Compare commits

...

57 commits
v1.2 ... master

Author SHA1 Message Date
GitHub Actions
4d21a05d80 Update CImg to v.3.3.6 2024-06-03 00:28:47 +00:00
GitHub Actions
a5b9c64c7a Update CImg to v.3.3.5 2024-04-17 20:26:24 +00:00
Aaron Liu
56bac07109
try checkoutv2 per https://github.com/actions/checkout/issues/664 2024-04-02 22:22:36 -04:00
Quentin Santos
92cf12d2ec Fix some minor typos 2024-03-23 10:36:25 +01:00
Aaron Liu
84a3616796
fix entry point display and remove spurious newlines 2024-03-20 16:29:33 -04:00
Stefan Haustein
38ec4249b8 small fixes 2024-03-20 20:59:16 +01:00
Stefan Haustein
df32d7e1c1 Updated README, pointing at the API. 2024-03-20 20:14:42 +01:00
Stefan Haustein
8aca9afaf7 cleanup / formatting 2024-03-06 18:51:31 +01:00
Stefan Haustein
917b381bc3 preparation for more character groups 2024-03-06 18:37:16 +01:00
GitHub Actions
a18c4e62c6 Update CImg to v.3.3.4 2024-03-03 00:27:09 +00:00
Stefan Haustein
e5787c0288 Slight reformatting 2024-02-21 20:41:27 +01:00
Stefan Haustein
2fa6ece556 Change getPixel to actually transfer a full pixel with all channels 2024-02-21 20:34:50 +01:00
Aaron Liu
4cb242b7f5
re-add error message formatting from 7b5d790 2024-02-17 14:14:32 -05:00
Stefan Haustein
b783c67f66 split tiv_lib.h/.cpp out of tiv.cpp, adressing issue #52 2024-02-14 20:36:34 +01:00
Stefan Haustein
8c9ca72902 slightly re-word the license. 2024-02-14 19:55:08 +01:00
Stefan Haustein
a901624905 Replace constants copied from sysexits.h with our own definitions 2024-02-14 19:51:05 +01:00
Stefan Haustein
1c2c1c27e8 Always define the exit codes locally -- even on posix. 2024-02-14 19:23:02 +01:00
Alessandro de Oliveira Faria (A.K.A.CABELO)
190689961e
Fix @cabelo's name (#139)
Sorry!
2024-02-11 21:29:12 -05:00
Aaron Liu
a216fc64eb
Workflow for updating CImg (#137)
Also updates CImg to 3.3.3... huh, coincidence? <!-- I think not -->
2024-02-02 16:28:03 -05:00
Aaron Liu
ca97659f83
redo 9969dd1 2024-02-01 18:30:28 -05:00
Aaron Liu
dedaa1b7c2
redo 7b5d790 and f51dc08 2024-02-01 18:27:29 -05:00
Aaron Liu
3373e02e4c
CI: say chosen image name and only trigger if src files are modified or we want it to 2024-02-01 18:21:13 -05:00
Aaron Liu
4e1709a7c5
downgrade needed c++ to 17 2024-02-01 18:20:56 -05:00
Stefan Haustein
cc4b4b30bd
Update README.md
Clean up the installation options
2024-02-01 18:20:55 -05:00
Stefan Haustein
516107ca4f
Update README.md
Add another hint to install ImageMagic to the common problems / troubleshooting section.
2024-02-01 18:20:55 -05:00
Aaron Liu
8d785902e6
merge 2024-02-01 18:20:49 -05:00
Stefan Haustein
bc2d21db09
change format to string streams, as format is only available in 0x20, which complicates compiling on a mac 2024-02-01 18:18:59 -05:00
Aaron Liu
a64e5527a9
Add news and license note 2024-02-01 18:18:58 -05:00
Aaron Liu
e70f05fdf6
test dir mode and add more images that could be tested 2024-02-01 18:18:58 -05:00
Aaron Liu
7b5d7907d1
fix spurious size detect error and format error messages from system
TODO: fix unexpanded tilde on Windows
2024-02-01 18:18:50 -05:00
Aaron Liu
497b533010
hopefully this fixes gcc 2024-02-01 18:16:53 -05:00
Aaron Liu
c6b8855930
fix CI to force gcc 13 2024-02-01 18:14:28 -05:00
Aaron Liu
ac334fdfaa
lots of changes after a lot of time
rewrote some functions to use an accompanying function that returns a string; this prepares for splitting us into a library
fixed weird bug caused by CImg update
some endls have been replaced with \n
slightly more consistent function naming
more comments

1.3 part 1 out of ?
2024-02-01 18:14:27 -05:00
Aaron Liu
d9c1b47816
fix last commit 2024-02-01 18:14:23 -05:00
Aaron Liu
7b3c8286fe
make positioning of flags in emitTermColor consistent with other funcs 2024-02-01 18:13:31 -05:00
Aaron Liu
f51dc08b0a
correct image size in warning
it's 20x6, not 80x24
2024-02-01 18:07:20 -05:00
Aaron Liu
e636a03c13
separate print and string generation functions
string generation during image processing now have their own functions
2024-02-01 18:05:50 -05:00
Aaron Liu
9969dd1cff
rename some functions
change emit to print; emit will be used for things that generate strings/streams
use camelCase, closer to google style
specify that color is terminal color
2024-02-01 18:04:23 -05:00
Aaron Liu
c001eca18d
minor optimizations & bump version
make flags an int8_t
replace excessive instances of endl with \n
make a boolean const with better assignment
2024-02-01 18:01:06 -05:00
Aaron Liu
e72b11d359
more comments 2024-02-01 17:52:19 -05:00
Aaron Liu
e358e372ec
fix weird CImg bug 2024-02-01 17:49:14 -05:00
Stefan Haustein
0bfc892f4f Use a function reference instead of a cimg object for the inner conversion methods. Preparation for factoring them out to a separate file (tiv_lib.h / .cpp) 2024-02-01 21:22:39 +01:00
Stefan Haustein
bba4a692db clarify comments some more 2024-01-24 20:13:28 +01:00
Stefan Haustein
512b3dd91c Align license header with README.md 2024-01-24 19:55:40 +01:00
Stefan Haustein
1154264c54 Clarify / clean up some comments 2024-01-24 19:49:07 +01:00
Aaron Liu
71d1a80d14
PR #136 from penguin359/penguin359-ci: Basic testing CI 2024-01-19 14:18:41 -05:00
Loren M. Lang
ba7bafe092
Dropped install and don't restrict color palette during test 2024-01-19 11:15:19 -08:00
Loren M. Lang
2b46a7a227 Specifiy specific size and colors 2024-01-16 21:41:27 -08:00
Loren M. Lang
8f91b297f1 Create a basic Continuous Integration test
Run a program test at the end.
