README.md

pyulgresample

Package that processes ulog-data from pyulog. It converts ulog-data into pandas dataframe through resampling and provides convenient functions to add and extract additional information from the ulog-data. To convert a .ulg file into ulog, please follow pyulog. With the ulogconv module the ulog can then be farther processed into a panda dict:

pandadict = ulogconv.createPandaDict(ulog)

pandaditct is a python dictionaires where each key is a panda dataframe that corresponds to a .ulg message.

To resample and merge all the dataframes of the pandadict:

dataframe = ulogconv.merge(pandadict)

By default, the merge-method uses linear interpolation for resampling. (Note: pandas merge-asof method is also supported)

dataframe is a pandas dataframe with index equal to the merged timestamps. Each column represents a message-field. For instance, the thrust-field of the message vehicle_local_position_setpoint message would be named as follow:

T_vehicle_local_position_setpoint_0__F_thrust_x

if the field x of vehicle_local_position_setpoint is a scalar or

T_vehicle_local_position_setpoint_0__F_x_0

if the field x is an array, where the 0 represents the index of the array.

The T stands for topic, which indicates the beginning of the topic. In this example, the topcic name is vehicle_local_position_setpoint. The topic name is followed by a number, which indicates the topic instance. If there is only one instance of a specific topic, then this number will be 0. The instance number is followed by two underlines and a capital letter F, which stands for field. In the example above, the field in question is x.

Informaiton can also be generated from the given .ulg data. If a new field is generated (for instance tilt-field can be generated from the attitude topic), then this new field message will have a NF instead of F to indicate the start of the field.

installation

To prevent any conflict with the system python version, it is suggested to use a virtual enrionment with python version 3.6 and higher. Otherwise, python 3.6 and higher must be the python system version. If you don't have 3.6 installed on your machinge, you can follow this tutorial.

virtualenvwrapper

First install virtualenv:

sudo apt install virtualenv

Install virtualenvrapper: this will install virtualenvwrapper.sh in ~/.local/bin

pip install virtualenvwrapper

Create a virtual environement directory

mkdir ~/.virtualenvs

Add virtual envrionment working-folder to bashrc and source virtualenvwrapper:

export WORKON_HOME=$HOME/.virtualenvs
source $HOME/.local/bin/virtualenvwrapper.sh

Open new terminal or source bashrc:

source ~/.bashrc

Create a virtual environment with python version 3 and no site packages included (python3 must be installed)

mkvirtualenv --python=python3 --no-site-packages [name-of-new-env]

You now created a new virtual environment with name [name-of-new-env].

To enter [name-of-new-env]:

workon [name-of-new-env]

To exit [name-of-new-env]:

deactivate

build setup

The build-system in use is flit

pip install flit

Now we can build the projct:

flit install -s

The -s stands for symlink which gives the option to test changes without reinstalling the package.

The projcet uses black for code-formatting and flake8 for style-guide enforcement. pre-commit-framework is used to ensure that each commit first gets adjusted through blake and then checked by flake8. PEP257 docstring style checker is used as well. First, we need to add pre-commit to our system:

pre-commit install

To test if everything is setup correctly, you can run the px4_attitude-script, which creates a pdf with a few plots in it. Make sure that the path to the ulg-file exists.

px4_attitude logs/cd6d8090-8c7f-41e5-bf62-3c95cc09fba1.ulg