feat: Add log_prob method to Model

Added log_prob method to Model instances, allowing users to calculate
the log probability of a list of unconstrained parameters.

This feature is accompanied by a test: the log_prob method is validated
by comparing the output against the log probability (lp__) extracted
from a model fit.

Closes #40

stan/model.py

@@ -308,6 +308,39 @@ async def go():
 
         return asyncio.run(go())
 
+    def log_prob(self, unconstrained_parameters: List[float], adjust_transform: bool = True) -> float:
+        """Calculate the log probability of a set of unconstrained parameters.
+
+        Arguments:
+            unconstrained_parameters: A sequence of unconstrained parameters.
+            adjust_transform: Apply jacobian adjust transform.
+
+        Returns: The log probability of the unconstrained parameters.
+
+        Notes:
+            The unconstrained parameters are passed to the log_prob
+            function in stan::model.
+
+        """
+        assert isinstance(self.data, dict)
+
+        payload = {
+            "data": self.data,
+            "unconstrained_parameters": unconstrained_parameters,
+            "adjust_transform": adjust_transform,
+        }
+
+        async def go():
+            async with stan.common.httpstan_server() as (host, port):
+                log_prob_url = f"http://{host}:{port}/v1/{self.model_name}/log_prob"
+                async with aiohttp.request("POST", log_prob_url, json=payload) as resp:
+                    response_payload = await resp.json()
+                    if resp.status != 200:
+                        raise RuntimeError(response_payload)
+                    return (await resp.json())["log_prob"]
+
+        return asyncio.run(go())
+
 
 def build(program_code, data=None, random_seed=None) -> Model:
     """Build (compile) a Stan program.

tests/test_log_prob.py

@@ -0,0 +1,54 @@
+"""Test model with array parameter."""
+import numpy as np
+import pytest
+
+import stan
+
+unrestricted_program = """
+    parameters {
+        real y;
+    }
+    model {
+        y ~ normal(0, 1);
+    }
+"""
+
+restricted_program = """
+    parameters {
+        real<lower=0> y;
+    }
+    model {
+        y ~ normal(0, 1);
+    }
+"""
+
+num_samples = 1000
+num_chains = 4
+
+
+@pytest.fixture
+def posterior(request):
+    return stan.build(request.param, random_seed=1)
+
+
+@pytest.mark.parametrize("posterior", [unrestricted_program], indirect=True)
+def test_log_prob(posterior):
+    """Test log probability against sampled model with unrestriction."""
+    fit = posterior.sample(num_chains=num_chains, num_samples=num_samples)
+    y = fit["y"][0][0]
+    lp__ = fit["lp__"][0][0]
+    lp = posterior.log_prob(unconstrained_parameters=[y])
+    assert np.allclose(lp__, lp)
+
+
+@pytest.mark.parametrize("posterior", [restricted_program], indirect=True)
+def test_log_prob_restricted(posterior):
+    """Test log probability against sampled model with restriction."""
+    fit = posterior.sample(num_chains=num_chains, num_samples=num_samples)
+    y = fit["y"][0][0]
+    y = posterior.unconstrain_pars({"y": y})[0]
+    lp__ = fit["lp__"][0][0]
+    lp = posterior.log_prob(unconstrained_parameters=[y], adjust_transform=False)
+    assert np.allclose(lp__, lp + y)
+    adjusted_lp = posterior.log_prob(unconstrained_parameters=[y], adjust_transform=True)
+    assert np.allclose(lp__, adjusted_lp)