2024-01-16 21:40:53 -08:00
Aaron Liu
7573de8a61
fix a stupid typo and clarify windows error code 2023-12-06 20:38:24 -05:00
Aaron Liu
c4e12b539a
bump version 2023-12-06 16:04:41 -05:00
Aaron Liu
e5c4677542
Merge branch 'master' of https://github.com/stefanhaustein/TerminalImageViewer 2023-12-06 15:58:41 -05:00
Aaron Liu
11d82bde63
try to help #132
also added a bit of docs
2023-12-06 15:57:55 -05:00
Aaron Liu
c87dd19ea1
should fix #132 2023-12-06 15:49:06 -05:00
Aaron Liu
c11b449672
Merge pull request #131 from Huell-Howitzer/patch-1 2023-11-30 20:34:20 -05:00
Huell-Howitzer
55e6f062a3
Fix typo 2023-11-30 16:45:52 -08:00
Tim Stewart
c29bdc7645 fix typo 2023-10-04 10:47:45 +02:00
10 changed files with 2936 additions and 1639 deletions

35
.github/workflows/ci.yml vendored Normal file
View file

@ -0,0 +1,35 @@
name: Build and run
on:
workflow_dispatch:
push:
branches: [ "master" ]
paths: ['src/**']
pull_request:
paths: ['src/**']
jobs:
build:
runs-on: ubuntu-latest
steps:
- name: Checkout
uses: actions/checkout@v4
- name: Install dependencies
run: sudo apt-get install -qy imagemagick
- name: Validate gcc version
run: |
if [[ $(gcc --version | awk '/gcc/ && ($3+0)>13{print "gcc-13+"}') != "gcc-13+" ]]; then
# Script courtesy of https://stackoverflow.com/a/67791068/16134571
sudo apt install gcc-13 g++-13
sudo update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-13 100 --slave /usr/bin/g++ g++ /usr/bin/g++-13 --slave /usr/bin/gcov gcov /usr/bin/gcov-13
sudo update-alternatives --set gcc /usr/bin/gcc-13
fi
- name: Build
run: make -C src
- name: Test
run: |
images=('/usr/local/share/icons/hicolor/128x128/apps/microsoft-edge.png' '/usr/local/share/icons/hicolor/128x128/apps/CMakeSetup.png' '/usr/local/doc/cmake/html/_static/file.png' '/usr/local/lib/android/sdk/extras/google/google_play_services/samples/tagmanager/cuteanimals/res/drawable/cat_1.jpg' '/usr/local/lib/android/sdk/extras/google/google_play_services/samples/wallet/res/drawable-ldpi/icon.png' '/usr/local/lib/android/sdk/extras/google/google_play_services/samples/wallet/res/drawable-hdpi/icon.png' '/usr/share/plymouth/themes/spinner/watermark.png' '/usr/share/apache2/icons/apache_pb.png' '/usr/share/doc/libpng-dev/examples/pngtest.png')
image=${images[ $RANDOM % ${#images[@]} ]} # Get random image
./src/tiv -w 160 -h 48 $image # Get random image
echo $image
./src/tiv -w 160 -h 48 /usr/share/pixmaps # Dir mode

43
.github/workflows/cimg.yml vendored Normal file
View file

@ -0,0 +1,43 @@
name: Update CImg
on:
schedule:
- cron: 0 0 3 * * # The first day after this workflow was merged
release:
types: [published]
workflow_dispatch:
jobs:
check-version:
name: Check for updates
runs-on: ubuntu-latest
outputs:
latest-tag: ${{ steps.latest-tag.outputs.result }}
steps:
- name: Fetch the latest tag (could be buggy)
uses: actions/github-script@v7
id: latest-tag
with:
result-encoding: string
script: |
return (await github.rest.repos.listTags({ owner: 'GreycLab', repo: 'CImg', per_page: 28 })).data[27].name
pull-file:
name: Update CImg.h
runs-on: ubuntu-latest
needs: check-version
if: ${{ needs.check-version.outputs.latest-tag != 'v.3.3.6' }}
steps:
- uses: actions/checkout@v2
name: Checkout this repository
with:
token: ${{ secrets.CIMG_UPDATE_TOKEN }}
- name: Download new CImg version
uses: carlosperate/download-file-action@v2
with:
file-url: 'https://github.com/GreycLab/CImg/raw/${{ needs.check-version.outputs.latest-tag }}/CImg.h'
location: 'src'
- name: Commit new CImg version (and update script)
run: |
sed -i 's/v.3.3.6/${{ needs.check-version.outputs.latest-tag }}/' .github/workflows/cimg.yml
git config user.name 'GitHub Actions'
git config user.email 'actions@github.com'
git commit -am "Update CImg to ${{ needs.check-version.outputs.latest-tag }}"
git push

4
.gitignore vendored
View file

@ -47,6 +47,10 @@ x64/
*.vcxproj.*
###Visual Studio Code###
.vscode/
###Java###
# Compiled class file

11
LICENSE
View file

@ -1,4 +1,4 @@
Copyright (c) 2023 Stefan Haustein and Aaron Liu
Copyright (c) 2024 Stefan Haustein and Aaron Liu
This program is free software: you may copy, redistribute and/or modify it
under the terms of (at your option) either the Apache License, Version 2.0,
@ -10,5 +10,10 @@ WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
For details, see either
<https://www.apache.org/licenses/LICENSE-20> for the Apache License, Version 2.0
, or <https://www.gnu.org/licenses/gpl-3.0.en.html> for the GNU GPL, version 3.
<https://www.apache.org/licenses/LICENSE-20> for the Apache License, Version 2.0,
or <https://www.gnu.org/licenses/gpl-3.0.en.html> for the GNU GPL, version 3.
Note that the code contained in this package contains the CImg library,
which is licensed under either [CeCILL 2.0](https://spdx.org/licenses/CECILL-2.0.html)
(close to GPL and compatible with it) or [CeCILL-C](https://spdx.org/licenses/CECILL-C)
(close to LGPL and compatible with Apache).

View file

@ -21,19 +21,22 @@ The shell will expand wildcards. By default, thumbnails and file names will be d
## News
- 2020-10-22: The Java version is now **deprecated**. Development has long shifted to the C++ version since that was created, and the last meaningful update to it was in 2016.
- 2021-05-22: We now support Apple Clang, thanks to the C++ filesystem library being no longer experimental. Issue forms have also been added to the GitHub repository.
- 2023-09-29: Today marks the 40th anniversary of the GNU project. If you haven't learned the news concerning it and Stallman, please do. In project news, @aaronliu0130 will probably be developing this project from now on as the original author has moved on to better things to do. Support for MSVC has been added and the repository is now under an Apache 2.0 or GPL3 dual license. CI building for each release will hopefully be setup soon. The main program has also adopted a mostly Google code-style because I (aaron) think it simply makes sense.
- 2024-03-20: Added a section on how to use the API.
- 2024-02-01: We are currently working on splitting the source code into dependency-free library files and a client that uses CImg.
- 2023-09-29: Today marks the 40th anniversary of the GNU project. If you haven't learned the news concerning it and Stallman, please do.
Support for MSVC has been added and the repository is now under an Apache 2.0 or GPL3 dual license. CI building for each release will hopefully be setup soon. The main program has also adopted a mostly Google code-style because I (aaron) think it simply makes sense.
`SPDX-License-Identifier: Apache-2.0 OR GPL-3.0-or-later`
- 2021-05-22: We now support Apple Clang, thanks to the C++ filesystem library being no longer experimental. Issue forms have also been added to the GitHub repository.
- 2020-10-22: The Java version is now **deprecated**. Development has long shifted to the C++ version since that was created, and the last meaningful update to it was in 2016.
## Installation
> [!IMPORTANT]
> All installation methods require installing ImageMagick, a required dependency. Most package managers should install it automatically.
### Build from source
### All platforms: Build from source
Our makefile currently only supports `g++`. It should be possible to compile `tiv` manually using any of your favorite compilers that support C++20 and Unix headers (`ioctl.h`, specifically) or `windows.h`. PRs are welcome.
Our makefile currently only supports `g++`. It should be possible to compile `tiv` manually using any of your favorite compilers that support C++17 and Unix headers (`ioctl.h` and `sysexits.h`, specifically) or `windows.h`. PRs are welcome.
```sh
git clone https://github.com/stefanhaustein/TerminalImageViewer.git
@ -44,52 +47,67 @@ make
sudo make install
```
### Homebrew
Please don't forget to install ImageMagick... On Debian based Linux via `sudo apt install imagemagick` and
on MacOS via `brew install imagemagick`.
Option 1:
### Mac: Homebrew
```sh
brew install tiv
```
Option 2 (deprecated, for macOS Mojave and below):
As the original Apple Shell only supports 256 color mode (-256) and there seems to be some extra
line spacing, distorting the image, we also recommend installing iTerm2:
```sh
brew tap stefanhaustein/TerminalImageViewer https://github.com/stefanhaustein/TerminalImageViewer
brew install terminalimageviewer
```
brew install --cask iterm2
```
### Snap (outdated)
### Third-Party Packages
sudo snap install --edge tiv
## Packages
- @aaronliu0130 has added homebrew support.
