UAVReaders

Parsers for UAV-related formats like DataFlash and Mavlink

Installation and Usage

This package is available from PyPI. You can install it by doing something like this: pip install UAVReaders.

Currently, prebuilt wheels are available for the majority of Linux distributions and no other operating systems at all.

Example

Here is an example of using data_flash_reader in a script.

import UAVReaders as u 

dfreader = u.data_flash_reader("test_data/dataflash_test.bin")

for packet in dfreader():
    # do something with the packet, which is a dict containing 
    # the values of each field in the original packet.

I would advise you not to keep the packets around for much longer than necessary. Do not save them into a Python list, for example; it will probably take too much memory if your DataFlash file is large. Expect a 50x memory penalty. A 1GB DataFlash file might take up 50GB when each packet is simultaneously in memory because of duplicating packet field names.

Usage from Python by cloning this repository

It is also easy to clone and use this repository as part of a larger project, for example by adding it as a git submodule. There are a few build-time dependencies that you need to install before building the package.

• regex, install with conda install regex
• future, install with conda install future
• cython, also available through conda with conda install cython
• A working Linux-ish C++ toolchain with make. GCC is what I have been using. It should technically be possible to build natively on windows, but I would just use WSL.

Now, you have to make sure that Git has pulled all the submodules this project requires by running git submodule update --init --recursive.

After that, just cd into the directory and run make pypackage. Import uavreaders.UAVReaders to load it into your python. This provides access to UAVReaders.data_flash_reader, UAVReaders.mavlink_reader, and a few auxiliary functions not needed during normal usage.

Commandline Usage

Running make alone will generate both the python package and two binaries that can be used to transform DataFlash or Mavlink files to JSON. They are called dataflashreader and mavlinkreader and are written to /bin under the project root.

They both have the same interface. They take data through stdin write to stdout and stderr, so you should use them through input redirection. To use dataflashreader under Bash, run something like this:

$ ./bin/dataflashreader <test_data/dataflash_test.bin | gzip >output.json.gz

This reads dataflash packets from testdata/dataflash_test.bin and compresses the output JSON, storing it in output.json.gz. JSON representations of MavLink and Dataflash packets are very large; you will almost always want to compress the output. A one-gigabyte DataFlash .bin file might expand by as much as 50 times on parsing to JSON.

Testing

Run make test to test the functionality of the DataFlash and Mavlink parsers. This compiles the command-line parsing utilities and compares their output against a known-good state.