| Age | Commit message (Collapse) | Author |
|---|
|
This entry indicates which files need to distributed with the package.
Closes #98.
|
|
for Image component (#96)
* change noisiness of terminal output
ffmpeg no longer prints everything into the terminal unless we're in `--verbose` mode. percentage progress text stays on one line while not in verbose mode.
* Added hint to run `avp --verbose` if `avp --log` is run with no avp_debug.log file present
* Classic Visualizer: add invert option
* Image component: fix path commandline option
* Image component: restrict file formats in CLI to match GUI
* Color component: add tooltip to color2 picker (second color of gradients)
* change tests to work with pytest-xdist
avp core stores its config (location of `settings.ini`) in temp directories if using multiple workers to run tests, so they don't interfere with each other. when using a single worker, the `tests/data/config` directory is still used
* check alt comp names when parsing cmdline
* rename `original.py` to `classic.py`
* move `component.py` into subpackage
* rename comp_original to comp_classic
* show traceback if renderFrame() raises exception
* do not try to insert non-existent components from project files
* add "composite" property for components
if a component returns "composite" then it will receive a frame to draw on during calls to previewRender and frameRender
* more tests of projects, actions, waveform, spectrum, image, color, classic
* do not change presetDir to "projects" within PresetManager
|
|
|
|
|
|
MainWindow.createNewProject (#94)
* search more than one filename for compName
* remove extra QVBoxLayout
* insert default components using name instead of index
* bump version 2.2.1 to 2.2.2
* tests do not use `.config` for settings
Core.storeSettings() is no longer called as a side effect any time avp.core is imported. Thus the tests use a new `initCore` method and the normal user path now relies on entering via `cli.py`. This means certain toolkit functions (e.g., ones using `FFMPEG_BIN`) no longer work if imported from a different python script, unless they call Core.storeSettings() themselves to initialize the settings.ini file
* fix UnboundLocalError in createNewProject
|
|
improve Conway default (#93)
* update gitignore
ignore profiling and coverage data
* F1 opens help window, create appName variable, move undostack class
* fix kaleidoscope effect, increase default Y values by +4
the increased y values allow the cells to continue animating for more than 60 minutes instead of 30 (at default 60f/t)
* update version number
* add minimumWidth to undo history window
* Classic Visualizer: option to include 64th bar
* Waveform component: fix #74 - new animation speed option
* move shared visualizer code into toolkit
* Waveform component: compress audio by default
* Waveform component: fix 100% animation speed
* new components receive random color
* update to Qt 6
* fix pushbutton stylesheet
* fix #92: replace ok/cancel with save/discard/cancel
* remove obsolete PaintColor subclass
* mv common shadow code into addShadow func
* add 3rd option of ok/cancel back to showMessage
the 3 options are:
- ok
- ok/cancel
- save/discard/cancel
* Image component: add shadow option
* small test of rgbFromString
* fix color tuple string
* test another way to get comp names from CLI
* rename component tests, add some more
* Image component: scale shadow based on resolution
* catch AttributeError if previewRender returns None
* Text component: fix blur radius only able to increase
the relativeWidgets system causes QDoubleSpinbox to only allow increases, because it really only works with integeres, so I changed the blur radius into a normal QSpinBox. I noted where the problem exists within component.py for future reference. This commit also removes an unneeded VerticalLayout from the ui file
* remove unnecessary QVBoxLayout
* paste shadow at x,y instead of using offset method
* fix tests due to shadow change
* don't print warning in connectWidget due to QFontComboBox
|
|
|
|
|
|
also adds a dissolve effect between frames and a kaleidoscope effect
the fancier shape types ignore audio for now. Fixes #91
|
|
* qtbot is needed in any test that uses a QObject
previously these tests would fail if they ran before qtbot was initialized by another test. I'm now running tests in a random order
* add tests for drawBars, readAudioFile, BlankFrame
* replace numpy.seterr with numpy.errstate
* fix incorrect comment
* add MockVideoWorker and imageDataSum
* test further into visualization (less likely to be a false positive)
* test FloodFrame function
* add failing test for Image component
one step towards fixing #89
* test component name CLI parsing
* prevent log warning when 1 setting changed
* correct tests to use widgets when needed
* test undo and blockSignals
* remove stretch_scale (use scale only)
* image ignores scale if stretch checkbox checked
fixes #89
* test Title Text component, ffmpeg command
* Image v2: replace stretched setting with resizeMode
3 resize modes are scale, cover, and stretch. Scale only applies when resizeMode is set to scale. Cover uses ImageOps.fit() to stretch while maintaining aspect ratio. Also, spinBox_scale was moved to be underneath comboBox_resizeMode.
* change transformData into staticmethod
the purpose is to allow easier reuse in other components
* add respondToAudio option to Image component
this causes the image to scale up and down slightly based on the input audio file
* cache static portion of image when animating
increases rendering speed of a 1-minute video by 12 seconds (based on two manual tests anyway)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Use pyproject.toml + uv_build
|
|
This replaces setup.py by a modern pyproject.toml using uv_build
backend.
Dependencies are being also managed by uv, so to install dependencies
and run the project one can execute:
```
uv sync
uv run pytest # optional
python -m avp
```
To build the both source and binary (wheel) distribution package run:
```
uv build
```
Uv can be installed with `pip install uv`.
The directory structure has been changed to reflect best practices.
- src/* -> src/avp/
- src/tests -> ../tests
|
|
|
|
|
|
and yeah, I accidentally ran black on the codebase. I don't want to spend more free time fixing that. All of these changes are simple renames or removals, nothing too major.
|
|
|
|
also remove outdated macOS instructions, which probably don't work anymore
|
|
|
|
instead of pre-computing the entire evolution during preFrameRender, it creates 60 ticks' worth of grid data at a time during the rendering process, deleting old data as it goes. This works with the current render process and should still work if the render process changes, as long as render threads don't become extremely de-synchronized (which would probably be an issue anyway)
|
|
|
|
|
|
added comments to preFrameRender and gridForTick methods
|
|
|
|
|
|
|
|
|
|
|
|
remove useless pytest fixtures
|
|
|
|
it's worth noting, but not important enough for 'error'
|
|
Looks less weird when window is maximized
|
|
Allows to decide how balanced the up/down smoothing is. The default (0) is the same as before; increasing sensitivity makes the up/down values closer to each other (less smooth, more sensitive)
|
|
|
|
|
|
Also remove unneeded instance variables `parser` and `args`, and quit if input/output can't be determined.
Also, --debug and --test are now mutually exclusive
|
|
|