- @megamaced has created [an RPM for SUSE](https://build.opensuse.org/package/show/home:megamaced/terminalimageviewer)
- @bperel has created [a Docker image](https://hub.docker.com/r/bperel/terminalimageviewer)
- @teresaejunior has created a snapcraft.yaml file, which can build a Snap package with `sudo docker run -it --rm -v "$PWD:$PWD" -w "$PWD" snapcore/snapcraft sh -c 'apt-get update && snapcraft'`, and then installed with `sudo snap install --dangerous ./*.snap`. This is no longer supported.
- An AUR package is also available, though the maintainer appears to be inactive. Replacements coming soon.
## Common problems / Troubleshooting
- Errors such as "unrecognized file format"? Make sure ImageMagic is installed.
- On some linux platforms, an extra flag seems to be required: `make LDLIBS=-lstdc++fs` (but it also breaks MacOs), see <https://github.com/stefanhaustein/TerminalImageViewer/issues/103>
- If you see strange horizontal lines, the characters don't fully fill the character cell. Remove additional line spacing in your terminal app
- Wrong colors? Try -256 to use a 256 color palette instead of 24 bit colors
- Strange characters? Try -0 or install an use full unicode font (e.g. inconsolata or firacode)
## Using the TIV API
Tiv can be used as an API. So if you always wanted to run your favorite FPS in a shell, this is the opportunity.
All the code useful as a library is isolated in [tiv_lib.h](https://github.com/stefanhaustein/TerminalImageViewer/blob/master/src/tiv_lib.h)
and [tiv_lib.cc](https://github.com/stefanhaustein/TerminalImageViewer/blob/master/src/tiv_lib.cc).
The main entry point is
```cpp
CharData findCharData(GetPixelFunction get_pixel, int x0, int y0, const int &flags)
```
The call takes a std::Function that allows the TIV code to request pixels from your framebuffer.
From this framebuffer, the call will query pixels for a 4x8 pixel rectangle, where x0 and y0
define the top left corner. The call searches the best unicode graphics character and colors to approximate this
cell of the image, and returns these in a CharData struct.
## Contributions
- 2019-03-26: Exciting week: @Cableo has fixed output redirection, @boretom has added cross-compilation support to the build file and @AlanDeSmet has fixed tall thumbnails and greyscale images.
- 2019-03-26: Exciting week: @cabelo has fixed output redirection, @boretom has added cross-compilation support to the build file and @AlanDeSmet has fixed tall thumbnails and greyscale images.
- 2020-07-05: @cxwx has fixed homebrew support.
I am happy to accept useful contributions under the Apache 2.0 licencse, but...
I am happy to accept useful contributions under the Apache 2.0 license, but...
- Before investing in larger contributions, please use an issue to discuss this
- Pull requests should be as "atomic" as possible. I won't accept any pull request doing multiple things at once.
- This program currently only depends on CImg and ImageMagick as image processing libraries and I'd prefer to keep it that way.
- Support for additional platforms, CPUs or similar will require somebody who is happy to help with maintenance, in particular if I don't have access to it.
## Common problems
- On some linux platforms, an extra flag seems to be required: `make LDLIBS=-lstdc++fs` (but it also breaks MacOs), see <https://github.com/stefanhaustein/TerminalImageViewer/issues/103>
- If you see strange horizontal lines, the characters don't fully fill the character cell. Remove additional line spacing in your terminal app
- Wrong colors? Try -256 to use a 256 color palette instead of 24 bit colors
- Strange characters? Try -0 or install an use full unicode font (e.g. inconsolata or firacode)
## Examples
Most examples were shot with the Java version of this program, which should have equivalent output but slower by millenia in CPU years.
@ -103,3 +121,8 @@ If multiple images match the filename spec, thumbnails are shown.
For the example below, the top image was generated with the character optimization disabled via the `-0` option.
![Comparison](https://i.imgur.com/OzdCeh6.png)
## Licensing
You are free to use this code under either the GPL (3 or later) or version 2.0 of the Apache license. We include the CImg library, which
is licensed under either [CeCILL 2.0](https://spdx.org/licenses/CECILL-2.0.html) (close to GPL and compatible with it) or [CeCILL-C] (https://spdx.org/licenses/CECILL-C) (close to LGPL and compatible with Apache).

3366
src/CImg.h

File diff suppressed because it is too large Load diff

View file

@ -1,7 +1,7 @@
PROGNAME = tiv
OBJECTS = tiv.o
OBJECTS = tiv.o tiv_lib.o
CXX ?= g++
CXXFLAGS ?= -O2 -fpermissive
INSTALL ?= install
@ -12,12 +12,14 @@ prefix ?= /usr/local
exec_prefix ?= $(prefix)
bindir ?= $(exec_prefix)/bin
override CXXFLAGS += -std=c++20 -Wall -fexceptions
override CXXFLAGS += -std=c++17 -Wall -fexceptions
override LDFLAGS += -pthread
all: $(PROGNAME)
tiv.o: CImg.h
tiv_lib.o: tiv_lib.h
tiv.o: CImg.h tiv_lib.h
$(PROGNAME): $(OBJECTS)
$(CXX) $(LDFLAGS) $^ -o $@ $(LOADLIBES) $(LDLIBS)

View file

@ -1,23 +1,21 @@
/*
* Copyright (c) 2023 Aaron Liu
* Copyright (c) 2017-2023, Stefan Haustein, Aaron Liu
*
* This file is free software: you may copy, redistribute and/or modify it
* under the terms of the GNU General Public License as published by the
* Free Software Foundation, either version 3 of the License, or (at your
* option) any later version.
* This file is free software: you may copy, redistribute and/or modify it
* under the terms of the GNU General Public License as published by the
* Free Software Foundation, either version 3 of the License, or (at your
* option) any later version.
*
* This file is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License for more details.
* This file is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <https://www.gnu.org/licenses/>.
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <https://www.gnu.org/licenses/>.
*
* This file incorporates work covered by the following copyright and
* permission notice:
*
* Copyright (c) 20172021, Stefan Haustein
* Alternatively, you may copy, redistribute and/or modify this file under
* the terms of the Apache License, version 2.0:
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
@ -37,382 +35,39 @@
#include <cmath>
#include <filesystem>
#include <fstream>
#include <functional>
#include <iostream>
#include <map>
#include <string>
#include <vector>
// CImg, the superior grafiks library
#include "tiv_lib.h"
// This #define tells CImg that we use the library without any display options
// -- just for loading images.
#define cimg_display 0
#include "CImg.h"
// First include for detecting console output size,
// everything else for exit codes
#ifdef _POSIX_VERSION
// Console output size detection
#include <sys/ioctl.h>
#include <sysexits.h>
// Error explanation, for some reason
#include <cstring>
#endif
#ifdef _WIN32
#include <windows.h>
// Following codes copied from /usr/include/sysexits.h,
// license: https://opensource.org/license/BSD-3-clause/
#define EX_OK 0 /* successful termination */
#define EX__BASE 64 /* base value for error messages */
#define EX_USAGE 64 /* command line usage error */
#define EX_DATAERR 65 /* data format error */
#define EX_NOINPUT 66 /* cannot open input */
#define EX_SOFTWARE 70 /* internal software error */
#define EX_CANTCREAT 73 /* can't create (user) output file */
#define EX_IOERR 74 /* input/output error */
#define EX_TEMPFAIL 75 /* temp failure; user is invited to retry */
#define EX_NOPERM 77 /* permission denied */
#define EX_CONFIG 78 /* configuration error */
// Error explanation
#include <system_error>
#endif
// using namespace std; // haha nope, bad style
// especially when we're also using the CImg namespace
// Program exit code constants compatible with sysexits.h.
