Data Plotter

Data Plotter is a simple tool useful to plot data, specifically streams of data sent by a microcontroller.

Data needs to be written on a file that is then passed as argument to the program, which will parse the file and display a graph.

Dependencies

The script depends on numpy, and other packages depending on the chosen backend:

• for the matplotlib backend: matplotlib, mplcairo and pyqt5 (or pyqt6)
• for the bokeh backend: bokeh

The user may use the provided requirements.txt file to install all dependencies (and backend dependencies) by running

pip install -r requirements.txt

Usage

usage: plot.py [-h] [--line-mode] [--maximum MAXIMUM] [--backend {matplotlib,bokeh}] filename

positional arguments:
  filename              Path to data file

options:
  -h, --help            show this help message and exit
  --line-mode           A boolean switch that controls whether the program is in line mode or in
                        scatter mode
  --maximum MAXIMUM     Maximum value for the signals, after which they get clipped (default: 4096)
  --backend {matplotlib,bokeh}
                        Sets the rendering backend (default: matplotlib)

where filename is the path to a text file with the following structure:

• Each signal is represented by a series of lines with the following structure:
  1. a line containing the label for the signal, to be displayed in the legend;
  2. a line for each sample, containing a whitespace-separated pair of coordinates to plot;
     • some x values may be omitted by specifying * instead, and they will be interpolated between adjacent ones. This feature is meant to be used with values obtained through DMA, which are harder to put a timestamp on.
       • Interpolation of the first/last value(s) is impossible, and will yield the x of the next/previous point;
       • to ensure correct interpolation x values must be monotonically nondecreasing.
     • if a sample line does not respect this syntax the script will try to ignore it, outputting a warning. this is useful for unreliable channels of communication where it is not crucial to never lose any samples;
  3. a line contanining the string "END_OF_SIGNAL"
• The file may contain more than one signal block with the same signal label; in that case the points will be added to the same label, allowing interleaving signals.

Comments and spacing:

• if an input line contains '#' followed by any other characters those will be considered a comment and ignored ('#' included, of course);
• empty lines and lines containing only whitespace or comments will be ignored;
• whitespace at the beginning and end of input lines will be ignored.

Example input file

1 # number of signals

name of signal # the name that will appear in the graph's legend
3 # number of samples
# list of "x y" tuples to plot. x should always increase
1 2
2 3
3 8
END_OF_SIGNAL

Some other more advanced example input files are provided under example_inputs/.

Serial

To connect correctly to the stm32 you need to configure some parameters of the serial terminal you are using. In the repo you can find the setup to use with moserial. You can copy the parameters directly from there under [port_settings].

sw-data-plotter/
|- example_inputs/
|- ...
|- *moserial-setup*
|- ...

Profiling:

You may profile the script with the following command:

python -m pyflame -a "--width 4000" plot.py example_inputs/many_points.txt

Note

Before running it pyflame and flamegraph need to be installed.

The command will generate a file "plot.py.svg" in the working directory. To visualize it you may use gimp or any other image viewer supporting svg.