From 3e895adbab8a9a44e9a0aa0fe2557797b8cd34ed Mon Sep 17 00:00:00 2001
From: Moon Sungjoon <sumoon@seoulsaram.org>
Date: Mon, 26 Jun 2023 18:21:53 +0900
Subject: [PATCH] Add hwp hwpx support

hwp/hwpx has several custom MIME types

.hwp:
 - application/x-hwp
 - application/haansofthwp
 - application/vnd.hancom.hwp

.hwpx:
 - application/haansofthwpx
 - application/vnd.hancom.hwpx,
 - application/hwp+zip

Fixes #243
---
 Dockerfile                             |  3 ++-
 README.md                              |  1 +
 dangerzone/conversion/doc_to_pixels.py | 20 ++++++++++++++++++++
 dangerzone/gui/main_window.py          |  2 +-
 install/pyinstaller/pyinstaller.spec   | 14 ++++++++++++++
 5 files changed, 38 insertions(+), 2 deletions(-)

diff --git a/Dockerfile b/Dockerfile
index f8c79f0..78b2926 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -14,7 +14,8 @@ RUN apk --no-cache -U upgrade && \
     poppler-data \
     python3 \
     py3-magic \
-    tesseract-ocr
+    tesseract-ocr \
+    font-noto-cjk
 
 # Download the trained models from the latest GitHub release of Tesseract, and
 # store them under /usr/share/tessdata. This is basically what distro packages
diff --git a/README.md b/README.md
index d443ab4..1f45faf 100644
--- a/README.md
+++ b/README.md
@@ -36,6 +36,7 @@ Dangerzone can convert these types of document into safe PDFs:
 - ODF Spreadsheet (`.ods`)
 - ODF Presentation (`.odp`)
 - ODF Graphics (`.odg`)
+- Hancom HWP (Hangul Word Processor) (`.hwp`, `.hwpx`)
 - Jpeg (`.jpg`, `.jpeg`)
 - GIF (`.gif`)
 - PNG (`.png`)
diff --git a/dangerzone/conversion/doc_to_pixels.py b/dangerzone/conversion/doc_to_pixels.py
index 2151500..ac6055a 100644
--- a/dangerzone/conversion/doc_to_pixels.py
+++ b/dangerzone/conversion/doc_to_pixels.py
@@ -77,6 +77,26 @@ class DocumentToPixels(DangerzoneConverter):
             "application/vnd.oasis.opendocument.text-template": {
                 "type": "libreoffice",
             },
+            # .hwp
+            "application/vnd.hancom.hwp": {
+                "type": "libreoffice",
+            },
+            "application/haansofthwp": {
+                "type": "libreoffice",
+            },
+            "application/x-hwp": {
+                "type": "libreoffice",
+            },
+            # .hwpx
+            "application/vnd.hancom.hwpx": {
+                "type": "libreoffice",
+            },
+            "application/haansofthwpx": {
+                "type": "libreoffice",
+            },
+            "application/hwp+zip": {
+                "type": "libreoffice",
+            },
             # At least .odt, .docx, .odg, .odp, .ods, and .pptx
             "application/zip": {
                 "type": "libreoffice",
diff --git a/dangerzone/gui/main_window.py b/dangerzone/gui/main_window.py
index 1aac928..71a7c14 100644
--- a/dangerzone/gui/main_window.py
+++ b/dangerzone/gui/main_window.py
@@ -553,7 +553,7 @@ class DocSelectionWidget(QtWidgets.QWidget):
         self.file_dialog.setFileMode(QtWidgets.QFileDialog.ExistingFiles)
         self.file_dialog.setNameFilters(
             [
-                "Documents (*.pdf *.docx *.doc *.docm *.xlsx *.xls *.pptx *.ppt *.odt *.odg *.odp *.ods *.jpg *.jpeg *.gif *.png *.tif *.tiff)"
+                "Documents (*.pdf *.docx *.doc *.docm *.xlsx *.xls *.pptx *.ppt *.odt *.odg *.odp *.ods *.hwp *.hwpx *.jpg *.jpeg *.gif *.png *.tif *.tiff)"
             ]
         )
 
diff --git a/install/pyinstaller/pyinstaller.spec b/install/pyinstaller/pyinstaller.spec
index f8d2380..186b1d2 100644
--- a/install/pyinstaller/pyinstaller.spec
+++ b/install/pyinstaller/pyinstaller.spec
@@ -143,6 +143,20 @@ if p == "Darwin":
                     "CFBundleTypeName": "ODF Graphics Document",
                     "CFBundleTypeRole": "Viewer",
                 },
+                {
+                    "CFBundleTypeExtensions": ["hwp", "hwpx"],
+                    "CFBundleTypeIconFile": "../macos/document.icns",
+                    "CFBundleTypeMIMETypes": [
+                        "application/vnd.hancom.hwp",
+                        "application/haansofthwp",
+                        "application/x-hwp",
+                        "application/vnd.hancom.hwpx",
+                        "application/haansofthwpx",
+                        "application/hwp+zip",
+                    ],
+                    "CFBundleTypeName": "Hancom Office Document",
+                    "CFBundleTypeRole": "Viewer",
+                },
                 {
                     "CFBundleTypeExtensions": ["jpg", "jpeg"],
                     "CFBundleTypeIconFile": "../macos/document.icns",