#define EXITCODE_OK 0
#define EXITCODE_COMMAND_LINE_USAGE_ERROR 64
#define EXITCODE_DATA_FORMAT_ERROR 65
#define EXITCODE_NO_INPUT_ERROR 66
// Implementation of flag representation for flags in the main() method
constexpr int FLAG_FG = 1;
constexpr int FLAG_BG = 2;
constexpr int FLAG_MODE_256 = 4;
constexpr int FLAG_24BIT = 8;
constexpr int FLAG_NOOPT = 16;
constexpr int FLAG_TELETEXT = 32;
// Steps (@TODO: Figure out what exactly they represent)
constexpr int COLOR_STEP_COUNT = 6;
constexpr int COLOR_STEPS[COLOR_STEP_COUNT] = {0, 0x5f, 0x87, 0xaf, 0xd7, 0xff};
constexpr int GRAYSCALE_STEP_COUNT = 24;
constexpr int GRAYSCALE_STEPS[GRAYSCALE_STEP_COUNT] = {
0x08, 0x12, 0x1c, 0x26, 0x30, 0x3a, 0x44, 0x4e, 0x58, 0x62, 0x6c, 0x76,
0x80, 0x8a, 0x94, 0x9e, 0xa8, 0xb2, 0xbc, 0xc6, 0xd0, 0xda, 0xe4, 0xee};
constexpr unsigned int BITMAPS[] = {
0x00000000, 0x00a0,
// Block graphics
// 0xffff0000, 0x2580, // upper 1/2; redundant with inverse lower 1/2
0x0000000f, 0x2581, // lower 1/8
0x000000ff, 0x2582, // lower 1/4
0x00000fff, 0x2583, 0x0000ffff, 0x2584, // lower 1/2
0x000fffff, 0x2585, 0x00ffffff, 0x2586, // lower 3/4
0x0fffffff, 0x2587,
// 0xffffffff, 0x2588, // full; redundant with inverse space
0xeeeeeeee, 0x258a, // left 3/4
0xcccccccc, 0x258c, // left 1/2
0x88888888, 0x258e, // left 1/4
0x0000cccc, 0x2596, // quadrant lower left
0x00003333, 0x2597, // quadrant lower right
0xcccc0000,
0x2598, // quadrant upper left
// 0xccccffff, 0x2599, // 3/4 redundant with inverse 1/4
0xcccc3333, 0x259a, // diagonal 1/2
// 0xffffcccc, 0x259b, // 3/4 redundant
// 0xffff3333, 0x259c, // 3/4 redundant
0x33330000, 0x259d, // quadrant upper right
// 0x3333cccc, 0x259e, // 3/4 redundant
// 0x3333ffff, 0x259f, // 3/4 redundant
// Line drawing subset: no double lines, no complex light lines
0x000ff000, 0x2501, // Heavy horizontal
0x66666666, 0x2503, // Heavy vertical
0x00077666, 0x250f, // Heavy down and right
0x000ee666, 0x2513, // Heavy down and left
0x66677000, 0x2517, // Heavy up and right
0x666ee000, 0x251b, // Heavy up and left
0x66677666, 0x2523, // Heavy vertical and right
0x666ee666, 0x252b, // Heavy vertical and left
0x000ff666, 0x2533, // Heavy down and horizontal
0x666ff000, 0x253b, // Heavy up and horizontal
0x666ff666, 0x254b, // Heavy cross
0x000cc000, 0x2578, // Bold horizontal left
0x00066000, 0x2579, // Bold horizontal up
0x00033000, 0x257a, // Bold horizontal right
0x00066000, 0x257b, // Bold horizontal down
0x06600660, 0x254f, // Heavy double dash vertical
0x000f0000, 0x2500, // Light horizontal
0x0000f000, 0x2500, //
0x44444444, 0x2502, // Light vertical
0x22222222, 0x2502,
0x000e0000, 0x2574, // light left
0x0000e000, 0x2574, // light left
0x44440000, 0x2575, // light up
0x22220000, 0x2575, // light up
0x00030000, 0x2576, // light right
0x00003000, 0x2576, // light right
0x00004444, 0x2577, // light down
0x00002222, 0x2577, // light down
// Misc technical
0x44444444, 0x23a2, // [ extension
0x22222222, 0x23a5, // ] extension
0x0f000000, 0x23ba, // Horizontal scanline 1
0x00f00000, 0x23bb, // Horizontal scanline 3
0x00000f00, 0x23bc, // Horizontal scanline 7
0x000000f0, 0x23bd, // Horizontal scanline 9
// Geometrical shapes. Tricky because some of them are too wide.
// 0x00ffff00, 0x25fe, // Black medium small square
0x00066000, 0x25aa, // Black small square
// 0x11224488, 0x2571, // diagonals
// 0x88442211, 0x2572,
// 0x99666699, 0x2573,
// 0x000137f0, 0x25e2, // Triangles
// 0x0008cef0, 0x25e3,
// 0x000fec80, 0x25e4,
// 0x000f7310, 0x25e5,
0, 0, // End marker for "regular" characters
// Teletext / legacy graphics 3x2 block character codes.
// Using a 3-2-3 pattern consistently, perhaps we should create automatic
// variations....
0xccc00000, 0xfb00, 0x33300000, 0xfb01, 0xfff00000, 0xfb02, 0x000cc000,
0xfb03, 0xccccc000, 0xfb04, 0x333cc000, 0xfb05, 0xfffcc000, 0xfb06,
0x00033000, 0xfb07, 0xccc33000, 0xfb08, 0x33333000, 0xfb09, 0xfff33000,
0xfb0a, 0x000ff000, 0xfb0b, 0xcccff000, 0xfb0c, 0x333ff000, 0xfb0d,
0xfffff000, 0xfb0e, 0x00000ccc, 0xfb0f,
0xccc00ccc, 0xfb10, 0x33300ccc, 0xfb11, 0xfff00ccc, 0xfb12, 0x000ccccc,
0xfb13, 0x333ccccc, 0xfb14, 0xfffccccc, 0xfb15, 0x00033ccc, 0xfb16,
0xccc33ccc, 0xfb17, 0x33333ccc, 0xfb18, 0xfff33ccc, 0xfb19, 0x000ffccc,
0xfb1a, 0xcccffccc, 0xfb1b, 0x333ffccc, 0xfb1c, 0xfffffccc, 0xfb1d,
0x00000333, 0xfb1e, 0xccc00333, 0xfb1f,
0x33300333, 0x1b20, 0xfff00333, 0x1b21, 0x000cc333, 0x1b22, 0xccccc333,
0x1b23, 0x333cc333, 0x1b24, 0xfffcc333, 0x1b25, 0x00033333, 0x1b26,
0xccc33333, 0x1b27, 0xfff33333, 0x1b28, 0x000ff333, 0x1b29, 0xcccff333,
0x1b2a, 0x333ff333, 0x1b2b, 0xfffff333, 0x1b2c, 0x00000fff, 0x1b2d,
0xccc00fff, 0x1b2e, 0x33300fff, 0x1b2f,
0xfff00fff, 0x1b30, 0x000ccfff, 0x1b31, 0xcccccfff, 0x1b32, 0x333ccfff,
0x1b33, 0xfffccfff, 0x1b34, 0x00033fff, 0x1b35, 0xccc33fff, 0x1b36,
0x33333fff, 0x1b37, 0xfff33fff, 0x1b38, 0x000fffff, 0x1b39, 0xcccfffff,
0x1b3a, 0x333fffff, 0x1b3b,
0, 1 // End marker for extended TELETEXT mode.
};
struct CharData {
std::array<int, 3> fgColor = std::array<int, 3>{0, 0, 0};
std::array<int, 3> bgColor = std::array<int, 3>{0, 0, 0};
int codePoint;
};
// Return a CharData struct with the given code point and corresponding averag
// fg and bg colors.
