Replace random values by hard-coded numbers in plot-types ...

... for the cases where we have only very few random numbers.
This makes the code a bit simpler, because we do not have to
seed. Also, a list of hard-coded numbers is slightly simpler to
grasp compared to a rng function. This is partly also in
anticipation of the switch to numpy's rng generator classes.
While conceptually superior, they are a bit more verbose and
introduce the additional local variable `rng` (IMHO this is the
benefit and the downside: It explicitly ties the rng state to
the local context -> clarity. But OTOH the user is bothered
with this local state even if it's not important to them - like
in our examples.)

Author: timhoffm

galleries/plot_types/basic/bar.py

@@ -11,9 +11,8 @@
 plt.style.use('_mpl-gallery')
 
 # make data:
-np.random.seed(3)
 x = 0.5 + np.arange(8)
-y = np.random.uniform(2, 7, len(x))
+y = [4.8, 5.5, 3.5, 4.6, 6.5, 6.6, 2.6, 3.0]
 
 # plot
 fig, ax = plt.subplots()

galleries/plot_types/basic/stem.py

@@ -11,9 +11,8 @@
 plt.style.use('_mpl-gallery')
 
 # make data
-np.random.seed(3)
 x = 0.5 + np.arange(8)
-y = np.random.uniform(2, 7, len(x))
+y = [4.8, 5.5, 3.5, 4.6, 6.5, 6.6, 2.6, 3.0]
 
 # plot
 fig, ax = plt.subplots()

galleries/plot_types/basic/step.py

@@ -11,9 +11,8 @@
 plt.style.use('_mpl-gallery')
 
 # make data
-np.random.seed(3)
 x = 0.5 + np.arange(8)
-y = np.random.uniform(2, 7, len(x))
+y = [4.8, 5.5, 3.5, 4.6, 6.5, 6.6, 2.6, 3.0]
 
 # plot
 fig, ax = plt.subplots()