Save Results¶
Save analysis results¶
There are two methods to save the analysis results of SptialTis.
Just directly access .result attribute:
result = st.cell_components(data).result
Or you can read it from anndata if you know the key:
result = st.get_result(data, 'your_key')
Save visualization results¶
If you set the CONFIG.AUTO_SAVE, the image can be saved automatically:
CONFIG.AUTO_SAVE = True # create a default saved path
CONFIG.AUTO_SAVE = "your_saved_folder" # provide your own saved path
Or you can manually save it:
sp.cell_components(data).save("saved_path/img.png")
The default will ensure a publication quality save (dpi>=300). To gain more access to the plot, you can get the plot instance:
sp.cell_components(data).fig # For matplotlib
sp.cell_components(data).plot # For bokeh & pyecharts
Note
Starting from 0.3.0, SpatialTis provides static visualization based on matplotlib for every analysis. The interactive visualizations are saved to .html by default. If you want to export them into static images, the best way is to use screen capture.
All the interactive visualization are rendered in HTML5 Canvas/SVG. To save these plots into static images, extra dependencies need to be installed. There are three visualization libraries used in spatialtis:
bokeh |
pyecharts |
matplotlib |
|
Renderer |
Canvas, SVG |
Canvas, SVG (Not perfect) |
WX, GTK, Cairo … |
Naive export |
.html, .svg |
.html, .png |
.png, .jpeg, .eps, .pdf … |
To export .svg in pyecharts, pass parameter renderer=’svg’ into plotting function to switch to svg backend, the default is ‘canvas’; However, the svg is not perfect, your exported image might has layout issue. This can only be solved if echarts.js support svg further in the future.
For bokeh, you need to install geckodriver with firefox or chromedriver with chromium, check selenium-python, download selenium afterwards; For pyecharts, phantom.js is recommended.
Installation of selenium¶
# use conda
conda install -c conda-forge selenium
# use pip
pip install selenium
Installation of phantomjs¶
# use conda
conda install -c conda-forge phantomjs
# use npm
npm install phantomjs
Installation of firefox and geckodriver¶
conda install -c conda-forge firefox geckodriver
Or you can download them from the official release site, remember to add geckodriver to your path if install manually.
Installation of chromium and chromedriver¶
If you have chrome on you system, you only need to install chromedriver, remember to add chromedriver to your path.