CharData createCharData(const cimg_library::CImg<unsigned char> &image, int x0,
int y0, int codepoint, int pattern) {
CharData result;
result.codePoint = codepoint;
int fg_count = 0;
int bg_count = 0;
unsigned int mask = 0x80000000;
for (int y = 0; y < 8; y++) {
for (int x = 0; x < 4; x++) {
int *avg;
if (pattern & mask) {
avg = result.fgColor.data();
fg_count++;
} else {
avg = result.bgColor.data();
bg_count++;
}
for (int i = 0; i < 3; i++) {
avg[i] += image(x0 + x, y0 + y, 0, i);
}
mask = mask >> 1;
}
}
// Calculate the average color value for each bucket
for (int i = 0; i < 3; i++) {
if (bg_count != 0) {
result.bgColor[i] /= bg_count;
}
if (fg_count != 0) {
result.fgColor[i] /= fg_count;
}
}
return result;
}
/**
* @brief Find the best character and colors
* for a 4x8 part of the image at the given position
*
* @param image
* @param x0
* @param y0
* @param flags
* @return CharData
*/
CharData findCharData(const cimg_library::CImg<unsigned char> &image, int x0,
int y0, const int &flags) {
int min[3] = {255, 255, 255};
int max[3] = {0};
std::map<long, int> count_per_color;
// Determine the minimum and maximum value for each color channel
for (int y = 0; y < 8; y++) {
for (int x = 0; x < 4; x++) {
long color = 0;
for (int i = 0; i < 3; i++) {
int d = image(x0 + x, y0 + y, 0, i);
min[i] = std::min(min[i], d);
max[i] = std::max(max[i], d);
color = (color << 8) | d;
}
count_per_color[color]++;
}
}
std::multimap<int, long> color_per_count;
for (auto i = count_per_color.begin(); i != count_per_color.end(); ++i) {
color_per_count.insert(std::pair<int, long>(i->second, i->first));
}
auto iter = color_per_count.rbegin();
int count2 = iter->first;
long max_count_color_1 = iter->second;
long max_count_color_2 = max_count_color_1;
if ((++iter) != color_per_count.rend()) {
count2 += iter->first;
max_count_color_2 = iter->second;
}
unsigned int bits = 0;
bool direct = count2 > (8 * 4) / 2;
if (direct) {
for (int y = 0; y < 8; y++) {
for (int x = 0; x < 4; x++) {
bits = bits << 1;
int d1 = 0;
int d2 = 0;
for (int i = 0; i < 3; i++) {
int shift = 16 - 8 * i;
int c1 = (max_count_color_1 >> shift) & 255;
int c2 = (max_count_color_2 >> shift) & 255;
int c = image(x0 + x, y0 + y, 0, i);
d1 += (c1 - c) * (c1 - c);
d2 += (c2 - c) * (c2 - c);
}
if (d1 > d2) {
bits |= 1;
}
}
}
} else {
// Determine the color channel with the greatest range.
int splitIndex = 0;
int bestSplit = 0;
for (int i = 0; i < 3; i++) {
if (max[i] - min[i] > bestSplit) {
bestSplit = max[i] - min[i];
splitIndex = i;
}
}
// We just split at the middle of the interval instead of computing the
// median.
int splitValue = min[splitIndex] + bestSplit / 2;
// Compute a bitmap using the given split and sum the color values for
// both buckets.
for (int y = 0; y < 8; y++) {
for (int x = 0; x < 4; x++) {
bits = bits << 1;
if (image(x0 + x, y0 + y, 0, splitIndex) > splitValue) {
bits |= 1;
}
}
}
}
// Find the best bitmap match by counting the bits that don't match,
// including the inverted bitmaps.
int best_diff = 8;
unsigned int best_pattern = 0x0000ffff;
int codepoint = 0x2584;
bool inverted = false;
unsigned int end_marker = flags & FLAG_TELETEXT ? 1 : 0;
for (int i = 0; BITMAPS[i + 1] != end_marker; i += 2) {
// Skip all end markers
if (BITMAPS[i + 1] < 32) {
continue;
}
unsigned int pattern = BITMAPS[i];
for (int j = 0; j < 2; j++) {
int diff = (std::bitset<32>(pattern ^ bits)).count();
if (diff < best_diff) {
best_pattern = BITMAPS[i]; // pattern might be inverted.
codepoint = BITMAPS[i + 1];
best_diff = diff;
inverted = best_pattern != pattern;
}
pattern = ~pattern;
}
}
if (direct) {
CharData result;
if (inverted) {
long tmp = max_count_color_1;
max_count_color_1 = max_count_color_2;
max_count_color_2 = tmp;
}
for (int i = 0; i < 3; i++) {
int shift = 16 - 8 * i;
result.fgColor[i] = (max_count_color_2 >> shift) & 255;
result.bgColor[i] = (max_count_color_1 >> shift) & 255;
result.codePoint = codepoint;
}
return result;
}
return createCharData(image, x0, y0, codepoint, best_pattern);
}
int clamp_byte(int value) {
return value < 0 ? 0 : (value > 255 ? 255 : value);
}
double sqr(double n) { return n * n; }
int best_index(int value, const int (&STEPS)[], int count) {
int best_diff = std::abs(STEPS[0] - value);
int result = 0;
for (int i = 1; i < count; i++) {
int diff = std::abs(STEPS[i] - value);
if (diff < best_diff) {
result = i;
best_diff = diff;
}
}
return result;
}
void emit_color(const int &flags, int r, int g, int b) {
void printTermColor(const int &flags, int r, int g, int b) {
r = clamp_byte(r);
g = clamp_byte(g);
b = clamp_byte(b);
@ -449,7 +104,7 @@ void emit_color(const int &flags, int r, int g, int b) {
std::cout << (bg ? "\x1B[48;5;" : "\u001B[38;5;") << color_index << "m";
}
void emitCodepoint(int codepoint) {
void printCodepoint(int codepoint) {
if (codepoint < 128) {
std::cout << static_cast<char>(codepoint);
} else if (codepoint < 0x7ff) {
@ -469,22 +124,28 @@ void emitCodepoint(int codepoint) {
}
}
void emit_image(const cimg_library::CImg<unsigned char> &image,
void printImage(const cimg_library::CImg<unsigned char> &image,
const int &flags) {
GetPixelFunction get_pixel = [&](int x, int y) -> unsigned long {
return (((unsigned long) image(x, y, 0, 0)) << 16)
| (((unsigned long) image(x, y, 0, 1)) << 8)
| (((unsigned long) image(x, y, 0, 2)));
};
CharData lastCharData;
for (int y = 0; y <= image.height() - 8; y += 8) {
for (int x = 0; x <= image.width() - 4; x += 4) {
CharData charData =
flags & FLAG_NOOPT
? createCharData(image, x, y, 0x2584, 0x0000ffff)
: findCharData(image, x, y, flags);
? createCharData(get_pixel, x, y, 0x2584, 0x0000ffff)
: findCharData(get_pixel, x, y, flags);
if (x == 0 || charData.bgColor != lastCharData.bgColor)
emit_color(flags | FLAG_BG, charData.bgColor[0],
charData.bgColor[1], charData.bgColor[2]);
printTermColor(flags | FLAG_BG, charData.bgColor[0],
charData.bgColor[1], charData.bgColor[2]);
if (x == 0 || charData.fgColor != lastCharData.fgColor)
emit_color(flags | FLAG_FG, charData.fgColor[0],
charData.fgColor[1], charData.fgColor[2]);
emitCodepoint(charData.codePoint);
printTermColor(flags | FLAG_FG, charData.fgColor[0],
charData.fgColor[1], charData.fgColor[2]);
printCodepoint(charData.codePoint);
lastCharData = charData;
}
std::cout << "\x1b[0m" << std::endl;
@ -532,14 +193,14 @@ cimg_library::CImg<unsigned char> load_rgb_CImg(const char *const &filename) {
}
// Implements --help
void emit_usage() {
void printUsage() {
std::cerr << R"(
Terminal Image Viewer v1.2
Terminal Image Viewer v1.2.1
usage: tiv [options] <image> [<image>...]
-0 : No block character adjustment, always use top half block char.
-2, --256 : Use 256-bit colors. Needed to display properly on macOS Terminal.
-c <num> : Number of thumbnail columns in 'dir' mode (3 by default).
-d, --dir : Force 'dir' mode. Automatially selected for more than one input.
-d, --dir : Force 'dir' mode. Automatically selected for more than one input.
-f, --full: Force 'full' mode. Automatically selected for one input.
--help : Display this help text.
-h <num> : Set the maximum output height to <num> lines.
