modified the .rst so the methods of certain modules are listed in a preferred order rather than alphabetically

Author: pavelkomarov

docs/source/code.rst

@@ -5,10 +5,9 @@ API documentation
    :maxdepth: 1
 
    finite_difference
-   kalman_smooth
-   linear_model
    smooth_finite_difference
    total_variation_regularization
+   kalman_smooth
+   linear_model
    optimize
    utils
-

docs/source/finite_difference.rst

@@ -2,4 +2,9 @@ finite_difference
 =================
 
 .. automodule:: pynumdiff.finite_difference
-    :members:
+    :no-members:
+
+.. autofunction:: pynumdiff.finite_difference.finite_difference
+.. autofunction:: pynumdiff.finite_difference.first_order
+.. autofunction:: pynumdiff.finite_difference.second_order
+.. autofunction:: pynumdiff.finite_difference.fourth_order

docs/source/kalman_smooth.rst

@@ -2,4 +2,10 @@ kalman_smooth
 =============
 
 .. automodule:: pynumdiff.kalman_smooth
-    :members:
+    :no-members:
+
+.. autofunction:: pynumdiff.kalman_smooth.rts_const_deriv
+.. autofunction:: pynumdiff.kalman_smooth.constant_velocity
+.. autofunction:: pynumdiff.kalman_smooth.constant_acceleration
+.. autofunction:: pynumdiff.kalman_smooth.constant_jerk
+.. autofunction:: pynumdiff.kalman_smooth.known_dynamics

docs/source/total_variation_regularization.rst

@@ -2,4 +2,12 @@ total_variation_regularization
 ==============================
 
 .. automodule:: pynumdiff.total_variation_regularization
-    :members:
+   :no-members:
+
+.. autofunction:: pynumdiff.total_variation_regularization.tvr
+.. autofunction:: pynumdiff.total_variation_regularization.velocity
+.. autofunction:: pynumdiff.total_variation_regularization.acceleration
+.. autofunction:: pynumdiff.total_variation_regularization.jerk
+.. autofunction:: pynumdiff.total_variation_regularization.iterative_velocity
+.. autofunction:: pynumdiff.total_variation_regularization.smooth_acceleration
+.. autofunction:: pynumdiff.total_variation_regularization.jerk_sliding

examples/1_basic_tutorial.ipynb

@@ -1,5 +1,3 @@
 """This module implements some common finite difference schemes
 """
 from ._finite_difference import finite_difference, first_order, second_order, fourth_order
-
-__all__ = ['finite_difference', 'first_order', 'second_order', 'fourth_order'] # So these get treated as direct members of the module by sphinx

examples/2a_optimizing_parameters_with_dxdt_known.ipynb

@@ -27,8 +27,8 @@ def finite_difference(x, dt, num_iterations, order):
             dxdt_hat[-1] = dxdt_hat[-2] # using stencil -1,0 vs stencil 0,1 you get an expression for the same value
         elif order == 2:
             dxdt_hat[1:-1] = (x_hat[2:] - x_hat[:-2])/2 # second-order center-difference formula
-            dxdt_hat[0] = x_hat[1] - x_hat[0]
-            dxdt_hat[-1] = x_hat[-1] - x_hat[-2] # use first-order endpoint formulas so as not to amplify noise. See #104
+            dxdt_hat[0] = (-3 * x_hat[0] + 4 * x_hat[1] - x_hat[2])/2 # use spaced out stencil to get endpoint formulas
+            dxdt_hat[-1] = (3 * x_hat[-1] - 4 * x_hat[-2] + x_hat[-3])/2          # that do not amplify noise. See #104
         elif order == 4:
             dxdt_hat[2:-2] = (8*(x_hat[3:-1] - x_hat[1:-3]) - x_hat[4:] + x_hat[:-4])/12 # fourth-order center-difference
             dxdt_hat[1] = (x_hat[2] - x_hat[0])/2

pynumdiff/finite_difference/__init__.py

@@ -1,5 +1,3 @@
 """This module implements Kalman filters
 """
 from ._kalman_smooth import rts_const_deriv, constant_velocity, constant_acceleration, constant_jerk, known_dynamics
-
-__all__ = ['rts_const_deriv', 'constant_velocity', 'constant_acceleration', 'constant_jerk', 'known_dynamics'] # So these get treated as direct members of the module by sphinx

pynumdiff/finite_difference/_finite_difference.py

@@ -78,8 +78,10 @@
                           {'gamma': (1e-4, 1e7),
                            'window_size': (3, 1000)}),
     rts_const_deriv: ({'forwardbackward': [True, False],
-                       'qr_ratio': [-16, -12, -9, -6, -3, 0, 3, 6, 9, 12, 16]},
-                      {'qr_ratio': [1e-20, 1e20]}),
+                       'order': [1, 3],
+                       'qr_ratio': [1e-16, 1e-12, 1e-9, 1e-6, 1e-3, 1, 1e3, 1e6, 1e9, 1e12, 1e16]},
+                      {'order': (1, 3),
+                       'qr_ratio': [1e-20, 1e20]}),
     constant_velocity: ({'forwardbackward': [True, False],
                          'q': [1e-8, 1e-4, 1e-1, 1e1, 1e4, 1e8],
                          'r': [1e-8, 1e-4, 1e-1, 1e1, 1e4, 1e8]},
@@ -169,7 +171,7 @@ def optimize(func, x, dt, search_space={}, dxdt_truth=None, tvgamma=1e-2, paddin
     categorical_params = {k for k,v in params.items() if isinstance(v, tuple)}
     categorical_combos = [dict(zip(categorical_params, combo)) for combo in product(*[params[k] for k in categorical_params])] # ends up [{}] if there are no categorical params
 
-    # The Nelder-Mead's search space is the Cartesian product of all dimensions where multiple options are given in a list
+    # The Nelder-Mead's search space is the dimensions where multiple numerical (float or castable to float) options are given in a list
     search_space_types = {k:type(v[0]) for k,v in params.items() if isinstance(v, list)} # map param name -> type, for converting to and from point
     if any(v not in [float, int, bool] for v in search_space_types.values()):
         raise ValueError("To optimize over categorical strings, put them in a tuple, not a list.")