Refactor AI daily report pipeline

ai_daily_report/assemble.py

@@ -0,0 +1,77 @@
+from __future__ import annotations
+
+import re
+from typing import Any
+
+from .classify import SECTION_ORDER
+from .models import NewsItem
+from .validate import validate_markdown
+
+
+END_PUNCTUATION = "。！？；.!?;"
+
+
+def _clean_text(text: str) -> str:
+    value = re.sub(r"^```(?:\w+)?\s*\n?", "", (text or "").strip())
+    value = re.sub(r"\n?```\s*$", "", value)
+    value = re.sub(r"^\s*>\s*", "", value)
+    value = re.sub(r"\[\d+\]|\[N\]", "", value)
+    value = re.sub(r"主线判断[：:]\s*", "", value)
+    value = re.sub(r"\s+", " ", value).strip()
+    return value
+
+
+def _ensure_sentence(text: str) -> str:
+    value = _clean_text(text)
+    if value and value[-1] not in END_PUNCTUATION:
+        value += "。"
+    return value
+
+
+def _source_link(item: NewsItem) -> str:
+    source = item.source_label or item.source_group or "来源"
+    if item.url:
+        return f"[{source} ↗]({item.url})"
+    return source
+
+
+def assemble_markdown(items: list[NewsItem], guide: dict[str, Any] | None = None) -> tuple[str, dict[str, Any]]:
+    guide = guide or {"theme": "", "threads": []}
+    lines: list[str] = []
+
+    theme = _clean_text(str(guide.get("theme") or ""))
+    if theme:
+        lines.extend(["## 导览", "", f"> {theme}", ""])
+
+    item_number = 1
+    for section in SECTION_ORDER:
+        section_items = [item for item in items if item.section == section]
+        if not section_items:
+            continue
+        lines.extend([f"## {section}", ""])
+        for item in section_items:
+            title = _clean_text(item.title or item.title_raw)
+            summary = _ensure_sentence(item.summary or item.summary_raw or "该条目暂无摘要。")
+            lines.extend(
+                [
+                    f"**{item_number}. {title}**",
+                    "",
+                    f"> {summary}{_source_link(item)}",
+                    "",
+                ]
+            )
+            item_number += 1
+
+    threads = guide.get("threads", []) or []
+    if threads:
+        lines.extend(["## 今日脉络", ""])
+        for thread in threads:
+            title = _clean_text(str(thread.get("title") or ""))
+            text = _ensure_sentence(str(thread.get("text") or ""))
+            if not title or not text:
+                continue
+            lines.extend([f"- **{title}**", f"  {text}", ""])
+
+    markdown = "\n".join(lines).strip()
+    report = validate_markdown(markdown, items)
+    return markdown, report