Final max commit

pymc/logprob/order.py

@@ -111,105 +111,7 @@ def find_measurable_max(fgraph: FunctionGraph, node: Node) -> Optional[List[Tens
 
 @_logprob.register(MeasurableMax)
 def max_logprob(op, values, base_rv, **kwargs):
-    r"""Compute the log-likelihood graph for the `Max` operation.
-
-    Parameters
-    ----------
-    op : Max-Op
-    values : tensor_like
-    rv : TensorVariable
-
-    Returns
-    -------
-    logprob : TensorVariable
-
-    Examples
-    --------
-    It is often desirable to find the log-probability of the maximum of i.i.d. random variables.
-
-    The "max of i.i.d. random variables" refers to finding the maximum value among a collection of random variables that are independent and identically distributed.
-    The example below illustrates how to find the Maximum from the distribution of random variables.
-
-    .. code-block:: python
-
-        import pytensor.tensor as pt
-
-        x = pt.random.normal(0, 1, size=(3,))
-        x.name = "x"
-        print(x.eval())
-        #[0.61748772 1.08723759 0.98970957]
-
-        x_max = pt.max(x, axis=None)
-        print(x_max.eval())
-        # 1.087237592696084
-
-    The log-probability of the maximum of i.i.d. random variables is a measure of the likelihood of observing a specific maximum value in a set of independent and identically distributed random variables.
-
-    The formula that we use here is :
-        \ln(f_{(n)}(x)) = \ln(n) + (n-1) \ln(F(x)) + \ln(f(x))
-    where f(x) represents the p.d.f and F(x) represents the c.d.f of the distribution respectively.
-
-    An example corresponding to this is illustrated below:
-
-    .. code-block:: python
-
-        import pytensor.tensor as pt
-        from pymc import logp
-
-        x = pt.random.uniform(0, 1, size=(3,))
-        x.name = "x"
-        # [0.09081509 0.84761712 0.59030273]
-
-        x_max = pt.max(x, axis=-1)
-        # 0.8476171198716373
-
-        x_max_value = pt.scalar("x_max_value")
-        x_max_logprob = logp(x_max, x_max_value)
-        test_value = x_max.eval()
-
-        x_max_logprob.eval({x_max_value: test_value})
-        # 0.7679597791946853
-
-    Currently our implementation has certain limitations which are mandated through some constraints.
-
-    We only consider a distribution of RandomVariables and the logp function fails for NonRVs.
-
-    .. code-block:: python
-
-        import pytensor.tensor as pt
-        from pymc import logp
-
-        x = pt.exp(pt.random.beta(0, 1, size=(3,)))
-        x.name = "x"
-        x_max = pt.max(x, axis=-1)
-        x_max_value = pt.vector("x_max_value")
-        x_max_logprob = logp(x_max, x_max_value)
-
-    The above code gives a Runtime error stating logprob method was not implemented as x in this case is not a pure random variable.
-    A pure random variable in PyMC represents an unknown quantity in a Bayesian model and is associated with a prior distribution that is combined with the likelihood of observed data to obtain the posterior distribution through Bayesian inference
-
-    We assume only univariate distributions as for multivariate variables, the concept of ordering is ambiguous since a "depth function" is required .
-
-    We only consider independent and identically distributed random variables, for now.
-    In probability theory and statistics, a collection of random variables is independent and identically distributed if each random variable has the same probability distribution as the others and all are mutually independent.
-
-    .. code-block:: python
-
-        import pytensor.tensor as pt
-        from pymc import logp
-
-        x = pm.Normal.dist([0, 1, 2, 3, 4], 1, shape=(5,))
-        x.name = "x"
-        x_max = pt.max(x, axis=-1)
-        x_max_value = pt.vector("x_max_value")
-        x_max_logprob = logp(x_max, x_max_value)
-
-    The above code gives a Runtime error stating logprob method was not implemented as x in this case is a Non-iid distribution.
-
-    Note: We assume a very fluid definition of i.i.d. here. We say that an RV belongs to an i.i.d. if that RVs do not have different stochastic ancestors.
-
-
-    """
+    r"""Compute the log-likelihood graph for the `Max` operation."""
     (value,) = values
 
     logprob = _logprob_helper(base_rv, value)