perf(analyzer): cache embedded arguments and some more little perf tweaks

Author: d-biehl

packages/robot/src/robotcode/robot/diagnostics/keyword_finder.py

@@ -1,3 +1,4 @@
+import functools
 import re
 from itertools import chain
 from typing import TYPE_CHECKING, Dict, Iterable, Iterator, List, NamedTuple, Optional, Sequence, Tuple
@@ -59,7 +60,6 @@ def __init__(self, namespace: "Namespace", library_doc: LibraryDoc) -> None:
         self._all_keywords: Optional[List[LibraryEntry]] = None
         self._resource_keywords: Optional[List[ResourceEntry]] = None
         self._library_keywords: Optional[List[LibraryEntry]] = None
-        self._bdd_prefix_regexp: Optional["re.Pattern[str]"] = None
 
     def reset_diagnostics(self) -> None:
         self.diagnostics = []
@@ -459,20 +459,18 @@ def _create_custom_and_standard_keyword_conflict_warning_message(
             f"or '{'' if standard[0] is None else standard[0].alias or standard[0].name}.{standard[1].name}'."
         )
 
-    @property
+    @functools.cached_property
     def bdd_prefix_regexp(self) -> "re.Pattern[str]":
-        if not self._bdd_prefix_regexp:
-            prefixes = (
-                "|".join(
-                    self.namespace.languages.bdd_prefixes
-                    if self.namespace.languages is not None
-                    else ["given", "when", "then", "and", "but"]
-                )
-                .replace(" ", r"\s")
-                .lower()
+        prefixes = (
+            "|".join(
+                self.namespace.languages.bdd_prefixes
+                if self.namespace.languages is not None
+                else ["given", "when", "then", "and", "but"]
             )
-            self._bdd_prefix_regexp = re.compile(rf"({prefixes})\s", re.IGNORECASE)
-        return self._bdd_prefix_regexp
+            .replace(" ", r"\s")
+            .lower()
+        )
+        return re.compile(rf"({prefixes})\s", re.IGNORECASE)
 
     def _get_bdd_style_keyword(self, name: str) -> Optional[KeywordDoc]:
         match = self.bdd_prefix_regexp.match(name)

packages/robot/src/robotcode/robot/diagnostics/library_doc.py

@@ -1,6 +1,7 @@
 from __future__ import annotations
 
 import ast
+import functools
 import hashlib
 import importlib
 import importlib.util
@@ -197,14 +198,29 @@ def convert_from_rest(text: str) -> str:
     return text
 
 
+if get_robot_version() >= (6, 0):
+
+    @functools.lru_cache(maxsize=None)
+    def _get_embedded_arguments(name: str) -> Any:
+        try:
+            return EmbeddedArguments.from_name(name)
+        except (VariableError, DataError):
+            return ()
+
+else:
+
+    @functools.lru_cache(maxsize=None)
+    def _get_embedded_arguments(name: str) -> Any:
+        try:
+            return EmbeddedArguments(name)
+        except (VariableError, DataError):
+            return ()
+
+
 def is_embedded_keyword(name: str) -> bool:
     try:
-        if get_robot_version() >= (6, 0):
-            if EmbeddedArguments.from_name(name):
-                return True
-        else:
-            if EmbeddedArguments(name):
-                return True
+        if _get_embedded_arguments(name):
+            return True
     except (VariableError, DataError):
         return True
 
@@ -235,18 +251,22 @@ def normalized_name(self) -> str:
     def embedded_arguments(self) -> Any:
         if self._embedded_arguments is None:
             if self._can_have_embedded:
-                try:
-                    if get_robot_version() >= (6, 0):
-                        self._embedded_arguments = EmbeddedArguments.from_name(self.name)
-                    else:
-                        self._embedded_arguments = EmbeddedArguments(self.name)
-                except (VariableError, DataError):
-                    self._embedded_arguments = ()
+                self._embedded_arguments = _get_embedded_arguments(self.name)
             else:
                 self._embedded_arguments = ()
 
         return self._embedded_arguments
 
+    if get_robot_version() >= (6, 0):
+
+        def __match_embedded(self, name: str) -> bool:
+            return self.embedded_arguments.match(name) is not None
+
+    else:
+
+        def __match_embedded(self, name: str) -> bool:
+            return self.embedded_arguments.name.match(name) is not None
+
     def __eq__(self, o: object) -> bool:
         if cached_isinstance(o, KeywordMatcher):
             if self._is_namespace != o._is_namespace:
@@ -261,10 +281,7 @@ def __eq__(self, o: object) -> bool:
             return False
 
         if self.embedded_arguments:
-            if get_robot_version() >= (6, 0):
-                return self.embedded_arguments.match(o) is not None
-
-            return self.embedded_arguments.name.match(o) is not None
+            return self.__match_embedded(o)
 
         return self.normalized_name == str(normalize_namespace(o) if self._is_namespace else normalize(o))
 
