@simplysm/sd-claude 14.0.76 → 14.0.77

package/claude/skills/sd-unpack/scripts/handlers/eml_handler.py

@@ -1,10 +1,21 @@
-"""EML 핸들러. 표준 라이브러리 email 모듈 사용."""
+"""EML 핸들러. 표준 라이브러리 email 모듈 사용.
+
+본문 일관화:
+- text/plain 있으면 그걸 body.md
+- text/plain 없고 text/html 있으면 HTML → 평문 변환(html2text) → body.md
+- HTML 변환 시 inline image (cid:) → 첨부 파일명 placeholder 치환
+
+CID 매핑:
+- Content-ID 헤더 있는 첨부는 cid_map 에 등록
+- images.rels.json 으로 CID↔파일명 양방향 추적
+"""
 from __future__ import annotations
 
 import email
 import hashlib
 import json
 import os
+import re
 from email.header import decode_header, make_header
 from pathlib import Path
 
@@ -29,24 +40,36 @@ def run(input_path: Path, out_dir: Path) -> None:
     raw = input_path.read_bytes()
     msg = email.message_from_bytes(raw)
 
-    headers = {
-        "From": _decode_header(msg.get("From")),
-        "To": _decode_header(msg.get("To")),
-        "Cc": _decode_header(msg.get("Cc")),
-        "Subject": _decode_header(msg.get("Subject")),
-        "Date": _decode_header(msg.get("Date")),
-        "Message-ID": _decode_header(msg.get("Message-ID")),
-    }
+    # 모든 헤더 보존 (envelope·X-* 등 원본 그대로). 동일 키 다수 등장 → list 누적.
+    headers: dict = {}
+    for key, raw in msg.items():
+        decoded = _decode_header(raw)
+        if key in headers:
+            existing = headers[key]
+            if isinstance(existing, list):
+                existing.append(decoded)
+            else:
+                headers[key] = [existing, decoded]
+        else:
+            headers[key] = decoded
     _common.write_text(
         out_dir / "headers.json",
         json.dumps(headers, ensure_ascii=False, indent=2),
     )
 
+    # README 의 헤더 섹션 표기용 envelope 키 (write_readme 의 dict 출력 한정)
+    envelope_keys = [
+        "From", "To", "Cc", "Bcc", "Subject", "Date", "Message-ID",
+        "Reply-To", "In-Reply-To", "References",
+    ]
+    readme_headers = {k: headers.get(k, "") for k in envelope_keys}
+
     body_text: str | None = None
     body_html: str | None = None
     attachments_dir = out_dir / "attachments"
     saved_attachments: list[Path] = []
     seen_hashes: set[str] = set()
+    cid_map: dict[str, str] = {}  # cid (without <>) → 첨부 파일명 (basename)
 
     for part in msg.walk():
         if part.is_multipart():
@@ -56,8 +79,10 @@ def run(input_path: Path, out_dir: Path) -> None:
         filename = part.get_filename()
         if filename:
             filename = _decode_header(filename)
+        cid_raw = (part.get("Content-ID") or "").strip()
+        cid = cid_raw.strip("<>") if cid_raw else ""
 
-        is_attachment = bool(filename) or "attachment" in disp
+        is_attachment = bool(filename) or "attachment" in disp or bool(cid)
 
         if is_attachment:
             payload = part.get_payload(decode=True) or b""
@@ -67,45 +92,96 @@ def run(input_path: Path, out_dir: Path) -> None:
                 continue
             seen_hashes.add(digest)
             _common.mkdir(attachments_dir)
-            dst = _common.unique_path(attachments_dir, filename or "attachment.bin")
+            dst_name = filename or (f"{cid}.bin" if cid else "attachment.bin")
+            dst = _common.unique_path(attachments_dir, dst_name)
             _common.write_bytes(dst, payload)
             saved_attachments.append(dst)
+            if cid:
+                cid_map[cid] = dst.name
         elif ctype == "text/plain" and body_text is None:
             body_text = _decode_payload(part)
         elif ctype == "text/html" and body_html is None:
             body_html = _decode_payload(part)
 
