BioLogger

biologger is a cross-platform keyboard and mouse event capture tool. It uses the jnativehook library to register system-wide hooks for keyboard and mouse events. The types of events captured include: keystrokes, mouse motion, mouse clicks, and scrolling (mouse wheel). The events are recorded to CSV files (one file for each event type).

To run, download the latest release and run the executable jar. From the command line,

$ java -jar biologger.jar

By default, the CSV files will be created in the current working directory. To specify a different output directory, use the -o option. Certain event types can also be ignored, such as -im to ignore mouse motion events. The full usage is:

usage: biologger
 -h,--help                      print this help message
 -ic,--ignore-click             ignore mouse click events
 -ik,--ignore-keystroke         ignore keystroke events
 -im,--ignore-motion            ignore mouse motion events
 -iw,--ignore-wheel             ignore mouse wheel events
 -mt,--motion-threshold <arg>   group motion events within mt milliseconds
 -nw,--no-window                don't start the user interface
 -o,--output <arg>              output directory
 -pk,--print-keys               print the full key map and exit
 -v,--verbose                   verbose mode

To be sure the events are being capture, run biologger in verbose mode:

$ java -jar -biologger.jar -v
Aug 27, 2020 10:50:32 AM com.vmonaco.bio.Listener mouseMoved
INFO: mousemove,1598550632149,842,783,40960,num_lock+scroll_lock,0
Aug 27, 2020 10:50:32 AM com.vmonaco.bio.Listener mouseMoved
INFO: mousemove,1598550632151,841,784,40960,num_lock+scroll_lock,0
Aug 27, 2020 10:50:32 AM com.vmonaco.bio.Listener mouseMoved
INFO: mousemove,1598550632153,841,786,40960,num_lock+scroll_lock,0

...

Event attributes and descriptions

The following tables describe each of the attributes (columns in the CSV files) for each event type.

Keystroke

Name	Description
press_time	key press timestamp (ms)
release_time	key release timestamp (ms)
key_code	numeric key code (see keycodes.csv for a table)
key_name	human-readable key name
modifier_code	modifier key codes pressed during the event
modifier_name	modifier key names pressed during the event (e.g., ctr+shift)
location	2 for left, 3 for right, 1 for keys that have a single location

Mouse motion

Name	Description
time	timestamp (ms)
x	absolute x location of the pointer (pixels)
y	absolute y location of the pointer (pixels)
modifier_code	modifier key codes pressed during the event
modifier_name	modifier key names pressed during the event
dragged	1 if a mouse button was held down during the event, 0 otherwise

Mouse track

Name	Description
start_time	timestamp (ms)
start_x	absolute x location of the pointer (pixels) at the start of the track
start_y	absolute y location of the pointer (pixels) at the start of the track
end_time	timestamp (ms)
end_x	absolute x location of the pointer (pixels) at the end of the track
end_y	absolute y location of the pointer (pixels) at the end of the track
distance_x	total horizontal distance traveled (pixels)
distance_y	total vertical distance traveled (pixels)
dragged	1 if a mouse button was held down during the event, 0 otherwise

Mouse click

Name	Description
press_time	press timestamp (ms)
release_time	release timestamp (ms)
button_code	mouse button numeric code (typically 1, 2, or 3)
press_x	absolute x position at the press time
press_y	absolute y position at the press time
release_x	absolute x position at the release time
release_y	absolute y position at the release time
modifier_code	modifier key codes pressed during the event
modifier_name	modifier key names pressed during the event
image	base64 encoded 200x200 image centered on the press position

The image attribute is a base64 encoded image centered on the press location. This can be used to provide context for where the user clicked the mouse, for example how close to the edge of a button the click occured.

The following python code decodes and displays the last click image in mouseclick.csv:

import io
import base64
import pandas as pd
from PIL import Image

df = pd.read_csv('mouseclick.csv')
i = df.iloc[-1].image
image = Image.open(io.BytesIO(base64.b64decode(i)))
image.show()

Mouse wheel

Name	Description
time	timestamp (ms)
amount	magnitude of the wheel event
direction	direction of the wheel event (either +1 or -1)
type	0 for unit, 1 for block
x	absolute x location of the pointer device
y	absolute y location of the pointer device
modifier_code	modifier key codes pressed during the event
modifier_name	modifier key names pressed during the event

Build instructions

See the Makefile. To build the jar:

$ make jar

To run the jar:

$ java -jar biologger.jar