Facepalm**2: fix array-read-slice, like actually..

We need to subtract the first index in the array segment read, not the
first index value in the time-sliced output, to get the correct offset
into the non-absolute (`ShmArray.array` read) array..

Further we **do** need the `&` between the advance indexing conditions
and this adds profiling to see that it is indeed real slow (like 20ms
ish even when using `np.where()`).

piker/data/flows.py

@@ -48,6 +48,10 @@ from ._sharedmem import (
 from ._sampling import (
     open_sample_stream,
 )
+from .._profile import (
+    Profiler,
+    pg_profile_enabled,
+)
 
 if TYPE_CHECKING:
     from pyqtgraph import PlotItem
@@ -251,6 +255,13 @@ class Flume(Struct):
         for the caller to use to slice the input array if needed.
 
         '''
+        profiler = Profiler(
+            msg='Flume.slice_from_time()',
+            disabled=not pg_profile_enabled(),
+            ms_threshold=4,
+            # ms_threshold=ms_slower_then,
+        )
+
         times = arr['time']
         index = arr['index']
 
@@ -272,12 +283,12 @@ class Flume(Struct):
 
         # use advanced indexing to map the
         # time range to the index range.
-        mask: np.ndarray = (
+        mask: np.ndarray = np.where(
             (times >= start_t)
-            |  # fml, i guess it's not an `&` ??
+            &
             (times < stop_t)
         )
-
+        profiler('advanced indexing slice')
         # TODO: if we can ensure each time field has a uniform
         # step we can instead do some arithmetic to determine
         # the equivalent index like we used to?
@@ -289,6 +300,8 @@ class Flume(Struct):
         i_by_t = index[mask]
         try:
             i_0 = i_by_t[0]
+            i_last = i_by_t[-1]
+            i_first_read = index[0]
         except IndexError:
             if (
                 start_t < times[0]
@@ -306,17 +319,20 @@ class Flume(Struct):
                     None,
                 )
 
-        abs_slc = slice(
-            i_0,
-            i_by_t[-1],
-        )
+        abs_slc = slice(i_0, i_last)
+
         # slice data by offset from the first index
         # available in the passed datum set.
         read_slc = slice(
-            i_0 - index[0],
-            i_by_t[-1] - i_0,
+            i_0 - i_first_read,
+            i_last - i_first_read + 1,
         )
 
+        profiler(
+            'slicing complete'
+            f'{start_t} -> {abs_slc.start} | {read_slc.start}\n'
+            f'{stop_t} -> {abs_slc.stop} | {read_slc.stop}\n'
+        )
         # also return the readable data from the timerange
         return (
             abs_slc,