listing_summary =
filesets | int | rows | len(rows)
size | filesize | rows | sum(x['values'][1] or 0 for x in rows)
...
listing_sort = Start time | descending
listing_columns =
Name | route | fileset | detail_route(fileset)
Size | filesize | fileset | sum(x.size for x in fileset.files)
...Rich dynamic web interface
MARV Robotics allows you to make vast amounts of data accessible by web. Your team can access all data and get an overview in a dynamic web GUI. Filter and find information, comment on and tag it for future reference.
@marv.node()
@marv.input('messages', filter=['*:sensor_msgs/NavSatFix'])
def utm(messages):
for _, msg, _ in messages:
east, north = utm.from_latlon(msg.longitude, msg.latitude)[:2]
yield {'x': east, 'y': north}
@marv.node()
@marv.input('utm')
@marv.param('magic', help='Magic render parameter')
def utm(utm, magic):
plt_path, plt_id = marv.make_file(name)
fig = render(plt_path, utm, magic)
return plt_idDriven by your algorithms
Your usecases define what information is relevant for you. MARV Robotics enables you to embed your own custom algorithms to extract this information from robot logs. Build complex networks of processing nodes to trigger algorithms, extract features, and visualize to your needs. And the best thing: it's easy.
[detail bag_meta]
title = Bag metadata
...
[detail topics]
title = Topics
...
[detail trajectory]
title = Trajectory
widget = osm
inputs = geo_json_trajectory
initial_zoom = 21Configurable to the core
Your algorithms extract all relevant information. MARV Robotics does all the rest. Adjust MARV's simple configuration file and create the powerful GUI that your developers deserve and your execs are going to love - no web development experience required.
import requests
url = 'https://ternaris.com/marv-robotics/demo/'
api = '/marv/api/2/fileset-listing'
query = '?filter={"tags": {"op": "any", "val": ["outdoor"]}}'
requests.get(url + api + query).json()Supercharge your workflows
- MARV Robotics is a team player
- Existing tools can interface with the API
- Command line interface for developers to e.g. playback robot logs
- Regression testing and benchmarking made easier by finding and streaming adequate data