Testing cosmosis connector with pure ccl.

tests/connector/cosmosis/test_cosmosis_module.py

@@ -105,15 +105,24 @@ def fixture_config_with_two_point_real() -> DataBlock:
     return result
 
 
-@pytest.fixture(name="config_with_two_point_harmonic")
-def fixture_config_with_two_point_harmonic() -> DataBlock:
+@pytest.fixture(name="config_with_two_point_harmonic", params=["default", "pure_ccl"])
+def fixture_config_with_two_point_harmonic(request) -> DataBlock:
     result = DataBlock()
+    likelihood_file = ""
+    if request.param == "default":
+        likelihood_file = (
+            "${FIRECROWN_DIR}/tests/likelihood/gauss_family/lkscript_two_point.py"
+        )
+    elif request.param == "pure_ccl":
+        likelihood_file = (
+            "${FIRECROWN_DIR}/tests/likelihood/"
+            "gauss_family/lkscript_two_point_pure_ccl.py"
+        )
+
     result.put_string(
         option_section,
         "Likelihood_source",
-        expandvars(
-            "${FIRECROWN_DIR}/tests/likelihood/gauss_family/lkscript_two_point.py"
-        ),
+        expandvars(likelihood_file),
     )
     result.put_string(
         option_section,

tests/likelihood/gauss_family/lkscript_two_point_pure_ccl.py

@@ -0,0 +1,59 @@
+"""
+Provides a trivial likelihood factory function for a ConstGaussian
+for testing purposes.
+"""
+
+import numpy as np
+import sacc
+
+from firecrown.likelihood.likelihood import NamedParameters
+from firecrown.likelihood.gaussian import ConstGaussian
+from firecrown.likelihood.two_point import TwoPoint
+from firecrown.likelihood.number_counts import (
+    NumberCounts,
+)
+from firecrown.modeling_tools import ModelingTools
+from firecrown.ccl_factory import (
+    CCLFactory,
+    PoweSpecAmplitudeParameter,
+    CCLCreationMode,
+)
+
+
+def build_likelihood(params: NamedParameters):
+    """Return a ConstGaussian (likelihood) object."""
+    sacc_data = sacc.Sacc()
+
+    src0 = NumberCounts(sacc_tracer="lens0")
+    z = np.linspace(0, 0.8, 50) + 0.05
+    dndz = np.exp(-0.5 * (z - 0.4) ** 2 / 0.02 / 0.02)
+    sacc_data.add_tracer("NZ", "lens0", z, dndz)
+
+    if params.get_string("projection") == "harmonic":
+        two_point = TwoPoint("galaxy_density_cl", source0=src0, source1=src0)
+        ells = np.unique(np.logspace(1, 3, 10).astype(np.int64))
+        Cells = np.random.normal(size=ells.shape[0])
+        sacc_data.add_ell_cl("galaxy_density_cl", "lens0", "lens0", ells, Cells)
+        cov = np.diag(np.ones_like(Cells) * 0.01)
+        sacc_data.add_covariance(cov)
+    elif params.get_string("projection") == "real":
+        two_point = TwoPoint("galaxy_density_xi", source0=src0, source1=src0)
+        thetas = np.linspace(0.0, np.pi, 10)
+        xis = np.random.normal(size=thetas.shape[0])
+        sacc_data.add_theta_xi("galaxy_density_xi", "lens0", "lens0", thetas, xis)
+        cov = np.diag(np.ones_like(xis) * 0.01)
+        sacc_data.add_covariance(cov)
+    else:
+        raise ValueError("Invalid projection")
+
+    likelihood = ConstGaussian(statistics=[two_point])
+    likelihood.read(sacc_data)
+
+    tools = ModelingTools(
+        ccl_factory=CCLFactory(
+            amplitude_parameter=PoweSpecAmplitudeParameter.SIGMA8,
+            creation_mode=CCLCreationMode.PURE_CCL_MODE,
+        )
+    )
+
+    return likelihood, tools