Update pipelines/anonymize.py

pipelines/anonymize.py

@@ -1,75 +1,63 @@
 import re
-import io
-from typing import Dict, Tuple
-from reportlab.pdfgen import canvas
+from typing import Tuple, Dict
 from reportlab.lib.pagesizes import A4
-from reportlab.pdfbase import pdfmetrics
-from reportlab.pdfbase.ttfonts import TTFont
-
-_MASK = "■" * 4  # マスク文字
+from reportlab.pdfgen import canvas
+from reportlab.lib.units import mm
+import io
 
-_EMAIL_RE = re.compile(r"[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}")
-_PHONE_RE = re.compile(r"(?:\+?\d{1,3}[ -]?)?(?:\(\d{2,4}\)[ -]?)?\d{2,4}[ -]?\d{2,4}[ -]?\d{3,4}")
-_NAME_HINTS = ["氏名", "Name", "名前"]
+# 簡易匿名化（メール・電話・氏名行のマスク）
+EMAIL_RE = re.compile(r"[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}")
+PHONE_RE = re.compile(r"(?:\+?\d{1,3}[ -]?)?(?:\(\d{2,4}\)[ -]?)?\d{2,4}[ -]?\d{2,4}[ -]?\d{3,4}")
+NAME_LINE_RE = re.compile(r"^(?:氏名|Name)[:：]?\s*([^\n\r]+)$", re.MULTILINE)
 
-def _mask_name_lines(text: str) -> Tuple[str, Dict[str, str]]:
+def anonymize_text(text: str) -> Tuple[str, Dict[str, str]]:
     mapping = {}
-    lines = text.splitlines()
-    for i, line in enumerate(lines):
-        if any(h in line for h in _NAME_HINTS):
-            mapping[f"LINE_{i}"] = line
-            lines[i] = re.sub(r".+", _MASK, line)
-    return "\n".join(lines), mapping
+    # メール
+    def _mask_email(m):
+        key = f"EMAIL_{len(mapping)+1}"
+        mapping[m.group(0)] = key
+        return key + "@masked"
+    text = EMAIL_RE.sub(_mask_email, text)
 
-def anonymize_text(text: str) -> Tuple[str, Dict[str, str]]:
-    """メール/電話/氏名行などをマスク。マッピングを返す。"""
-    mapping: Dict[str, str] = {}
+    # 電話
+    def _mask_phone(m):
+        ph = m.group(0)
+        # 短すぎるノイズはスキップ
+        if len(re.sub(r"\D", "", ph)) < 7:
+            return ph
+        key = f"TEL_{len(mapping)+1}"
+        mapping[ph] = key
+        return key
+    text = PHONE_RE.sub(_mask_phone, text)
 
-    def _apply(pat: re.Pattern, t: str, key_prefix: str) -> str:
-        nonlocal mapping
-        def _rep(m):
-            original = m.group(0)
-            mapping[f"{key_prefix}:{original}"] = original
-            return _MASK
-        return pat.sub(_rep, t)
+    # 氏名行
+    def _mask_name(m):
+        val = m.group(1).strip()
+        key = f"NAME_{len(mapping)+1}"
+        mapping[val] = key
+        return m.group(0).replace(val, key)
+    text = NAME_LINE_RE.sub(_mask_name, text)
 
-    out = text
-    out = _apply(_EMAIL_RE, out, "EMAIL")
-    out = _apply(_PHONE_RE, out, "PHONE")
-    out, name_map = _mask_name_lines(out)
-    mapping.update(name_map)
-    return out, mapping
+    return text, mapping
 
 def render_anonymized_pdf(text: str) -> bytes:
-    """
-    シンプルなテキストPDFを生成。
-    日本語フォントがある場合は `fonts/NotoSansCJKjp-Regular.otf` を使用。
-    無い場合はHelvetica（□が出る可能性あり）。
-    """
     buf = io.BytesIO()
     c = canvas.Canvas(buf, pagesize=A4)
     width, height = A4
 
-    # フォント登録（任意）
-    try:
-        pdfmetrics.registerFont(TTFont("NotoSansJP", "fonts/NotoSansCJKjp-Regular.otf"))
-        font_name = "NotoSansJP"
-    except Exception:
-        font_name = "Helvetica"
-
-    c.setFont(font_name, 10)
-
-    margin = 40
-    y = height - margin
-    line_height = 14
+    # 素朴なテキスト描画（自動改頁）
+    margin_x = 15 * mm
+    margin_y = 15 * mm
+    y = height - margin_y
+    c.setFont("Helvetica", 10)
 
     for line in text.splitlines():
-        if y < margin + line_height:
+        if y < margin_y:
             c.showPage()
-            c.setFont(font_name, 10)
-            y = height - margin
-        c.drawString(margin, y, line[:1800])
-        y -= line_height
+            c.setFont("Helvetica", 10)
+            y = height - margin_y
+        c.drawString(margin_x, y, line[:180])  # 超長行は素直に切る
+        y -= 12
 
     c.showPage()
     c.save()