Skip to content
Tools to decode the data produced by JPI engine monitors
Java Shell
Branch: master
Clone or download

Latest commit

Fetching latest commit…
Cannot retrieve the latest commit at this time.


Type Name Latest commit message Commit time
Failed to load latest commit information.


edmtools are a set of analysis tools built on the unattributed for the EDM-series of engine data monitors produced for general aviation aircraft by J.P. Instruments, Inc.


$ git clone


  • Java > 1.5
  • Ant
  • The Google protocol buffer compiler >= 3.0.0. As of this writing it is not yet released; clone the repository and build it.


  • $ ant
  • $ scripts/JpiDecode --list file.JPI
  • $ scripts/JpiDecode --flight xxx file.JPI
  • $ scripts/JpiDecode --v 3 --flight xxx file.JPI
  • $ scripts/JpiDecode --flight xxx file.JPI


Prints a JPI binary file in a human-readable (protocol buffer) format to STDOUT.

verbosity, 0-3
list the flight numbers in the JPI file to standard out
select a particular flight number to convert


Extracts some or all of a JPI file to another JPI file. This was written to extract samples for functional tests.

verbosity, 0-3
comma-separated list of flight numbers
replace the registration (eg tail or serial number) with an arbitrary string


import edmtools.JpiDecoder;
import edmtools.JpiDecoder.JpiDecoderConfiguration;
import edmtools.JpiInputStream;
import edmtools.Proto.JpiFile;

JpiFile jpiFile = JpiDecoder.decode(
    new JpiInputStream("DOWNLOADED.JPI"),

JPI File Format notes

The JPI binary file format stores a series of value deltas. Data corruption therefore insidiously affects a metric for the remainder of the flight. To guard against this, JPI prefixes each data record with a repeated marker and also checksums each data record with a postfix byte. JPI tools will abort processing upon encountering checksum errors; this toolkit will continue processing but log the error in the parse_warning field. The user should treat any remaining output following a parse_warning with suspicion. Data corruption in the masks at the beginning of each record will render the record unparseable as the expected vs actual quantity of deltas will differ and the program will under- or over-read the record.


You can’t perform that action at this time.