-    body_inline = None
+    # body.md: text/plain 우선, 없으면 HTML→평문
+    # body.from_html.md: HTML 있으면 항상 별도 생성 (인라인 이미지 위치 placeholder 포함)
+    if body_text:
+        body_md = body_text
+    elif body_html:
+        body_md = _html_to_md(body_html, cid_map)
+    else:
+        body_md = ""
+
     body_file_link = None
     body_html_link = None
-    if body_text:
-        if len(body_text) < 1000:
-            body_inline = body_text
-        else:
-            _common.write_text(out_dir / "body.md", body_text)
-            body_file_link = "body.md"
+    body_from_html_link = None
+    if body_md:
+        _common.write_text(out_dir / "body.md", body_md)
+        body_file_link = "body.md"
     if body_html:
         _common.write_text(out_dir / "body.html", body_html)
         body_html_link = "body.html"
+        # text/plain 우선 정책으로 body.md 가 plain 인 경우, HTML→평문 변환본도 별도로
+        # 보존 (인라인 이미지 위치 단서). body.md 자체가 from_html 이면 중복 회피.
+        if body_text:
+            from_html_md = _html_to_md(body_html, cid_map)
+            _common.write_text(out_dir / "body.from_html.md", from_html_md)
+            body_from_html_link = "body.from_html.md"
+
+    # CID↔파일명 매핑 (인라인 이미지 있을 때만)
+    if cid_map:
+        rels = {cid: f"attachments/{fname}" for cid, fname in cid_map.items()}
+        _common.write_text(
+            out_dir / "images.rels.json",
+            json.dumps(rels, ensure_ascii=False, indent=2),
+        )
 
     attachment_links: list[str] = []
     for ap in saved_attachments:
+        size = ap.stat().st_size
         recursed = maybe_recurse_attachment(ap, attachments_dir)
         if recursed is not None:
             os.unlink(_common.long_str(ap))
-            attachment_links.append(f"attachments/{recursed.name}/")
+            attachment_links.append(f"attachments/{recursed.name}/ ({_common.format_size(size)})")
         else:
-            attachment_links.append(f"attachments/{ap.name}")
+            attachment_links.append(f"attachments/{ap.name} ({_common.format_size(size)})")
 
     _common.write_readme(
         out_dir,
         source_name=input_path.name,
         source_size=input_path.stat().st_size,
-        tool="email (표준 라이브러리)",
-        loss_notes="없음 (이메일 본문 텍스트는 손실 없이 보존됨)",
-        body_inline=body_inline,
+        tool="email (표준 라이브러리) + html2text",
+        loss_notes=(
+            "본문은 body.md (text/plain 우선, 없으면 HTML→평문). "
+            "text/plain·HTML 둘 다 있을 때 HTML→평문(인라인 이미지 위치 placeholder 포함)은 body.from_html.md 별도. "
+            "원본 HTML 은 body.html, CID↔첨부 매핑은 images.rels.json (인라인 이미지 있을 때)."
+        ),
         body_file_link=body_file_link,
         body_html_link=body_html_link,
-        headers=headers,
+        body_from_html_link=body_from_html_link,
+        headers=readme_headers,
         attachments=attachment_links,
     )
+
+
+def _html_to_md(html: str, cid_map: dict[str, str]) -> str:
+    """HTML 본문 → 평문 md 변환. cid: 이미지 src 는 첨부 파일명 placeholder 로 치환."""
+    _common.ensure_pip("html2text")
+    import html2text
+
+    h = html2text.HTML2Text()
+    h.body_width = 0
+    h.ignore_links = False
+    h.ignore_images = False
+    md = h.handle(html)
+
+    # 1) `![alt](cid:XXX)` → `![<filename>](attachments/<filename>)`
+    def replace_cid_img(m: re.Match) -> str:
+        alt, cid_value = m.group(1), m.group(2).strip()
+        fname = cid_map.get(cid_value) or cid_map.get(cid_value.split("@")[0])
+        if fname:
+            return f"![{fname}](attachments/{fname})"
+        return f"![{alt}](cid:{cid_value})"
+
+    md = re.sub(r"!\[([^\]]*)\]\(cid:([^)]+)\)", replace_cid_img, md)
+    return md

package/claude/skills/sd-unpack/scripts/handlers/msg_handler.py

@@ -1,8 +1,13 @@
-"""MSG (Outlook) 핸들러. extract-msg 라이브러리 사용."""
+"""MSG (Outlook) 핸들러. extract-msg 라이브러리 사용.
+
+본문·CID·envelope 헤더 규약은 eml_handler 와 동일.
+"""
 from __future__ import annotations
 
