fix: Tensor conversion for expected aux and actual data (#1038)

* Make args of Model.expected_actualdata and Model.expected_auxdata tensors
   - Gives parity with Model.expected_data
* Add tests for Model object interaction with own API

Author: alexander-held

src/pyhf/pdf.py

@@ -609,6 +609,8 @@ def expected_auxdata(self, pars):
             Tensor: The expected data of the auxiliary pdf
 
         """
+        tensorlib, _ = get_backend()
+        pars = tensorlib.astensor(pars)
         return self.make_pdf(pars)[1].expected_data()
 
     def _modifications(self, pars):
@@ -630,6 +632,8 @@ def expected_actualdata(self, pars):
             Tensor: The expected data of the main model (no auxiliary data)
 
         """
+        tensorlib, _ = get_backend()
+        pars = tensorlib.astensor(pars)
         return self.make_pdf(pars)[0].expected_data()
 
     def expected_data(self, pars, include_auxdata=True):

tests/test_pdf.py

@@ -133,6 +133,13 @@ def test_pdf_basicapi_tests(backend):
         [60.0, 51.020408630], 1e-08
     )
 
+    assert tensorlib.tolist(pdf.expected_actualdata(pars)) == pytest.approx(
+        [60.0], 1e-08
+    )
+    assert tensorlib.tolist(pdf.expected_auxdata(pars)) == pytest.approx(
+        [51.020408630], 1e-08
+    )
+
     pdf = pyhf.simplemodels.hepdata_like(
         source['bindata']['sig'],
         source['bindata']['bkg'],