Mapping function plugin system (#152)

# Description:
Closes #128 

See #128 for design and reasoning behind this PR. Needed for using
mapping functions in OTEAPI pipelines.

## Type of change:
- [ ] Bug fix.
- [x] New feature.
- [ ] Documentation update.
- [ ] Testing.

tests/conftest.py

@@ -88,12 +88,15 @@ def expected_function_triplestore(
 @prefix ex: <http://example.com/onto#> .
 @prefix fno: <https://w3id.org/function/ontology#> .
 @prefix map: <http://emmo.info/domain-mappings#> .
+@prefix oteio: <http://emmo.info/oteio#> .
 @prefix owl: <http://www.w3.org/2002/07/owl#> .
 @prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
 @prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
 
 ex:sum__{fid} a fno:Function ;
     rdfs:label "sum_"@en ;
+    oteio:hasPythonFunctionName "sum_" ;
+    oteio:hasPythonModuleName "conftest" ;
     dcterms:description "Returns the sum of `first_param` and `second_param`."@en ;
     fno:expects ( ex:sum__{fid}_parameter1_first_param ex:sum__{fid}_parameter2_second_param ) ;
     fno:returns ( ex:sum__{fid}_output1 ) .

tests/mappings/test_cost.py

@@ -0,0 +1,29 @@
+"""Test cost."""
+import pytest
+
+pytest.importorskip("rdflib")
+
+# pylint: disable=wrong-import-position
+from tripper import Triplestore
+
+ts = Triplestore(backend="rdflib")
+
+# Define some prefixed namespaces
+CHEM = ts.bind("chem", "http://onto-ns.com/onto/chemistry#")
+MOL = ts.bind("mol", "http://onto-ns.com/meta/0.1/Molecule#")
+SUB = ts.bind("sub", "http://onto-ns.com/meta/0.1/Substance#")
+
+
+def formula_cost(ts, input_iris, output_iri):
+    """Returns a cost."""
+    # pylint: disable=unused-argument
+    return 2.72
+
+
+# Add mappings from data models to ontology
+ts.map(MOL.name, CHEM.Identifier, cost=3.14)
+ts.map(SUB.formula, CHEM.Formula, cost=formula_cost)
+
+# pylint: disable=protected-access
+assert ts._get_cost(CHEM.Identifier) == 3.14
+assert ts._get_cost(CHEM.Formula) == 2.72

tests/test_add_function.py

@@ -31,11 +31,14 @@ def func(a, b):
 @prefix ex: <http://example.com/ex#> .
 @prefix fno: <https://w3id.org/function/ontology#> .
 @prefix map: <http://emmo.info/domain-mappings#> .
+@prefix oteio: <http://emmo.info/oteio#> .
 @prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
 @prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
 
 <:func_{f_id}> a fno:Function ;
     rdfs:label "func"@en ;
+    oteio:hasPythonFunctionName "func" ;
+    oteio:hasPythonModuleName "test_add_function" ;
     dcterms:description "Returns the sum of `a` and `b`."@en ;
     fno:expects ( <:func_{f_id}_parameter1_a> <:func_{f_id}_parameter2_b> ) ;
     fno:returns ( <:func_{f_id}_output1> ) .
@@ -67,13 +70,16 @@ def func(a, b):
 @prefix dcterms: <http://purl.org/dc/terms/> .
 @prefix emmo: <http://emmo.info/emmo#> .
 @prefix ex: <http://example.com/ex#> .
+@prefix oteio: <http://emmo.info/oteio#> .
 @prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
 
 <:func_{f_id}> a emmo:EMMO_4299e344_a321_4ef2_a744_bacfcce80afc ;
     rdfs:label "func"@en ;
     emmo:EMMO_36e69413_8c59_4799_946c_10b05d266e22 ex:arg1,
         ex:arg2 ;
     emmo:EMMO_c4bace1d_4db0_4cd3_87e9_18122bae2840 ex:sum ;
+    oteio:hasPythonFunctionName "func" ;
+    oteio:hasPythonModuleName "test_add_function" ;
     dcterms:description "Returns the sum of `a` and `b`."@en .
 
 ex:arg1 a emmo:EMMO_194e367c_9783_4bf5_96d0_9ad597d48d9a ;
@@ -100,13 +106,16 @@ def func(a, b):
 @prefix dcterms: <http://purl.org/dc/terms/> .
 @prefix emmo: <http://emmo.info/emmo#> .
 @prefix ex: <http://example.com/ex#> .
+@prefix oteio: <http://emmo.info/oteio#> .
 @prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
 
 <:func_{f_id}> a emmo:EMMO_4299e344_a321_4ef2_a744_bacfcce80afc ;
     rdfs:label "func"@en ;
     emmo:EMMO_36e69413_8c59_4799_946c_10b05d266e22 ex:arg1,
         ex:arg2 ;
     emmo:EMMO_c4bace1d_4db0_4cd3_87e9_18122bae2840 ex:sum ;
+    oteio:hasPythonFunctionName "func" ;
+    oteio:hasPythonModuleName "test_add_function" ;
     dcterms:description "Returns the sum of `a` and `b`."@en .
 