+import email as stdemail
 import json
 import os
+import re
 from pathlib import Path
 
 from . import _common
@@ -15,18 +20,60 @@ def run(input_path: Path, out_dir: Path) -> None:
 
     msg = extract_msg.Message(str(input_path))
     try:
-        headers = {
-            "From": msg.sender or "",
-            "To": msg.to or "",
-            "Cc": msg.cc or "",
-            "Subject": msg.subject or "",
-            "Date": str(msg.date) if msg.date else "",
-        }
+        raw_header = getattr(msg, "header", None) or ""
+        # extract-msg 일부 버전 header 는 EmailMessage 객체 — str() 로 정규화
+        if not isinstance(raw_header, str):
+            raw_header = str(raw_header)
+
+        # 모든 헤더 보존: msg.header 의 raw RFC822 파싱 + extract-msg 의 정형 필드 보강
+        headers: dict = {}
+        if raw_header:
+            parsed = stdemail.message_from_string(raw_header)
+            for key, val in parsed.items():
+                if key in headers:
+                    existing = headers[key]
+                    if isinstance(existing, list):
+                        existing.append(val)
+                    else:
+                        headers[key] = [existing, val]
+                else:
+                    headers[key] = val
+        # extract-msg 정형 필드 (raw header 없을 때 fallback)
+        if not headers:
+            headers = {
+                "From": msg.sender or "",
+                "To": msg.to or "",
+                "Cc": msg.cc or "",
+                "Bcc": getattr(msg, "bcc", "") or "",
+                "Subject": msg.subject or "",
+                "Date": str(msg.date) if msg.date else "",
+            }
         _common.write_text(
             out_dir / "headers.json",
             json.dumps(headers, ensure_ascii=False, indent=2),
         )
 
+        envelope_keys = [
+            "From", "To", "Cc", "Bcc", "Subject", "Date", "Message-ID",
+            "Reply-To", "In-Reply-To", "References",
+        ]
+        readme_headers: dict = {}
+        for k in envelope_keys:
+            v = headers.get(k)
+            if v:
+                readme_headers[k] = v
+        # extract-msg 정형 필드로 envelope 보강 (raw header 에 없을 때)
+        if not readme_headers.get("From"):
+            readme_headers["From"] = msg.sender or ""
+        if not readme_headers.get("To"):
+            readme_headers["To"] = msg.to or ""
+        if not readme_headers.get("Cc"):
+            readme_headers["Cc"] = msg.cc or ""
+        if not readme_headers.get("Subject"):
+            readme_headers["Subject"] = msg.subject or ""
+        if not readme_headers.get("Date"):
+            readme_headers["Date"] = str(msg.date) if msg.date else ""
+
         body_text = msg.body or ""
         body_html_raw = getattr(msg, "htmlBody", None)
         body_html: str | None = None
@@ -36,21 +83,10 @@ def run(input_path: Path, out_dir: Path) -> None:
             else:
                 body_html = body_html_raw
 
