Add live_view module to qutil.plotting.

The live_view module provides a matplotlib interface for concurrent data viewing using animations and blitting. For line plots without axis rescaling, frame rates of ~100 fps can be reached. A 1080x1920 pixel image can be displayed at ~20 fps. Data sources are pushed to a background thread from where they can fill a queue.

See mjolnir commits a771a5cc..205318e6 for earlier history.

qutil/plotting/live_view.py

@@ -1080,9 +1080,10 @@ class BatchedLiveView1D(LiveView1DBase):
     Define a function that generates random data. The function will be
     pushed to a background thread so that the interpreter is unblocked.
 
-    >>> rng = np.random.default_rng()
     >>> def produce_data(interval=1e-1):
-    ...     import time
+    ...     import time  # imports for tests in child process
+    ...     import numpy as np
+    ...     rng = np.random.default_rng()
     ...     t0 = t1 = time.perf_counter()
     ...     while True:
     ...         dt = time.perf_counter() - t1