Quivi is an image viewer (specialized for comic/manga reading) which supports many file formats and compressed (zip, rar) files. It is aimed for fast & easy file browsing with keyboard or mouse.

The home page for the original distribution of Quivi is http://quivi.sourceforge.net/

This fork was made with the primary purpose of adding 64-bit compatibility. The original Python code has been ported to Python 3. Most of the original functionality is unchanged, and some additional features have been added.

The 64-bit executable prevents issues with loading large files. The upgrade to Python 3 fixes some Unicode issues that could prevent opening files depending on the name. Upgrading the image libraries used for displaying images adds support for webp images.

• Quivi will now remember if you closed the application while in fullscreen mode and automatically re-open in fullscreen mode. This can be disabled in the configuration ("Remember full screen on close").
• Added support for mouse auxiliary buttons (Aux1 and Aux2). These can be assigned behavior in the settings menu. By default they do nothing.
• Internal change to how the Config file (pyquivi.ini) is written to should reduce the chances of it becoming corrupted.
  • If the config file does become corrupt, it will re-create a new file from scratch instead of making the application inoperable.
  • The config file is now stored as UTF-8 instead of the system default. It will attempt to open the file both as UTF-8 and the system default, but always write as UTF-8.
• New feature: Placeholders. These act like Favorites, but with some changes. The behavior can be tweaked in Settings.
  • Placeholders save a path and a specific page. They are intended to quickly resume reading something you're in the middle of.
  • Only one placeholder can exist for an item; saving a new placeholder will replace the existing one.
  • (Optionally) Only a single place holder can be used for anything. Saving any new placeholder will replace the existing one.
  • (Optionally) Placeholders will be deleted upon opening.
  • (Optionally) Opening a folder/archive that has a placeholder will automatically load the placeholder.
  • Placeholders and Favorites can optionally be separated into separate submenus of the Favorites menu on the menu bar. This is a configuration option in settings. Placeholders and favorites are also included in a right-click context menu and are always separated there.
• Holding down shift while using the scrollwheel will scroll horizontally.
• Holding down ctrl while using the scrollwheel zoom in/out on the mouse cursor's location (instead of the center of the image).
• Added "Drag image" as an explicit command, instead of the default behavior of always dragging with left click. This means left click can be reassigned to a different action, such as Next image, and it will never attempt to move the image. The old behavior can be restored via an option in the Mouse tab.
  • Changed the default always drag behavior to have a threshold. If the mouse is moved more than x pixels, it is treated as a drag command. Otherwise it is a left button click (the other behavior was effectively a threshold of 0 pixels).
• Changed how the list of commands for keyboard/mouse are populated. Some commands are now marked as keyboard or mouse only. This technically removes functionality, but it makes the menu slightly easier to work with.
• Reworked how Cairo resizes image. If Cairo is enabled, images will be rescaled via a matrix operation instead of creating a new image. This is massively faster for the initial zoom, but slower for panning. The current approach is to use a high-quality resample while zooming, but a fast resample while panning. In the background, a high-quality resized image is created and used for panning when it is available. This is (roughly) the same approach taken by Eye of Gnome. It should be possible to always smoothing scale/pan at high quality, as GIMP does this, but I haven't figured it out.
• 16-bit precision images should work with PIL.
• New feature: Right-to-left viewing. When opening an image that is longer than the screen, start at the top-right corner instead of the top-left. This is for images that use a right-to-left reading order (i.e. Japanese manga). It will have no effect if the image is resized to the width of the viewer.
  • This will also invert the direction of horizontal mouse wheel scrolling. Mouse wheel up will always move to the "start" of the image.
  • This can be toggled in the options. To the best of my knowledge, there is no way to detect the intended reading direction.
• New feature: Vertical scrolling past the bottom of the image will scroll to the right (or left - RTL flips this).
  • This can be toggled in the options. There's no effect if the image is resized to fit the screen.
  • This is intended mostly for mousewheel scrolling, but it will work for other scrolling as well. It will not do anything for dragging an image.
  • To prevent accidental horizontal scrolling, there's a small "buffer" that needs to be overcome; scroll while at the bottom 3+ times and it should start.
• New feature: Spread page viewing. If enabled and the current image is wider than it is tall (width > height), fit-to-width will be calculating using half of the image width instead of the full width.
  • The intent is for when viewing standard page images that includes image files that are two physical pages joined together, i.e. full-page spreads. This should keep the zoom level roughly consistent with the rest of the book.
  • This will lead to false positives if viewing landscape pages, or any digital art that doesn't try to adhere to a standard page layout. It can be toggled via a hotkey. This will automatically resize the image.
  • There's no indication that this is being done while in fullscreen, so if two pages are joined together but don't have shared art and contain ample margins, it will be easy to accidentally skip pages.
• Increased the size of the image cache to better account for modern hardware (it's still hardcoded and doesn't check system RAM available or anything).
• New feature: Move the currently opened archive to another folder.
  • Folders can be saved to settings to quickly move zip archives to a specific "archive" folder
  • There is currently no way to reorder or delete these saved folders, only add.
• New feature: Bindable context menu. This can be assigned as any other shortcut to a mouse key, e.g. right click or middle click.
  • The menu itself is not customizable, but will include most navigation/display options that may be useful when in full screen mode.
  • This resulted in a significant backend rework of the menu system to enable nested menus. For now, the main menu bar is still exclusively "flat".
• The delete option in the menu can now delete zip (or rar) archive files. The delete option in the menu will either delete individual images (when viewing a folder) or the entire archive (when viewing an archive). Separate bindable commands exist that will only attempt to delete one type, e.g. for backwards compatibility.
• Added support for animated GIFs. Other animated formats should also work (e.g. webp) but haven't been tested.
  • Support is limited to PIL. FreeImage does (should) support animation, but the FreeImage project is no longer active so I'm considering replacing the library.
  • Zooming is not enabled. Images are always 100%, regardless of what the status bar says. I'll revisit this later to see if Cairo can be used for fast zooming on an animated image.
  • GIFs can specify a maximum number of times to play the animation. This is not supported - images will always play endlessly.
  • Windows uses some tricks to work around inaccurate sleep/timer issues. I believe I have a good implementation in place.
  • This is less tested than most other features simply because most manga are not animated.
• New configuration option: Hide mouse cursor if not moved for x seconds. Setting to 0 (default) will disable this feature.
  • If the mouse is not moved, the cursor will eventually disappear. This only applies to movement, not mouse clicks or any other interaction.
  • This only applies to the image viewing canvas, not other parts of the application. If the cursor is the hand, it will disappear.
  • The check for this is purposely low-accuracy, so it may take up to a second more than the configured value to disappear.

