docs: fix some documentation comments (#52)

Author: danieleades

arithmetic-coding-core/src/model/max_length.rs

@@ -135,7 +135,7 @@ pub trait Model {
     fn max_length(&self) -> usize;
 }
 
-/// A wrapper which converts a [`fixed_length::Model`](Model) to a
+/// A wrapper which converts a [`max_length::Model`](Model) to a
 /// [`crate::Model`].
 #[derive(Debug, Clone)]
 pub struct Wrapper<M>
@@ -150,7 +150,7 @@ impl<M> Wrapper<M>
 where
     M: Model,
 {
-    /// Construct a new wrapper from a [`fixed_length::Model`](Model)
+    /// Construct a new wrapper from a [`Model`]
     pub fn new(model: M) -> Self {
         let remaining = model.max_length();
         Self { model, remaining }

arithmetic-coding-core/src/model/one_shot.rs

@@ -7,10 +7,12 @@ use crate::{fixed_length, BitStore};
 
 /// A [`Model`] is used to calculate the probability of a given symbol occuring
 /// in a sequence. The [`Model`] is used both for encoding and decoding. A
-/// 'fixed-length' model always expects an exact number of symbols, and so does
+/// 'one-shot' only ever encodes a single symbol, and so does
 /// not need to encode an EOF symbol.
 ///
-/// A fixed length model can be converted into a regular model using the
+/// A one-shot [`Model`] is a special case of the [`fixed_length::Model`].
+///
+/// A one-shot model can be converted into a regular model using the
 /// convenience [`Wrapper`] type.
 ///
 /// The more accurately a [`Model`] is able to predict the next symbol, the