refs #732: separate report data and rendering requests

ai-web-agent-console/app.py

@@ -2894,6 +2894,104 @@ def _clean_presentation_title(value: Any) -> str:
     return text or "업무 현황"
 
 
+def _fallback_presentation_data_query(question: str) -> str:
+    text = " ".join(str(question or "").split()).strip()
+    text = re.sub(
+        r"\s*(?:그리고\s*)?(?:이걸|이를|그\s*결과를|결과를)\s*"
+        r"(?:HMM\s*)?(?:HTML|리포트|보고서|대시보드|차트)\s*"
+        r"(?:형식)?(?:으로|로)?\s*(?:만들어|생성해|작성해|보여)\s*"
+        r"(?:줘|주세요)?\s*[.!?]?\s*$",
+        "",
+        text,
+        flags=re.IGNORECASE,
+    )
+    text = re.sub(
+        r"(?i)HTML\s*(?:형식)?(?:으로|로)?\s*",
+        "",
+        text,
+    )
+    text = re.sub(
+        r"(?:HMM\s*)?(?:리포트|보고서|대시보드|차트)\s*"
+        r"(?:형식)?(?:으로|로)?\s*(?:만들어|생성해|작성해)\s*"
+        r"(?:줘|주세요)?",
+        "보여줘",
+        text,
+        flags=re.IGNORECASE,
+    )
+    return " ".join(text.split()).strip() or str(question or "").strip()
+
+
+def _plan_presentation_data_query(
+    *,
+    question: str,
+    route: RoutedMcpTool,
+    model_profile_key: str,
+) -> str:
+    """Separate the data request from its presentation-format request."""
+
+    fallback = _fallback_presentation_data_query(question)
+    try:
+        profile = resolve_model_profile(model_profile_key)
+        client = build_oci_genai_completion_client(
+            profile.model_id,
+            profile.answer_model_region,
+            profile.answer_model_endpoint,
+        )
+        text = client.complete(
+            system_prompt=(
+                "Rewrite the Korean request as a data-retrieval question for the selected "
+                "MCP tool. Preserve every business subject, identifier, filter, period, and "
+                "requested metric. Remove only presentation-format and artifact-generation "
+                "instructions such as HTML, report, dashboard, chart, render, 보여주는 형식, "
+                "or 만들어줘. Do not answer the question, write SQL, or request a token. "
+                "The rewritten query must ask for ordinary structured data rows, never HTML "
+                "tags, Markdown, or preformatted table rows. Return only schema JSON."
+            ),
+            user_prompt=json.dumps(
+                {
+                    "question": question,
+                    "selected_tool": route.tool.name,
+                    "tool_description": route.tool.description[:1200],
+                },
+                ensure_ascii=False,
+            ),
+            response_schema={
+                "type": "object",
+                "additionalProperties": False,
+                "required": ["data_query"],
+                "properties": {"data_query": {"type": "string"}},
+            },
+            max_tokens=300,
+            temperature=temperature_for_model_profile(profile),
+        )
+        parsed = json.loads(text)
+        if isinstance(parsed, Mapping):
+            planned = _clean_agent_tool_query(parsed.get("data_query"), fallback)
+            return _fallback_presentation_data_query(planned or fallback)
+    except Exception:
+        pass
+    return fallback
+
+
+def _mcp_rows_contain_presentation_markup(value: Any) -> bool:
+    rows = _mcp_structured_rows(value)
+    if not rows:
+        return False
+    for row in rows:
+        if not isinstance(row, Mapping):
+            continue
+        keys = {str(key).casefold() for key in row}
+        if keys & {"htmlrow", "html", "markdown"}:
+            return True
+        if any(
+            isinstance(item, str)
+            and re.search(r"<(?:tr|td|table|div)\b", item, flags=re.IGNORECASE)
+            for item in row.values()
+        ):
+            return True
+    return False
+
+
 def _plan_presentation_title(
     *,
     question: str,
@@ -3995,6 +4093,12 @@ def run_mcp_agent_loop(
 
         started = perf_counter()
         presentation_title = ""
+        if wants_presentation and not observations and not _is_presentation_route(route):
+            tool_query = _plan_presentation_data_query(
+                question=question,
+                route=route,
+                model_profile_key=model_profile_key,
+            )
         if _is_presentation_route(route):
             source_step = next(
                 (
@@ -4025,6 +4129,14 @@ def run_mcp_agent_loop(
         )
         parsed = _content_text_json(raw_result)
         mcp_result = parsed if parsed is not None else raw_result
+        if (
+            wants_presentation
+            and not _is_presentation_route(route)
+            and _mcp_rows_contain_presentation_markup(mcp_result)
+        ):
+            raise PublicMcpError(
+                "데이터 조회 결과가 구조화 행이 아니라 표현용 마크업으로 반환됐습니다."
+            )
         elapsed = perf_counter() - started
         summary = _mcp_summary(mcp_result)
         step = {

ai-web-agent-console/tests/test_scenarios.py

@@ -211,6 +211,30 @@ class DemoScenarioConfigTest(unittest.TestCase):
             source,
         )
 
+    def test_hmm_report_data_query_drops_html_format_request(self) -> None:
+        source = (Path(__file__).parents[1] / "app.py").read_text(encoding="utf-8")
+        tree = ast.parse(source)
+        helper = next(
+            node
+            for node in tree.body
+            if isinstance(node, ast.FunctionDef)
+            and node.name == "_fallback_presentation_data_query"
+        )
+        namespace: dict[str, Any] = {"re": re}
+        exec(compile(ast.Module(body=[helper], type_ignores=[]), "app.py", "exec"), namespace)
+
+        query = namespace["_fallback_presentation_data_query"](
+            "E1001 팀장의 담당 선사 최신 매출, 매출총이익, 정시 운항률, "
+            "위험 등급을 HTML로 보여줘"
+        )
+
+        self.assertNotIn("HTML", query.upper())
+        self.assertIn("E1001", query)
+        self.assertIn("매출총이익", query)
+        self.assertIn("정시 운항률", query)
+        self.assertIn("위험 등급", query)
+        self.assertIn("_mcp_rows_contain_presentation_markup(mcp_result)", source)
+
     def test_hmm_report_template_contains_only_dynamic_payload_slot(self) -> None:
         template = (
             Path(__file__).parents[1]