• Version 2.0.9 changes how the automatic zoom setting is stored in the configuration file. Existing configuration files will still read the old value, but if switching back to an older version of the application, the application will fail to load
  • Workaround is to manually edit pyquivi.ini to change fittype = FitType.WIDTH_IF_LARGER (as an example) to fittype = 1
  • This is likely a one-off; other values that use magic numbers are not planned for a similar switch to text.
• Version 2.0.9 changes the behavior of the 'Fit x' options in the View header menu to be 'Fit x if larger'.

For the most part, existing functionality is being kept intact. A few things were dropped either due to difficulty porting the code from Python2 to Python3, or because they aren't needed any more.

• Support for mangafox and onemanga was removed. Neither site is still active. The relevant code could be repurposed for other website(s) with a public API, but I would rather focus on being an offline reader.
• Embedded RAR support was removed. RAR files can still be opened as long as winrar is installed.
• Internally, GDI is longer supported. I have not been able to get it to work with embedded files.

Most of the 2 -> 3 conversion was automatic, which did leave some artifacts that need to be cleaned up. It's also likely there's some real division that needs to be corrected. Some of the Unicode hacks were removed, but others may remain. Python3 should give much better Unicode support overall.

Type hints have been added to some methods. This is mostly added to subclasses to ensure they adhere to the common interface. I don't intend on adding type hints everywhere, only where "most needed" (or just trivial functions)

• Updated to support Python 3.13.3. Minimum required version is Python 3.6 due to the use of fstrings.
• wx updated to 4.2.4
• wx.lib.pubsub was split off of Wx as Pypubsub; version 4.0.3 is used.
• Image display supports FreeImage and PIL (Pillow). GDI works for local files only, not files within compressed archives. Cairo can be used to speed up zooming operations.
• Removed online manga reader support. This is mostly to simplify the conversion process, as it allowed dropping httplib and beautifulsoup from the project
• Removed third party path utlity; pathlib (core Python module) is used instead.
• Tests haven't been modified, except for the automated conversion.

• 16-bit precision images do not work with FreeImage. PIL support is hackish.

A large number of text strings have been added or modified. The existing translations have not been updated. All of the supported languages are only partially translated as a result.