-        body_inline = None
-        body_file_link = None
-        body_html_link = None
-        if body_text:
-            if len(body_text) < 1000:
-                body_inline = body_text
-            else:
-                _common.write_text(out_dir / "body.md", body_text)
-                body_file_link = "body.md"
-        if body_html:
-            _common.write_text(out_dir / "body.html", body_html)
-            body_html_link = "body.html"
-
         attachments_dir = out_dir / "attachments"
         attachment_links: list[str] = []
+        cid_map: dict[str, str] = {}
+
         for att in msg.attachments:
             _common.mkdir(attachments_dir)
             filename = (
@@ -58,6 +94,13 @@ def run(input_path: Path, out_dir: Path) -> None:
                 or getattr(att, "shortFilename", None)
                 or "attachment.bin"
             )
+            cid = (
+                getattr(att, "cid", None)
+                or getattr(att, "contentId", None)
+                or ""
+            )
+            if cid:
+                cid = str(cid).strip("<>")
             data = att.data
             if isinstance(data, str):
                 data = data.encode("utf-8")
@@ -65,24 +108,94 @@ def run(input_path: Path, out_dir: Path) -> None:
                 data = b""
             dst = _common.unique_path(attachments_dir, filename)
             _common.write_bytes(dst, data)
+            size = dst.stat().st_size
+            if cid:
+                cid_map[cid] = dst.name
             recursed = maybe_recurse_attachment(dst, attachments_dir)
             if recursed is not None:
                 os.unlink(_common.long_str(dst))
-                attachment_links.append(f"attachments/{recursed.name}/")
+                attachment_links.append(f"attachments/{recursed.name}/ ({_common.format_size(size)})")
             else:
-                attachment_links.append(f"attachments/{dst.name}")
+                attachment_links.append(f"attachments/{dst.name} ({_common.format_size(size)})")
+
+        # body.md: text 우선, 없으면 HTML→평문
+        # body.from_html.md: text·HTML 둘 다 있을 때 HTML→평문 변환본 별도 (이미지 위치 placeholder)
+        if body_text:
+            body_md = body_text
+        elif body_html:
+            body_md = _html_to_md(body_html, cid_map)
+        else:
+            body_md = ""
+
+        body_file_link = None
+        body_html_link = None
+        body_from_html_link = None
+        if body_md:
+            _common.write_text(out_dir / "body.md", body_md)
+            body_file_link = "body.md"
+        if body_html:
+            _common.write_text(out_dir / "body.html", body_html)
+            body_html_link = "body.html"
+            if body_text:
+                from_html_md = _html_to_md(body_html, cid_map)
+                _common.write_text(out_dir / "body.from_html.md", from_html_md)
+                body_from_html_link = "body.from_html.md"
+
+        if cid_map:
+            rels = {cid: f"attachments/{fname}" for cid, fname in cid_map.items()}
+            _common.write_text(
+                out_dir / "images.rels.json",
+                json.dumps(rels, ensure_ascii=False, indent=2),
+            )
 
         _common.write_readme(
             out_dir,
             source_name=input_path.name,
             source_size=input_path.stat().st_size,
-            tool="extract-msg",
-            loss_notes="없음 (Outlook 본문 + 헤더 + 첨부 보존)",
-            body_inline=body_inline,
+            tool="extract-msg + html2text",
+            loss_notes=(
+                "본문은 body.md (text 우선, 없으면 HTML→평문). "
+                "text·HTML 둘 다 있을 때 HTML→평문(인라인 이미지 위치 placeholder 포함)은 body.from_html.md 별도. "
+                "원본 HTML 은 body.html, CID↔첨부 매핑은 images.rels.json (인라인 이미지 있을 때)."
+            ),
             body_file_link=body_file_link,
             body_html_link=body_html_link,
-            headers=headers,
+            body_from_html_link=body_from_html_link,
+            headers=readme_headers,
             attachments=attachment_links,
         )
     finally:
         msg.close()
+
+
+def _header_field(raw_header: str, key: str) -> str:
+    """raw rfc822 header 문자열에서 특정 키 추출. 없으면 빈 문자열."""
+    if not raw_header:
+        return ""
+    try:
+        parsed = stdemail.message_from_string(raw_header)
+        return parsed.get(key, "") or ""
+    except Exception:
+        return ""
+
+
+def _html_to_md(html: str, cid_map: dict[str, str]) -> str:
+    """HTML 본문 → 평문 md 변환. cid: 이미지 src 는 첨부 파일명 placeholder 로 치환."""
+    _common.ensure_pip("html2text")
+    import html2text
+
+    h = html2text.HTML2Text()
+    h.body_width = 0
+    h.ignore_links = False
+    h.ignore_images = False
+    md = h.handle(html)
+
+    def replace_cid_img(m: re.Match) -> str:
+        alt, cid_value = m.group(1), m.group(2).strip()
+        fname = cid_map.get(cid_value) or cid_map.get(cid_value.split("@")[0])
+        if fname:
+            return f"![{fname}](attachments/{fname})"
+        return f"![{alt}](cid:{cid_value})"
+
+    md = re.sub(r"!\[([^\]]*)\]\(cid:([^)]+)\)", replace_cid_img, md)
+    return md