@ -552,42 +213,12 @@ enum Mode { AUTO, THUMBNAILS, FULL_SIZE };
int main(int argc, char *argv[]) {
std::ios::sync_with_stdio(false); // apparently makes printing faster
bool detectSize = true;
// Platform-specific implementations for determining console size, better
// implementations are welcome Fallback sizes when unsuccesful
// implementations are welcome Fallback sizes when unsuccessful
int maxWidth = 80;
int maxHeight = 24;
#ifdef __POSIX_VERSION
struct winsize w;
// If redirecting STDOUT to one file ( col or row == 0, or the previous
// ioctl call's failed )
if (ioctl(STDOUT_FILENO, TIOCGWINSZ, &w) != 0 ||
(w.ws_col | w.ws_row) == 0) {
std::cerr << "Warning: failed to determine most reasonable size, "
"defaulting to 80x24"
<< std::endl;
} else {
maxWidth = w.ws_col * 4;
maxHeight = w.ws_row * 8;
}
#elif defined _WIN32
CONSOLE_SCREEN_BUFFER_INFO w;
if (GetConsoleScreenBufferInfo(
GetStdHandle(STD_OUTPUT_HANDLE),
&w)) { // just like powershell, but without the hyphens, hooray
maxWidth = w.dwSize.X * 4;
maxHeight = w.dwSize.Y * 8;
} else {
std::cerr << GetLastError();
std::cerr << "Warning: failed to determine most reasonable size, "
"defaulting to 80x24"
<< std::endl;
}
#else
std::cerr << "Warning: failed to determine most reasonable sie: "
"unrecognized system, defaulting to 80x24"
<< std::endl;
#endif
// Reading input
char flags = 0; // bitwise representation of flags,
@ -596,11 +227,11 @@ int main(int argc, char *argv[]) {
int columns = 3;
std::vector<std::string> file_names;
int ret = EX_OK; // The return code for the program
int ret = EXITCODE_OK; // The return code for the program
if (argc <= 1) {
emit_usage();
return 0;
printUsage();
return EXITCODE_COMMAND_LINE_USAGE_ERROR;
}
for (int i = 1; i < argc; i++) {
@ -612,7 +243,7 @@ int main(int argc, char *argv[]) {
columns = std::stoi(argv[++i]);
} else {
std::cerr << "Error: -c requires a number" << std::endl;
ret = EX_USAGE;
ret = EXITCODE_COMMAND_LINE_USAGE_ERROR;
}
} else if (arg == "-d" || arg == "--dir") {
mode = THUMBNAILS;
@ -620,25 +251,25 @@ int main(int argc, char *argv[]) {
mode = FULL_SIZE;
} else if (arg == "-w") {
if (i < argc - 1) {
maxWidth = 4 * std::stoi(argv[++i]);
maxWidth = 4 * std::stoi(argv[++i]), detectSize = false;
} else {
std::cerr << "Error: -w requires a number" << std::endl;
ret = EX_USAGE;
ret = EXITCODE_COMMAND_LINE_USAGE_ERROR;
}
} else if (arg == "-h") {
if (i < argc - 1)
maxHeight = 8 * std::stoi(argv[++i]);
maxHeight = 8 * std::stoi(argv[++i]), detectSize = false;
else
emit_usage();
printUsage(); // people might confuse this with help
} else if (arg == "--256" || arg == "-2" || arg == "-256") {
flags |= FLAG_MODE_256;
} else if (arg == "--help" || arg == "-help") {
emit_usage();
printUsage();
} else if (arg == "-x") {
flags |= FLAG_TELETEXT;
} else if (arg[0] == '-') {
std::cerr << "Error: Unrecognized argument: " << arg << std::endl;
ret = EX_USAGE;
ret = EXITCODE_COMMAND_LINE_USAGE_ERROR;
} else {
// Arguments that will be displayed
if (std::filesystem::is_directory(arg)) {
@ -646,19 +277,51 @@ int main(int argc, char *argv[]) {
if (std::filesystem::is_regular_file(p.path()))
file_names.push_back(p.path().string());
} else {
// Check if file can be opened
// Check if file can be opened, @TODO find better way
std::ifstream fin(arg.c_str());
if (fin) {
file_names.push_back(arg);
} else {
std::cerr << "Error: Cannot open '" << arg
<< "', permission issue?" << std::endl;
ret = EX_NOINPUT;
ret = EXITCODE_NO_INPUT_ERROR;
}
}
}
}
if (detectSize) {
#ifdef _POSIX_VERSION
struct winsize w;
// If redirecting STDOUT to one file ( col or row == 0, or the previous
// ioctl call's failed )
if (ioctl(STDOUT_FILENO, TIOCGWINSZ, &w) != 0 ||
(w.ws_col | w.ws_row) == 0) {
std::cerr << "Warning: failed to determine most reasonable size: "
<< strerror(errno) << ", defaulting to 20x6" << std::endl;
} else {
maxWidth = w.ws_col * 4;
maxHeight = w.ws_row * 8;
}
#elif defined _WIN32
CONSOLE_SCREEN_BUFFER_INFO w;
if (GetConsoleScreenBufferInfo(
GetStdHandle(STD_OUTPUT_HANDLE),
&w)) { // just like PowerShell, but without the hyphens, hooray
maxWidth = w.dwSize.X * 4;
maxHeight = w.dwSize.Y * 8;
} else {
std::cerr << "Warning: failed to determine most reasonable size: "
<< std::system_category().message(GetLastError())
<< ", defaulting to 80x24" << std::endl;
}
#else
std::cerr << "Warning: failed to determine most reasonable size: "
"unrecognized system, defaulting to 80x24"
<< std::endl;
#endif
}
if (mode == FULL_SIZE || (mode == AUTO && file_names.size() == 1)) {
for (const auto &filename : file_names) {
try {
@ -671,12 +334,12 @@ int main(int argc, char *argv[]) {
image.resize(new_size.width, new_size.height, -100, -100,
5);
}
// the acutal magic which generates the output
emit_image(image, flags);
// the actual magic which generates the output
printImage(image, flags);
} catch (cimg_library::CImgIOException &e) {
std::cerr << "Error: '" << filename
<< "' has an unrecognized file format" << std::endl;
ret = EX_DATAERR;
ret = EXITCODE_DATA_FORMAT_ERROR;
}
}
} else { // Thumbnail mode
@ -712,7 +375,7 @@ int main(int argc, char *argv[]) {
// Probably no image; ignore.
}
}
if (count) emit_image(image, flags);
if (count) printImage(image, flags);
std::cout << sb << std::endl << std::endl;
}
}

397
src/tiv_lib.cpp Normal file
View file

@ -0,0 +1,397 @@
/*
* Copyright (c) 2017-2023, Stefan Haustein, Aaron Liu
*
* This file is free software: you may copy, redistribute and/or modify it
* under the terms of the GNU General Public License as published by the
* Free Software Foundation, either version 3 of the License, or (at your
* option) any later version.
*
* This file is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <https://www.gnu.org/licenses/>.
