_loader/types: cimport yaml functions for better speed

- _yaml: export node_validate function as Cython, as it was not done
before. This requires rewriting the function to remove a closure.
- Optimize node check by not calling is_node().

4 files changed, 17 insertions, 15 deletions

diff --git a/setup.py b/setup.py
index e393096b8..ab3c6f30d 100755
--- a/setup.py
+++ b/setup.py
@@ -399,7 +399,7 @@ def register_cython_module(module_name, dependencies=None):
 BUILD_EXTENSIONS = []
 
 register_cython_module("buildstream._loader._loader")
-register_cython_module("buildstream._loader.types")
+register_cython_module("buildstream._loader.types", dependencies=["buildstream._yaml"])
 register_cython_module("buildstream._yaml")
 register_cython_module("buildstream._variables", dependencies=["buildstream._yaml"])
 
diff --git a/src/buildstream/_loader/types.pyx b/src/buildstream/_loader/types.pyx
index a9aa3ab32..5e4e12666 100644
--- a/src/buildstream/_loader/types.pyx
+++ b/src/buildstream/_loader/types.pyx
@@ -18,7 +18,7 @@
 #        Tristan Van Berkom <tristan.vanberkom@codethink.co.uk>
 
 from .._exceptions import LoadError, LoadErrorReason
-from .. import _yaml
+from .. cimport _yaml
 
 
 # Symbol():
@@ -51,14 +51,14 @@ class Symbol():
 # A simple object describing a dependency
 #
 # Args:
-#    name (str): The element name
+#    name (str or Node): The element name
 #    dep_type (str): The type of dependency, can be
 #                    Symbol.ALL, Symbol.BUILD, or Symbol.RUNTIME
 #    junction (str): The element name of the junction, or None
-#    provenance (Provenance): The YAML node provenance of where this
-#                             dependency was declared
+#    provenance (ProvenanceInformation): The YAML node provenance of where this
+#                                        dependency was declared
 #
-class Dependency():
+class Dependency:
     def __init__(self, dep, provenance, default_dep_type=None):
         self.provenance = provenance
 
@@ -67,7 +67,7 @@ class Dependency():
             self.dep_type = default_dep_type
             self.junction = None
 
-        elif _yaml.is_node(dep):
+        elif type(dep) is _yaml.Node and type(dep.value) is dict:
             if default_dep_type:
                 _yaml.node_validate(dep, ['filename', 'junction'])
                 dep_type = default_dep_type
@@ -75,7 +75,7 @@ class Dependency():
                 _yaml.node_validate(dep, ['filename', 'type', 'junction'])
 
                 # Make type optional, for this we set it to None
-                dep_type = _yaml.node_get(dep, str, Symbol.TYPE, default_value=None)
+                dep_type = _yaml.node_get(dep, str, Symbol.TYPE, None, None)
                 if dep_type is None or dep_type == Symbol.ALL:
                     dep_type = None
                 elif dep_type not in [Symbol.BUILD, Symbol.RUNTIME]:
@@ -86,7 +86,7 @@ class Dependency():
 
             self.name = _yaml.node_get(dep, str, Symbol.FILENAME)
             self.dep_type = dep_type
-            self.junction = _yaml.node_get(dep, str, Symbol.JUNCTION, default_value=None)
+            self.junction = _yaml.node_get(dep, str, Symbol.JUNCTION, None, None)
 
         else:
             raise LoadError(LoadErrorReason.INVALID_DATA,
diff --git a/src/buildstream/_yaml.pxd b/src/buildstream/_yaml.pxd
index 27a1a888e..6a12fa7b3 100644
--- a/src/buildstream/_yaml.pxd
+++ b/src/buildstream/_yaml.pxd
@@ -39,6 +39,7 @@ cdef class ProvenanceInformation:
 
 
 cpdef object node_get(Node node, object expected_type, str key, list indices=*, object default_value=*, bint allow_none=*)
+cpdef void node_validate(Node node, list valid_keys) except *
 cpdef void node_set(Node node, object key, object value, list indices=*) except *
 cpdef list node_keys(object node)
 cpdef ProvenanceInformation node_get_provenance(Node node, str key=*, list indices=*)
diff --git a/src/buildstream/_yaml.pyx b/src/buildstream/_yaml.pyx
index 4505e2f95..f14b55e51 100644
--- a/src/buildstream/_yaml.pyx
+++ b/src/buildstream/_yaml.pyx
@@ -1119,16 +1119,17 @@ cpdef object node_sanitize(object node, object dict_type=OrderedDict):
 #    LoadError: In the case that the specified node contained
 #               one or more invalid keys
 #
-def node_validate(Node node, list valid_keys):
+cpdef void node_validate(Node node, list valid_keys) except *:
 
     # Probably the fastest way to do this: https://stackoverflow.com/a/23062482
     cdef set valid_keys_set = set(valid_keys)
-    invalid = next((key for key in node.value if key not in valid_keys_set), None)
+    cdef str key
 
-    if invalid:
-        provenance = node_get_provenance(node, key=invalid)
-        raise LoadError(LoadErrorReason.INVALID_DATA,
-                        "{}: Unexpected key: {}".format(provenance, invalid))
+    for key in node.value:
+        if key not in valid_keys_set:
+            provenance = node_get_provenance(node, key=key)
+            raise LoadError(LoadErrorReason.INVALID_DATA,
+                            "{}: Unexpected key: {}".format(provenance, key))
 
 
 # Node copying