 ex:arg1 a emmo:EMMO_194e367c_9783_4bf5_96d0_9ad597d48d9a ;

tests/test_eval_function.py

@@ -0,0 +1,53 @@
+"""Test Triplestore.eval_function()"""
+import pytest
+
+from tripper import Triplestore
+
+pytest.importorskip("rdflib")
+
+
+def func(a, b):
+    """Returns the sum of `a` and `b`."""
+    # pylint: disable=invalid-name
+    return a + b
+
+
+ts = Triplestore(backend="rdflib")
+EX = ts.bind("ex", "http://example.com/ex#")
+
+# Test to add function in current scope
+iri = ts.add_function(
+    func,
+    expects=[EX.arg1, EX.arg2],
+    returns=EX.sum,
+    standard="emmo",
+)
+assert ts.eval_function(func_iri=iri, args=(2, 3)) == 5
+
+
+# Test to add a function from the standard library. The hashlib module
+# is not expected to be imported in the current scope
+iri2 = ts.add_function(
+    EX.shape256,
+    expects=[EX.Bytes],
+    returns=EX.ShakeVar,
+    func_name="shake_256",
+    module_name="hashlib",
+)
+shakevar = ts.eval_function(iri2, (b"a",))
+assert shakevar.hexdigest(4) == "867e2cb0"
+
+
+# Test to add a function from a pypi package.
+iri3 = ts.add_function(
+    EX.UFloat,
+    expects=[EX.Uncertainty],
+    returns=EX.UUID,
+    func_name="ufloat",
+    module_name="uncertainties",
+    # package_name="uncertainties",
+    pypi_package_name="uncertainties==3.1.7",
+)
+val = ts.eval_function(iri3, args=(1, 0.1))
+assert val.n == 1
+assert val.s == 0.1

tripper/errors.py

@@ -18,6 +18,10 @@ class NoSuchIRIError(NamespaceError):
     """Namespace has no such IRI."""
 
 
+class CannotGetFunctionError(TriplestoreError):
+    """Not able to get function documented in the triplestore."""
+
+
 # === Warnings ===
 class UnusedArgumentWarning(Warning):
     """Argument is unused."""

tripper/mappings/mappings.py

@@ -20,6 +20,7 @@
 from pint import Quantity  # remove
 
 from tripper import DM, EMMO, FNO, MAP, RDF, RDFS
+from tripper.errors import CannotGetFunctionError
 from tripper.triplestore import hasAccessFunction, hasDataValue
 from tripper.utils import parse_literal
 
@@ -889,17 +890,29 @@ def mapping_routes(
     soAFun = dict(triplestore.subject_objects(hasAccessFunction))
     soDVal = dict(triplestore.subject_objects(hasDataValue))
 
+    def getfunc(func_iri, default=None):
+        """Returns callable function corresponding to `func_iri`.
+        Raises CannotGetFunctionError if func_iri cannot be found."""
+        if func_iri is None:
+            return None
+        if func_iri in function_repo and function_repo[func_iri]:
+            return function_repo[func_iri]
+        try:
+            return (
+                triplestore._get_function(  # pylint: disable=protected-access
+                    func_iri
+                )
+            )
+        except CannotGetFunctionError:
+            return default
+
     def getcost(target, stepname):
         """Returns the cost assigned to IRI `target` for a mapping step
         of type `stepname`."""
         cost = soCost.get(target, default_costs[stepname])
-        if cost is None:
-            return None
-        return (
-            function_repo[cost]
-            if cost in function_repo
-            else float(parse_literal(cost))
-        )
+        if cost is None or callable(cost) or isinstance(cost, float):
+            return cost
+        return getfunc(cost, float(parse_literal(cost)))
 
     def walk(target, visited, step):
         """Walk backward in rdf graph from `node` to sources."""
@@ -914,7 +927,7 @@ def addnode(node, steptype, stepname):
             step.cost = getcost(target, stepname)
             if node in soAFun:
                 value = value_class(
-                    value=triplestore.function_repo[soAFun[node]],
+                    value=getfunc(soAFun[node]),
                     unit=soUnit.get(node),
                     iri=node,
                     property_iri=soInst.get(node),
@@ -962,10 +975,10 @@ def addnode(node, steptype, stepname):
             addnode(node, StepType.INV_SUBCLASSOF, "subClassOf")
 
         for fmap in function_mappers:
-            for func, input_iris in fmap(triplestore)[target]:
+            for func_iri, input_iris in fmap(triplestore)[target]:
                 step.steptype = StepType.FUNCTION
-                step.cost = getcost(func, "function")
-                step.function = function_repo.get(func)
+                step.cost = getcost(func_iri, "function")
+                step.function = getfunc(func_iri)
                 step.join_mode = True
                 for input_iri in input_iris:
                     step0 = mappingstep_class(

tripper/namespace.py

@@ -199,3 +199,4 @@ def __eq__(self, other):
 EMMO = Namespace("http://emmo.info/emmo#")
 MAP = Namespace("http://emmo.info/domain-mappings#")
 DM = Namespace("http://emmo.info/datamodel#")
+OTEIO = Namespace("http://emmo.info/oteio#")