*
* Alternatively, you may copy, redistribute and/or modify this file under
* the terms of the Apache License, version 2.0:
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* https://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "tiv_lib.h"
#include <array>
#include <bitset>
#include <cmath>
#include <filesystem>
#include <fstream>
#include <functional>
#include <iostream>
#include <map>
#include <string>
#include <vector>
const int END_MARKER = 0;
// An interleaved map of 4x8 bit character bitmaps (each hex digit represents a
// row) to the corresponding Unicode character code point.
constexpr unsigned int BITMAPS[] = {
0x00000000, 0x00a0, 0,
// Block graphics
// 0xffff0000, 0x2580, 0, // upper 1/2; redundant with inverse lower 1/2
0x0000000f, 0x2581, 0, // lower 1/8
0x000000ff, 0x2582, 0, // lower 1/4
0x00000fff, 0x2583, 0,
0x0000ffff, 0x2584, 0, // lower 1/2
0x000fffff, 0x2585, 0,
0x00ffffff, 0x2586, 0, // lower 3/4
0x0fffffff, 0x2587, 0,
// 0xffffffff, 0x2588, // full; redundant with inverse space
0xeeeeeeee, 0x258a, 0, // left 3/4
0xcccccccc, 0x258c, 0, // left 1/2
0x88888888, 0x258e, 0, // left 1/4
0x0000cccc, 0x2596, 0, // quadrant lower left
0x00003333, 0x2597, 0, // quadrant lower right
0xcccc0000, 0x2598, 0, // quadrant upper left
// 0xccccffff, 0x2599, // 3/4 redundant with inverse 1/4
0xcccc3333, 0x259a, 0, // diagonal 1/2
// 0xffffcccc, 0x259b, // 3/4 redundant
// 0xffff3333, 0x259c, // 3/4 redundant
0x33330000, 0x259d, 0, // quadrant upper right
// 0x3333cccc, 0x259e, // 3/4 redundant
// 0x3333ffff, 0x259f, // 3/4 redundant
// Line drawing subset: no double lines, no complex light lines
0x000ff000, 0x2501, 0, // Heavy horizontal
0x66666666, 0x2503, 0, // Heavy vertical
0x00077666, 0x250f, 0, // Heavy down and right
0x000ee666, 0x2513, 0, // Heavy down and left
0x66677000, 0x2517, 0, // Heavy up and right
0x666ee000, 0x251b, 0, // Heavy up and left
0x66677666, 0x2523, 0, // Heavy vertical and right
0x666ee666, 0x252b, 0, // Heavy vertical and left
0x000ff666, 0x2533, 0, // Heavy down and horizontal
0x666ff000, 0x253b, 0, // Heavy up and horizontal
0x666ff666, 0x254b, 0, // Heavy cross
0x000cc000, 0x2578, 0, // Bold horizontal left
0x00066000, 0x2579, 0, // Bold horizontal up
0x00033000, 0x257a, 0, // Bold horizontal right
0x00066000, 0x257b, 0, // Bold horizontal down
0x06600660, 0x254f, 0, // Heavy double dash vertical
0x000f0000, 0x2500, 0, // Light horizontal
0x0000f000, 0x2500, 0, //
0x44444444, 0x2502, 0, // Light vertical
0x22222222, 0x2502, 0,
0x000e0000, 0x2574, 0, // light left
0x0000e000, 0x2574, 0, // light left
0x44440000, 0x2575, 0, // light up
0x22220000, 0x2575, 0, // light up
0x00030000, 0x2576, 0, // light right
0x00003000, 0x2576, 0, // light right
0x00004444, 0x2577, 0, // light down
0x00002222, 0x2577, 0, // light down
// Misc technical
0x44444444, 0x23a2, 0, // [ extension
0x22222222, 0x23a5, 0, // ] extension
0x0f000000, 0x23ba, 0, // Horizontal scanline 1
0x00f00000, 0x23bb, 0, // Horizontal scanline 3
0x00000f00, 0x23bc, 0, // Horizontal scanline 7
0x000000f0, 0x23bd, 0, // Horizontal scanline 9
// Geometrical shapes. Tricky because some of them are too wide.
// 0x00ffff00, 0x25fe, 0, // Black medium small square
0x00066000, 0x25aa, 0, // Black small square
// 0x11224488, 0x2571, 0, // diagonals
// 0x88442211, 0x2572, 0,
// 0x99666699, 0x2573, 0,
// 0x000137f0, 0x25e2, 0, // Triangles
// 0x0008cef0, 0x25e3, 0,
// 0x000fec80, 0x25e4, 0,
// 0x000f7310, 0x25e5, 0,
// Teletext / legacy graphics 3x2 block character codes.
// Using a 3-2-3 pattern consistently, perhaps we should create automatic
// variations....
0xccc00000, 0xfb00, FLAG_TELETEXT,
0x33300000, 0xfb01, FLAG_TELETEXT,
0xfff00000, 0xfb02, FLAG_TELETEXT,
0x000cc000, 0xfb03, FLAG_TELETEXT,
0xccccc000, 0xfb04, FLAG_TELETEXT,
0x333cc000, 0xfb05, FLAG_TELETEXT,
0xfffcc000, 0xfb06, FLAG_TELETEXT,
0x00033000, 0xfb07, FLAG_TELETEXT,
0xccc33000, 0xfb08, FLAG_TELETEXT,
0x33333000, 0xfb09, FLAG_TELETEXT,
0xfff33000, 0xfb0a, FLAG_TELETEXT,
0x000ff000, 0xfb0b, FLAG_TELETEXT,
0xcccff000, 0xfb0c, FLAG_TELETEXT,
0x333ff000, 0xfb0d, FLAG_TELETEXT,
0xfffff000, 0xfb0e, FLAG_TELETEXT,
0x00000ccc, 0xfb0f, FLAG_TELETEXT,
0xccc00ccc, 0xfb10, FLAG_TELETEXT,
0x33300ccc, 0xfb11, FLAG_TELETEXT,
0xfff00ccc, 0xfb12, FLAG_TELETEXT,
0x000ccccc, 0xfb13, FLAG_TELETEXT,
0x333ccccc, 0xfb14, FLAG_TELETEXT,
0xfffccccc, 0xfb15, FLAG_TELETEXT,
0x00033ccc, 0xfb16, FLAG_TELETEXT,
0xccc33ccc, 0xfb17, FLAG_TELETEXT,
0x33333ccc, 0xfb18, FLAG_TELETEXT,
0xfff33ccc, 0xfb19, FLAG_TELETEXT,
0x000ffccc, 0xfb1a, FLAG_TELETEXT,
0xcccffccc, 0xfb1b, FLAG_TELETEXT,
0x333ffccc, 0xfb1c, FLAG_TELETEXT,
0xfffffccc, 0xfb1d, FLAG_TELETEXT,
0x00000333, 0xfb1e, FLAG_TELETEXT,
0xccc00333, 0xfb1f, FLAG_TELETEXT,
0x33300333, 0x1b20, FLAG_TELETEXT,
0xfff00333, 0x1b21, FLAG_TELETEXT,
0x000cc333, 0x1b22, FLAG_TELETEXT,
0xccccc333, 0x1b23, FLAG_TELETEXT,
0x333cc333, 0x1b24, FLAG_TELETEXT,
0xfffcc333, 0x1b25, FLAG_TELETEXT,
0x00033333, 0x1b26, FLAG_TELETEXT,
0xccc33333, 0x1b27, FLAG_TELETEXT,
0xfff33333, 0x1b28, FLAG_TELETEXT,
0x000ff333, 0x1b29, FLAG_TELETEXT,
0xcccff333, 0x1b2a, FLAG_TELETEXT,
0x333ff333, 0x1b2b, FLAG_TELETEXT,
0xfffff333, 0x1b2c, FLAG_TELETEXT,
0x00000fff, 0x1b2d, FLAG_TELETEXT,
0xccc00fff, 0x1b2e, FLAG_TELETEXT,
0x33300fff, 0x1b2f, FLAG_TELETEXT,
0xfff00fff, 0x1b30, FLAG_TELETEXT,
0x000ccfff, 0x1b31, FLAG_TELETEXT,
0xcccccfff, 0x1b32, FLAG_TELETEXT,
0x333ccfff, 0x1b33, FLAG_TELETEXT,
0xfffccfff, 0x1b34, FLAG_TELETEXT,
0x00033fff, 0x1b35, FLAG_TELETEXT,
0xccc33fff, 0x1b36, FLAG_TELETEXT,
0x33333fff, 0x1b37, FLAG_TELETEXT,
0xfff33fff, 0x1b38, FLAG_TELETEXT,
0x000fffff, 0x1b39, FLAG_TELETEXT,
0xcccfffff, 0x1b3a, FLAG_TELETEXT,
0x333fffff, 0x1b3b, FLAG_TELETEXT,
0, END_MARKER, 0 // End marker
};
// The channel indices are 0, 1, 2 for R, G, B
unsigned char get_channel(unsigned long rgb, int index) {
return (unsigned char) ((rgb >> ((2 - index) * 8)) & 255);
}
CharData createCharData(GetPixelFunction get_pixel, int x0, int y0,
int codepoint, int pattern) {
CharData result;
result.codePoint = codepoint;
int fg_count = 0;
int bg_count = 0;
unsigned int mask = 0x80000000;
for (int y = 0; y < 8; y++) {
for (int x = 0; x < 4; x++) {
int *avg;
if (pattern & mask) {
avg = result.fgColor.data();
fg_count++;
} else {
avg = result.bgColor.data();
bg_count++;
}
long rgb = get_pixel(x0 + x, y0 + y);
