Fast acquisition from a Vimba camera in Python

pymba_demo.py

"""
This demonstration assumes you have already installed Pymba and followed the installation instructions there:
https://github.com/morefigs/pymba

MoviePy is used to save the frames to a video file for later viewing, while the frame's mean intensity is
printed in real time. MoviePy can be installed from pip nicely.
"""
import pymba
from moviepy.video.io.ffmpeg_writer import FFMPEG_VideoWriter as VidWriter
from skimage.color import gray2rgb
import time

# Start by getting the Vimba API set up
# Using the context (i.e. 'with' syntax) is a good idea
# If something crashes, vimba.shutdown() will get called automatically
# and your camera will not get left in an unusable state!
with pymba.Vimba() as vimba:
    print vimba.getVersion()
    
    # Find any cameras that are attached
    system = pymba.getSystem()
    system.runFeatureCommand("GeVDiscoveryAllOnce")
    time.sleep(0.2) # feature commands run in another thread, but need a moment to execute
    camera_ids = vimba.getCameraIds()
    
    # Just use the first camera we found
    vmb_cam = vimba.getCamera(camera_ids[0])
    vmb_cam.openCamera()
    
    # Set up the camera with the right parameters
    # Includes examples of setting up the output TTLs
    vmb_cam.PixelFormat = 'Mono8'
    vmb_cam.AcquisitionMode = 'Continuous'
    vmb_cam.BinningHorizontal = 4
    vmb_cam.BinningVertical = 4
    vmb_cam.Height = vmb_cam.HeightMax
    vmb_cam.Width = vmb_cam.WidthMax
    vmb_cam.AcquisitionFrameRateAbs = 60
    vmb_cam.SyncOutSelector = 'SyncOut1'
    vmb_cam.SyncOutSource = 'Exposing'
    vmb_cam.TriggerSource = 'FixedRate'
    
    # Set up the video file where our frames will be logged
    video_logger = VidWriter(filename="vimba_test.mp4",
                             size=(vmb_cam.Width, vmb_cam.Height),
                             fps=vmb_cam.AcquisitionFrameRateAbs,
                             code='mpeg4',
                             preset='ultrafast')
    
    # Now we set up a callback that will process frames as they arrive
    def frame_callback(frame):
        frame_data = frame.getBufferByteData()
        img = np.ndarray(buffer=frame_data,
                         dtype=np.uint8,
                         shape=(frame.height, frame.width))
        print img.mean()
        video_logger.write_frame(gray2rgb(img))
        frame.queueFrameCapture(frame_callback)
    
    # Finally, create a buffer of frames. Vimba's way of doing this
    # is odd: you just make a bunch of frame objects and Vimba decides
    # what order to fill them in.
    n_vimba_frames = 10
    frame_pool = [vmb_cam.getFrame() for _ in xrange(n_vimba_frames)]
    for frame in frame_pool:
        frame.announceFrame() # now Vimba knows about the frame... also odd design choice
        frame.queueFrameCapture(frame_callback)
        
    # Start capturing frames!
    vmb_cam.startCapture()
    vmb_cam.runFeatureCommand('AcquisitionStart')
    time.sleep(30) # 30 seconds of acquisition, proves that frames are written on another thread
    vmb_cam.runFeatureCommand('AcquisitionStop')
    time.sleep(0.2)
    vmb_cam.flushCaptureQueue() # unqueues all frames, must be called before endCapture()
    vmb_cam.endCapture()
    vmb_cam.revokeAllFrames() # makes Vimba forget about the frames it's just unqueued

Hi,
Minor edits:
Line 12: add import numpy as np
Line 21: system = pymba.getSystem() should be system = vimba.getSystem()

I am getting an error

File "C:/Users/######/Desktop/Python Video Recorder/untitled0.py", line 20, in
system = vimba.getSystem()

AttributeError: 'Vimba' object has no attribute 'getSystem'

I have Pymba and the Vimba ADK installed using conda, any idea?

I believe the method names have changed since this code was published.
vimba.getSystem() -> vimba.system()
system.runFeatureCommand() -> system.run_feature_command()
vimba.getCameraIds() -> vimba.camera_ids()
...and so on.
Just type, for example vimba. in your console, and go over the suggested methods and use them instead.

I was scratching my head on how to modify some methods for the vmb_cam object. A convenient approach to figuring out which methods can be called for a particular object is to employ dir():

obj_mth = [mth_nam for mth_nam in dir(vmb_cam) if callable(getattr(vmb_cam, mth_nam))]

For vmb_cam, this yields:
['__class__', '__delattr__', '__dir__', '__eq__', '__format__', '__ge__', '__getattr__', '__getattribute__', '__gt__', '__hash__', '__init__', '__init_subclass__', '__le__', '__lt__', '__ne__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__sizeof__', '__str__', '__subclasshook__', '_feature_info', '_feature_infos', '_streaming_callback', 'acquire_frame', 'arm', 'close', 'disarm', 'end_capture', 'feature', 'feature_command_is_done', 'feature_names', 'flush_capture_queue', 'new_frame', 'open', 'read_register', 'register_feature_invalidation_callback', 'revoke_all_frames', 'run_feature_command', 'start_capture', 'start_frame_acquisition', 'stop_frame_acquisition', 'unregister_all_feature_invalidation_callbacks', 'unregister_feature_invalidation_callback', 'write_register']