@@ -935,8 +952,6 @@ def __getitem__(self, key: str) -> KeywordDoc:
         )
 
     def __contains__(self, _x: object) -> bool:
-        if not isinstance(_x, KeywordMatcher):
-            _x = KeywordMatcher(str(_x))
         return any(k == _x for k in self._matchers.keys())
 
     def __len__(self) -> int:

packages/robot/src/robotcode/robot/diagnostics/namespace.py

@@ -893,6 +893,7 @@ def get_namespaces(self) -> Dict[KeywordMatcher, List[LibraryEntry]]:
                 self._namespaces[KeywordMatcher(v.alias or v.name or v.import_name, is_namespace=True)].append(v)
             for v in (self.get_resources()).values():
                 self._namespaces[KeywordMatcher(v.alias or v.name or v.import_name, is_namespace=True)].append(v)
+
         return self._namespaces
 
     def get_resources(self) -> Dict[str, ResourceEntry]:
@@ -1793,11 +1794,10 @@ def iter_all_keywords(self) -> Iterator[KeywordDoc]:
 
         libdoc = self.get_library_doc()
 
-        for doc in itertools.chain(
+        yield from itertools.chain(
             self.get_imported_keywords(),
             libdoc.keywords if libdoc is not None else [],
-        ):
-            yield doc
+        )
 
     @_logger.call
     def get_keywords(self) -> List[KeywordDoc]:

packages/robot/src/robotcode/robot/diagnostics/namespace_analyzer.py

@@ -531,6 +531,20 @@ def _append_diagnostics(
             )
         )
 
+    KEYWORDS_WITH_EXPRESSIONS = [
+        "BuiltIn.Evaluate",
+        "BuiltIn.Should Be True",
+        "BuiltIn.Should Not Be True",
+        "BuiltIn.Skip If",
+        "BuiltIn.Continue For Loop If",
+        "BuiltIn.Exit For Loop If",
+        "BuiltIn.Return From Keyword If",
+        "BuiltIn.Run Keyword And Return If",
+        "BuiltIn.Pass Execution If",
+        "BuiltIn.Run Keyword If",
+        "BuiltIn.Run Keyword Unless",
+    ]
+
     def _analyze_keyword_call(
         self,
         node: ast.AST,
@@ -708,19 +722,7 @@ def _analyze_keyword_call(
             )
 
         if result is not None:
-            if result.longname in [
-                "BuiltIn.Evaluate",
-                "BuiltIn.Should Be True",
-                "BuiltIn.Should Not Be True",
-                "BuiltIn.Skip If",
-                "BuiltIn.Continue For Loop If",
-                "BuiltIn.Exit For Loop If",
-                "BuiltIn.Return From Keyword If",
-                "BuiltIn.Run Keyword And Return If",
-                "BuiltIn.Pass Execution If",
-                "BuiltIn.Run Keyword If",
-                "BuiltIn.Run Keyword Unless",
-            ]:
+            if result.longname in self.KEYWORDS_WITH_EXPRESSIONS:
                 tokens = argument_tokens
                 if tokens and (token := tokens[0]):
                     self._analyze_token_expression_variables(token)

packages/robot/src/robotcode/robot/utils/match.py

@@ -3,13 +3,13 @@
 _transform_table = str.maketrans("", "", "_ ")
 
 
-@lru_cache(maxsize=5000)
+@lru_cache(maxsize=None)
 def normalize(text: str) -> str:
     # return text.lower().replace("_", "").replace(" ", "")
     return text.casefold().translate(_transform_table)
 
 
-@lru_cache(maxsize=5000)
+@lru_cache(maxsize=None)
 def normalize_namespace(text: str) -> str:
     return text.lower().replace(" ", "")