vector-go-sdk

vector-go-sdk

A go SDK for Vector! This is an early alpha and not all features are present.

Usage

Currently. the SDK expects to find in your home directory the Python SDK data. It consists of a .anki_vector folder where a sdk_config.ini file is placed. This file contains the configuration for each robot:

[ROBOT SERIAL NO]

cert = path to certificate (not needed by the go SDK)

ip = robot ip

name = robot name

guid = token for connecting to the robot

When you instantiate the robot you provide the serial number, then the SDK code will parse the .anki_vector/sdk_config.ini and get the data it needs. I develop and test this version of the go SDK using a production Vector 1.0 and wirepod

Examples

Please see the cmd/examples directory for usage examples.

Known Issues

• FaceEnrollmentStart(personName string)
  • Adding a new face enrollment via SDK doesn't seem to work. The procedure completes but the face is not saved on robot
• SaveHiResCameraPicture(fileName string)
  • This doesn't seem to work on my production Vector on Wirepod. Probably vector-cloud on the robot needs to be updated. But since it's a production robot I'm stuck to 1.8... Anyway the image is saved at the regular 360p size.

Changelog

+RELEASE_ALPHA_22

• Added a custom setting for game in progress. This way a game can progress towards multiple STT invocations and its state can be made persistent

+RELEASE_ALPHA_21

• Added sdk-wrapper-oskr for OSKR-only functions:
  • TriggerWakeWord() to trigger a voice input prompt on Vector
• Added oskr.proto / oskr.pb.go, oskr_grpc.pb.go to allow communication with a custom grpc server to be run on OSKR See an example in examples/navigator, this needs to be used in conjunction with the server counterpart that you can find here: https://github.com/kirillgrishin-tech/vector-go-sdk-oskr-extensions

+RELEASE_ALPHA_20

• Improved TTS support. Added custom voice server that provides custom voices from different providers (currently only using FakeYou.com)

+RELEASE_ALPHA_19

• GetEyeColor() in case of custom eye colors returns always full white

+RELEASE_ALPHA_18

• Fix InitSDKForWirepod() function

+RELEASE_ALPHA_17

• Added a couple of extra fields in custom settings to use with VIM chat

+RELEASE_ALPHA_16

• Added SetChatTarget() and GetChatTarget() to save username with whom this bot is chatting with

+RELEASE_ALPHA_15

• Added GetStaticCameraPicture()
• Exposed a few image functions
• Fixed eSpeak TTS: now the proper language voice is used
• Bugfix: Using events with wirepod crashes the application

+RELEASE_ALPHA_14

• Bugfix: TextOnImg uses an hardcoded data path
• Bugfix: InitSDK doesn't correctly initialize event stream (event example broken!)
• Bugfix: temporary path not correctly used

+RELEASE_ALPHA_13

• Set TTS output always at the maximum volume

+RELEASE_ALPHA_12

• Bugfix: Language not correctly set to robot's locale language when initializing the sdk

+RELEASE_ALPHA_11

• Bugfix: SetRobotName doesn't write permanent data

+RELEASE_ALPHA_10

• Bugfix: SetRobotName doesn't write permanent data

+RELEASE_ALPHA_09 Integration with VectorX (https://github.com/fforchino/vectorx)

+RELEASE_ALPHA_08 Integration with Wirepod

+RELEASE_ALPHA_07 CODE REFACTORING aimed at:

• Make all paths os-independent
• Make all temporary files bot-dependent, to allow usage in a scenario where multiple robots are connected to an escape pod
• Organize external files in 3 categories:
  • (TMP) Temporary files living in a TmpPath, they are relative to the bot invoking a certain function
  • (DATA) Data files, they are read only files that the SDK needs, so they are shared among all bots
  • (NVM) Persistent cloud storage available to the bot +- The introduction of NVM allows the implementation of CustomSettings, an extra data structure that every bot now owns and contains new settings, like robot_name

Forgive me for the very quick description, I will soon document all properly in a wiki. The important thing is that the GO SDK the way it is now is ready for the integration with Wirepod. Hooray! Remember to get the updated version: go get -u github.com/kirillgrishin-tech/vector-go-sdk/pkg/sdk-wrapper

RELEASE_ALPHA_06

• Image functions
  • Add the option to use Vector's eye color to blend the images with (kind of near monochrome effect)
• Audio functions
  • Use master volume to play audio files. Set the master volume higher if you want to play audio louder
• Settings
  • Added utility to get Vector's eye colors as RGB, and audio volume in the range 0..100
• Examples
  • Add full example of "Roll a die" (uses the features above)

RELEASE_ALPHA_05

• Image functions
  • Added animated gif processing
  • Fixed a bug in image resize
  • Added fade in/fade out (from/to black) image transition

RELEASE_ALPHA_04

• Added voice functions (internationalized TTS using htgo or espeak TTS engines)
  • Some voices are not directly supported in espeak (just need to figure out what dependencies to install)

RELEASE_ALPHA_03

• Removed dependency from OpenCV, because goCV sucks and requires user to download a lot of stuff. Better use OpenCv in Python!
• Bugfix: WriteText() / DisplayImage() / DisplayImageWithTransition() Image and text displayed have garbled colors

RELEASE_ALPHA_02

• To try to support hires camera image snapshot I added OpenCV as a dependency. To install OpenCV on a Raspberry Pi or other Linux environment, run setup.sh
• Added camera functions
• Added face enrollment functions
• Added events with examples (Vector roars when touched)

RELEASE_ALPHA_01

I have taken the go SDK and did some changes.

• Integrated the functions of the great Wire's vector-web-api-app as sdk_wrapper functions
• Added many examples:
• The SDK supports features in these fields:
  • Audio playback and volume settings
  • Custom eye colors
  • Animations
  • Settings