Minor docstring/formatting touchups.

networkx/drawing/layout.py

@@ -1137,24 +1137,20 @@ def forceatlas2_layout(
     linlog=False,
     dim=2,
 ):
-    """Forceatlas2 layout for networkx
+    """Forceatlas2 layout.
 
-    The ForceAtlas2 layout is a force-directed approach that
-    improves upon  ForceAtlas by simplifying  the parameters
-    [1]. It offers a  good compromise between clustering and
-    spatially separating  parts of  the network.  The layout
-    was originally  designed and used by  Gephi to visualize
+    The ForceAtlas2 layout is a force-directed approach that improves upon
+    ForceAtlas by simplifying the parameters [1]_. It offers a good compromise
+    between clustering and spatially separating parts of  the network.
+    The layout was originally designed and used by Gephi to visualize
     graphs in a continuous and interactive manner.
 
-    In  the  layout,  nodes  repulse  each other  similar  to
-    oppositely charged  particles, while edges  attract like
-    to  springs.  The  balance between  the  attractive  and
-    repulsive  forces  produce   visually  pleasing  layouts,
-    particularly for large graphs.
+    In the layout, nodes repulse each other similar to oppositely charged particles,
+    while edges attract like springs. The balance between the attractive and
+    repulsive forces produce visually pleasing layouts, particularly for large graphs.
 
-    The  algorithm  includes parameters  for  customization,
-    such as gravity (pull towards the center), scaling (size
-    of the layout) and jitter (random perturbations).
+    The algorithm includes parameters for customization, such as gravity (pull
+    towards the center), scaling (size of the layout) and jitter (random perturbations).
 
     Parameters
     ----------
@@ -1165,47 +1161,44 @@ def forceatlas2_layout(
     n_iter: number (default: 100)
         Simulation steps
     jitter_tolerance : number (default: 1.0)
-        Jitter  tolerance  for  adjusting  speed  of  layout convergence;
-        lower values are slower but yield more accuracy. Jitter controls
-        how much swinging is allowed. Values above 1 are discouraged.
-        Jitter helps break symmetries between similar parts of the graph
-        through random perturbations in the layout generation.
+        Jitter tolerance for adjusting speed of layout convergence; lower values
+        are slower but yield more accuracy. Jitter controls how much swinging is
+        allowed. Values above 1 are discouraged. Jitter helps break symmetries
+        between similar parts of the graph through random perturbations in the
+        layout generation.
     scaling_ratio : number (default: 2.0)
-        Controls  force scaling  constants k_attraction  and
-        k_repulsion
+        Controls force scaling constants k_attraction and k_repulsion
     distributed_action : bool (default: False)
         Normalizes the attraction forces according to the degree of the nodes
     strong_gravity : bool (default: false)
-        Controls the  "pull" to  the center  of mass  of the
-        plot (0,0)
+        Controls the "pull" to the center of mass of the plot
     adjust_sizes: bool (default: False)
         Prevent node overlapping in the layout
-    dissuade_hubs : bool (default: false)
+    dissuade_hubs : bool (default: False)
         Prevent hub clustering
     edge_weight_influence : bool (default: False)
-        Generate layout with or without considering the edge
-        weights
+        Generate layout with or without considering the edge weights
     linlog : bool (default: False)
         Use log attraction rather than linear attraction
     dim: number (default: 2)
-       Sets the dimensions of the layout. This parameter is ignored if pos is given.
+       Sets the dimensions of the layout. This parameter is ignored if `pos` is given.
 
     Examples
     --------
     >>> import networkx as nx
     >>> G = nx.florentine_families_graph()
     >>> nx.draw(G, pos=nx.forceatlas2_layout(G))
 
-    [1] https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0098679&type=printable
+    References
+    ----------
+
+    .. [1] https://journals.plos.org/plosone/article/file?id=10.1371/journal.pone.0098679&type=printable
 
     """
     import numpy as np
 
     def estimate_factor(n, swing, traction, speed, speed_efficiency, jitter_tolerance):
-        """
-        ForceAtlas2 helper function
-        Computes scaling factor for force
-        """
+        """Computes scaling factor for force."""
 
         # estimate jitter
         opt_jitter = 0.05 * np.sqrt(n)
@@ -1241,7 +1234,7 @@ def estimate_factor(n, swing, traction, speed, speed_efficiency, jitter_toleranc
     else:
         dim = len(next(iter(pos.values())))
 
-    # Scale the position of the node proportional to the @pos
+    # Scale the position of the node proportional to the pos
     max_pos_range = 1
     min_pos_range = 0
     # check if we have a valid pos else just return (empty graph)
@@ -1275,7 +1268,7 @@ def estimate_factor(n, swing, traction, speed, speed_efficiency, jitter_toleranc
     speed_efficiency = 1
     swing = 1
     traction = 1
-    for idx in range(n_iter):
+    for _ in range(n_iter):
         # compute pairwise difference
         diff = pos_arr[:, None] - pos_arr[None]
         # compute pairwise distance