dora-robomaster/graphs/dataflow.yml

nodes:
  - id: robot
    operator:
      # python: /home/peiji/anaconda3/envs/dora38/bin/python
      # args:   ../operators/robot.py
      python: ../operators/robot.py
      inputs:
        blaster:
          source: llm/blaster
          queue_size: 1
        led:
          source: llm/led
          queue_size: 1
        control:
          source: planning/control
          queue_size: 1
        gimbal_control:
          source: planning/gimbal_control
          queue_size: 1
        tick:
          source: dora/timer/millis/100
          queue_size: 1
      outputs:
        - position

  - id: bot_webcam
    custom:
      source: ../operators/opencv_stream.py
      outputs:
        - image

  - id: object_detection
    operator:
      python: ../operators/object_detection.py
      inputs:
        image: bot_webcam/image
      outputs:
        - bbox

  ### Second Camera
  - id: webcam
    operator:
      python: ../operators/webcam.py
      inputs:
        tick:
          source: dora/timer/millis/50
          queue_size: 1
      outputs:
        - image

  - id: plot_webcam
    operator:
      python: ../operators/plot.py
      inputs:
        image: webcam/image
        text: whisper/text  


  - id: plot_bot
    operator:
      python: ../operators/plot.py
      inputs:
        image: bot_webcam/image
        keyboard_buffer: keyboard/buffer
        user_message: keyboard/submitted
        assistant_message: llm/assistant_message
        bbox: object_detection/bbox


  - id: planning
    operator:
      python: ../operators/planning_op.py
      inputs:
        position: robot/position
        bbox: object_detection/bbox
        tick: dora/timer/millis/100
      outputs:
        - control
        - gimbal_control
        - led
        - blaster

  ## Speech to text
  - id: keyboard
    custom:
      source: ../operators/keyboard_op.py
      outputs:
        - buffer
        - submitted
        - record
        - ask
        - send
        - change
      inputs:
        recording: whisper/text

  - id: microphone
    operator:
      python: ../operators/microphone_op.py
      inputs:
        record: keyboard/record
      outputs:
        - audio

  - id: whisper
    operator:
      python: ../operators/whisper_op.py
      inputs:
        audio: microphone/audio
      outputs:
        - text

  # ## Code Modifier
  # - id: vectordb
  #   operator:
  #     python: ../operators/sentence_transformers_op.py
  #     inputs:
  #       query: keyboard/change
  #       saved_file: file_saver/saved_file
  #     outputs:
  #       - raw_file

  - id: llm
    operator:
      python: ../operators/llm_op.py
      inputs:
        # code_modifier: vectordb/raw_file
        assistant: keyboard/ask
        message_sender: keyboard/send
      outputs:
        - modified_file
        - assistant_message
        - line
        - control
        - led
        - blaster
        - rotation

  - id: file_saver
    operator:
      python: ../operators/file_saver_op.py
      inputs:
        llm_output_file: llm/modified_file
      outputs:
        - saved_file

  - id: dora-record
    custom:
      source: dora-record
      inputs:
        llm_output_file: llm/modified_file
        # raw_file: vectordb/raw_file
        saved_file: file_saver/saved_file
        audio: microphone/audio
        whisper_text: whisper/text
        bbox: object_detection/bbox
        image: cv2_encoder/encoded_image
        position: robot/position
        control: planning/control
        gimbal_control: planning/gimbal_control

  ### Second Camera
  - id: cv2_encoder
    operator:
      python: ../operators/cv2_encoder_op.py
      inputs:
        image: bot_webcam/image
      outputs:
        - encoded_image