for (int i = 0; i < 3; i++) {
avg[i] += get_channel(rgb, i);
}
mask = mask >> 1;
}
}
// Calculate the average color value for each bucket
for (int i = 0; i < 3; i++) {
if (bg_count != 0) {
result.bgColor[i] /= bg_count;
}
if (fg_count != 0) {
result.fgColor[i] /= fg_count;
}
}
return result;
}
CharData findCharData(GetPixelFunction get_pixel, int x0, int y0,
const int &flags) {
int min[3] = {255, 255, 255};
int max[3] = {0};
std::map<long, int> count_per_color;
// Determine the minimum and maximum value for each color channel
for (int y = 0; y < 8; y++) {
for (int x = 0; x < 4; x++) {
long color = 0;
long rgb = get_pixel(x0 + x, y0 + y);
for (int i = 0; i < 3; i++) {
int d = get_channel(rgb, i);
min[i] = std::min(min[i], d);
max[i] = std::max(max[i], d);
color = (color << 8) | d;
}
count_per_color[color]++;
}
}
std::multimap<int, long> color_per_count;
for (auto i = count_per_color.begin(); i != count_per_color.end(); ++i) {
color_per_count.insert(std::pair<int, long>(i->second, i->first));
}
auto iter = color_per_count.rbegin();
int count2 = iter->first;
long max_count_color_1 = iter->second;
long max_count_color_2 = max_count_color_1;
if ((++iter) != color_per_count.rend()) {
count2 += iter->first;
max_count_color_2 = iter->second;
}
unsigned int bits = 0;
bool direct = count2 > (8 * 4) / 2;
if (direct) {
for (int y = 0; y < 8; y++) {
for (int x = 0; x < 4; x++) {
bits = bits << 1;
int d1 = 0;
int d2 = 0;
unsigned long rgb = get_pixel(x0 + x, y0 + y);
for (int i = 0; i < 3; i++) {
int shift = 16 - 8 * i;
int c1 = (max_count_color_1 >> shift) & 255;
int c2 = (max_count_color_2 >> shift) & 255;
int c = get_channel(rgb, i);
d1 += (c1 - c) * (c1 - c);
d2 += (c2 - c) * (c2 - c);
}
if (d1 > d2) {
bits |= 1;
}
}
}
} else {
// Determine the color channel with the greatest range.
int splitIndex = 0;
int bestSplit = 0;
for (int i = 0; i < 3; i++) {
if (max[i] - min[i] > bestSplit) {
bestSplit = max[i] - min[i];
splitIndex = i;
}
}
// We just split at the middle of the interval instead of computing the
// median.
int splitValue = min[splitIndex] + bestSplit / 2;
// Compute a bitmap using the given split and sum the color values for
// both buckets.
for (int y = 0; y < 8; y++) {
for (int x = 0; x < 4; x++) {
bits = bits << 1;
if (get_channel(get_pixel(x0 + x, y0 + y), splitIndex) > splitValue) {
bits |= 1;
}
}
}
}
// Find the best bitmap match by counting the bits that don't match,
// including the inverted bitmaps.
int best_diff = 8;
unsigned int best_pattern = 0x0000ffff;
int codepoint = 0x2584;
bool inverted = false;
for (int i = 0; BITMAPS[i + 1] != END_MARKER; i += 3) {
if ((BITMAPS[i + 2] & flags) != BITMAPS[i + 2]) {
continue;
}
unsigned int pattern = BITMAPS[i];
for (int j = 0; j < 2; j++) {
int diff = (std::bitset<32>(pattern ^ bits)).count();
if (diff < best_diff) {
best_pattern = BITMAPS[i]; // pattern might be inverted.
codepoint = BITMAPS[i + 1];
best_diff = diff;
inverted = best_pattern != pattern;
}
pattern = ~pattern;
}
}
if (direct) {
CharData result;
if (inverted) {
long tmp = max_count_color_1;
max_count_color_1 = max_count_color_2;
max_count_color_2 = tmp;
}
for (int i = 0; i < 3; i++) {
int shift = 16 - 8 * i;
result.fgColor[i] = (max_count_color_2 >> shift) & 255;
result.bgColor[i] = (max_count_color_1 >> shift) & 255;
result.codePoint = codepoint;
}
return result;
}
return createCharData(get_pixel, x0, y0, codepoint, best_pattern);
}
int clamp_byte(int value) {
return value < 0 ? 0 : (value > 255 ? 255 : value);
}
double sqr(double n) { return n * n; }
int best_index(int value, const int STEPS[], int count) {
int best_diff = std::abs(STEPS[0] - value);
int result = 0;
for (int i = 1; i < count; i++) {
int diff = std::abs(STEPS[i] - value);
if (diff < best_diff) {
result = i;
best_diff = diff;
}
}
return result;
}

99
src/tiv_lib.h Normal file
View file

@ -0,0 +1,99 @@
/*
* Copyright (c) 2017-2023, Stefan Haustein, Aaron Liu
*
* This file is free software: you may copy, redistribute and/or modify it
* under the terms of the GNU General Public License as published by the
* Free Software Foundation, either version 3 of the License, or (at your
* option) any later version.
*
* This file is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program. If not, see <https://www.gnu.org/licenses/>.
*
* Alternatively, you may copy, redistribute and/or modify this file under
* the terms of the Apache License, version 2.0:
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* https://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#ifndef TIV_LIB_H
#define TIV_LIB_H
#include <array>
#include <functional>
// Implementation of flag representation for flags in the main() method
constexpr int FLAG_FG = 1;
constexpr int FLAG_BG = 2;
constexpr int FLAG_MODE_256 = 4; // Limit colors to 256-color mode
constexpr int FLAG_24BIT = 8; // 24-bit color mode
constexpr int FLAG_NOOPT = 16; // Only use the same half-block character
constexpr int FLAG_TELETEXT = 32; // Use teletext characters
// Color saturation value steps from 0 to 255
constexpr int COLOR_STEP_COUNT = 6;
constexpr int COLOR_STEPS[COLOR_STEP_COUNT] = {0, 0x5f, 0x87, 0xaf, 0xd7, 0xff};
// Grayscale saturation value steps from 0 to 255
constexpr int GRAYSCALE_STEP_COUNT = 24;
constexpr int GRAYSCALE_STEPS[GRAYSCALE_STEP_COUNT] = {
0x08, 0x12, 0x1c, 0x26, 0x30, 0x3a, 0x44, 0x4e, 0x58, 0x62, 0x6c, 0x76,
0x80, 0x8a, 0x94, 0x9e, 0xa8, 0xb2, 0xbc, 0xc6, 0xd0, 0xda, 0xe4, 0xee};
typedef std::function<unsigned long(int, int)> GetPixelFunction;
int clamp_byte(int value);
int best_index(int value, const int STEPS[], int count);
double sqr(double n);
/**
* @brief Struct to represent a character to be drawn.
* @param fgColor RGB
* @param bgColor RGB
* @param codePoint The code point of the character to be drawn.
*/
struct CharData {
std::array<int, 3> fgColor = std::array<int, 3>{0, 0, 0};
std::array<int, 3> bgColor = std::array<int, 3>{0, 0, 0};
int codePoint;
};
// Return a CharData struct with the given code point and corresponding averag
// fg and bg colors.
CharData createCharData(GetPixelFunction get_pixel, int x0, int y0,
int codepoint, int pattern);
/**
* @brief Find the best character and colors
* for a 4x8 part of the image at the given position
*
* @param image
* @param x0
* @param y0
* @param flags
* @return CharData
*/
CharData findCharData(GetPixelFunction get_pixel, int x0, int y0,
const int &flags);
#endif