx-locked-role="direct"' in final_html
+        assert 'data-fcx-locked-role="reference"' in final_html
+        assert "② 直接答案逐知识点测评" not in final_html
+        assert "③ 补充参考逐知识点测评" not in final_html
+
+    if os.getenv("FACT_CHECK_X_ASSERTIONS_OUTPUT"):
+        Path(os.environ["FACT_CHECK_X_ASSERTIONS_OUTPUT"]).write_text(json.dumps({
+            "schemaVersion": "fact-check-x/test-assertions@1",
+            "actualAssertionIds": [
+                "platform.minimum_one",
+                "platform.single_full_pipeline",
+                "platform.three_full_pipeline",
+                "platform.dynamic_layout",
+            ],
+        }), encoding="utf-8")
+    print("PASS Fact-Check-X N≥1 动态单/多平台完整流水线")
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())