[Documentation] [TitleIndex] [WordIndex

Only released in EOL distros:  

roboframenet: arbitrator | executor | frame_registrar | imperative_to_declarative | location_memorizer | moo | move_action_server | move_base_rfn | pr2_props_rfn | rfnserver | roboframenet_bringup | roboframenet_msgs | semantic_framer | stanford_parser | stanford_parser_msgs | stanford_parser_ros | stop_server | turtlebot_follower_rfn | utter

Package Summary

rfnserver (aka, RFNServer, or RoboFrameNet Server) is a class which provides bindings to RoboFrameNet for the purpose of adding semantic frames, adding lexical units, and registering callbacks to semantic frames.

Overview

rfnserver is a child class of actionlib's SimpleActionServer which requires a special type of action -- namely, FilledSemanticFrame.action in roboframenet_msgs.

A relatively stable version of RFNServer is working in Python.

Initializing an RFNServer requires a three lines of code -- one each for importing, initializing, and starting.

Adding semantic frames, and adding lexical units each require one line of code. (For adding semantic frames and lexical units, a corresponding yaml file is also needed. See: yaml files for lexical units, yaml files for semantic frames.)

Registration with a semantic frame requires one line of code for registration plus a function to handle the callback.

Initializing an RFNServer

from rfnserver import RFNServer

[...]

server = RFNServer("moo", moo)
# Add semantic frames and lexical units here.
# Register with semantic frames here.
server.start()

The first argument is the topic where you want the RFNServer (behind-the-scenes: SimpleActionServer) to publish goal, feedback, and preemption messages.

The second argument is a callback function which takes one argument -- namely, a filled semantic frame. (The message type is of type roboframenet_msgs/FilledSemanticFrame .) The frame passed into this callback will be from one of the semantic frames with which the RFNServer has registered.

To get arguments for a specific frame element of the semantic frame, the helper function frame_argument is provided:

def callback(filled_semantic_frame):
  [...]
  arg_str = RFNServer.frame_argument(filled_semantic_frame, "parameter")
  [...]

Adding Semantic Frames

server.add_frame("../frames/mooing.yaml")

The filepath is relative to the directory from which the program was launched. Typically, this means that before running, you'll want to either move to the node's directory, or you'll want to add the attribute cwd="node" to the node's tag in its roslaunch file.

Adding Lexical Units

server.add_lexical_unit("../lu/mooing.yaml")

The filepath is relative to the directory from which the program was launched. Typically, this means that before running, you'll want to either move to the node's directory, or you'll want to add the attribute cwd="node" to the node's tag in its roslaunch file.

Registering with a Semantic Frame

server.register_with_frame("mooing")

Upon registration, the callback specified in the initialization of the RFNServer may receive a filled semantic frame of the specified type.


2019-07-13 13:03