Refactor AI daily report pipeline

ai_daily_report/validate.py

@@ -0,0 +1,46 @@
+from __future__ import annotations
+
+import re
+from typing import Any
+
+from .classify import SECTION_ORDER
+from .models import NewsItem
+
+
+def validate_report_markdown(markdown: str, items: list[NewsItem]) -> dict[str, Any]:
+    return validate_markdown(markdown, items)
+
+
+def validate_markdown(markdown: str, items: list[NewsItem]) -> dict[str, Any]:
+    blocking_errors: list[str] = []
+    auto_fixes: list[str] = []
+    warnings: list[dict[str, str]] = []
+
+    if not items:
+        blocking_errors.append("no_items")
+    if len((markdown or "").strip()) < 80:
+        blocking_errors.append("markdown_too_short")
+    if items and "## " not in markdown:
+        blocking_errors.append("no_sections")
+    if re.search(r"\{[^{}]*\}", markdown or ""):
+        blocking_errors.append("json_fragment_detected")
+    if "> >" in (markdown or ""):
+        auto_fixes.append("double_blockquote_detected")
+    if re.search(r"\[\d+\]|\[N\]", markdown or ""):
+        auto_fixes.append("reference_marker_detected")
+
+    for item in items:
+        if not item.url:
+            warnings.append({"type": "missing_url", "item_id": item.id})
+        if item.section not in SECTION_ORDER:
+            blocking_errors.append("invalid_section")
+            break
+
+    return {
+        "item_count": len(items),
+        "section_count": len({item.section for item in items if item.section}),
+        "markdown_length": len(markdown or ""),
+        "auto_fixes": auto_fixes,
+        "warnings": warnings,
+        "blocking_errors": blocking_errors,
+    }