Create anonyize.py

pipelines/anonyize.py

@@ -0,0 +1,58 @@
+import io
+import re
+from reportlab.pdfgen import canvas
+from reportlab.lib.pagesizes import A4
+from reportlab.lib.units import mm
+
+RE_EMAIL = re.compile(r"[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}")
+RE_PHONE = re.compile(r"(?:\+?\d{1,3}[ -]?)?(?:\(\d{2,4}\)[ -]?)?\d{2,4}[ -]?\d{2,4}[ -]?\d{3,4}")
+RE_NAME = re.compile(r"(?:(?:氏名[:：]?)[\t ]*)(\S+)")
+
+
+def anonymize_text(text: str):
+    anon_map = {}
+
+    def repl_email(m):
+        val = m.group(0)
+        key = "<EMAIL>"
+        anon_map.setdefault(key, []).append(val)
+        return key
+
+    def repl_phone(m):
+        val = m.group(0)
+        key = "<PHONE>"
+        anon_map.setdefault(key, []).append(val)
+        return key
+
+    def repl_name(m):
+        val = m.group(1)
+        key = "<NAME>"
+        anon_map.setdefault(key, []).append(val)
+        return m.group(0).replace(val, key)
+
+    t = RE_EMAIL.sub(repl_email, text)
+    t = RE_PHONE.sub(repl_phone, t)
+    t = RE_NAME.sub(repl_name, t)
+
+    return t, anon_map
+
+
+def render_anonymized_pdf(text: str) -> bytes:
+    buf = io.BytesIO()
+    c = canvas.Canvas(buf, pagesize=A4)
+    width, height = A4
+
+    margin = 15 * mm
+    line_height = 5 * mm
+    x = margin
+    y = height - margin
+
+    for line in text.splitlines():
+        if y < margin:
+            c.showPage()
+            y = height - margin
+        c.drawString(x, y, line[:1200])
+        y -= line_height
+
+    c.save()
+    return buf.getvalue()