[GR-39984] Performance issues with regex and nltk modules.

PullRequest: graalpython/2354

Author: lukasstadler

graalpython/com.oracle.graal.python/src/com/oracle/graal/python/builtins/modules/BinasciiModuleBuiltins.java

@@ -127,7 +127,6 @@ int getBufferLength(
             }
 
             @ExportMessage
-            @TruffleBoundary
             byte readByte(int byteOffset,
                             @Cached TruffleString.CodePointAtIndexNode codePointAtIndexNode) {
                 int ch = codePointAtIndexNode.execute(str, byteOffset, TS_ENCODING);

graalpython/com.oracle.graal.python/src/com/oracle/graal/python/builtins/objects/common/EconomicMapStorage.java

@@ -97,6 +97,12 @@ static final class DictKey {
         public int hashCode() {
             return (int) hash;
         }
+
+        @Override
+        public String toString() {
+            CompilerAsserts.neverPartOfCompilation();
+            return value + "@" + hash;
+        }
     }
 
     private final PEMap map;

graalpython/com.oracle.graal.python/src/com/oracle/graal/python/nodes/call/special/CallBinaryMethodNode.java

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2018, 2021, Oracle and/or its affiliates. All rights reserved.
+ * Copyright (c) 2018, 2022, Oracle and/or its affiliates. All rights reserved.
  * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
  *
  * The Universal Permissive License (UPL), Version 1.0
@@ -201,6 +201,17 @@ static Object callSelfMethod(VirtualFrame frame, @SuppressWarnings("unused") PBu
         return builtinNode.execute(frame, func.getSelf(), arg1, arg2);
     }
 
+    @Specialization(guards = {"builtinNode != null", "getCallTarget(func, getCt) == ct", "!takesSelfArg",
+                    "frame != null || unusedFrame"}, limit = "getCallSiteInlineCacheMaxDepth()")
+    static Object callMethod(VirtualFrame frame, @SuppressWarnings("unused") PBuiltinMethod func, Object arg1, Object arg2,
+                    @SuppressWarnings("unused") @Cached GetCallTargetNode getCt,
+                    @SuppressWarnings("unused") @Cached("getCallTarget(func, getCt)") RootCallTarget ct,
+                    @SuppressWarnings("unused") @Cached("takesSelfArg(func)") boolean takesSelfArg,
+                    @Cached("getTernary(frame, func.getFunction())") PythonTernaryBuiltinNode builtinNode,
+                    @SuppressWarnings("unused") @Cached("frameIsUnused(builtinNode)") boolean unusedFrame) {
+        return builtinNode.execute(frame, arg1, arg2, PNone.NO_VALUE);
+    }
+
     /**
      * In case the function takes less or equal to 2 arguments (so it is <it>at least</it> binary)
      * we also try to call a ternary function.

graalpython/lib-graalpython/_sre.py

@@ -170,7 +170,7 @@ def setup(sre_compiler, error_class, flags_table):
     (FLAG_DOTALL, "DOTALL"),
     (FLAG_UNICODE, "UNICODE"),
     (FLAG_VERBOSE, "VERBOSE"),
-    (FLAG_DEBUG, "DEBUG"),
+    # (FLAG_DEBUG, "DEBUG"), # there is no DEBUG flag in tregex
     (FLAG_ASCII, "ASCII"),
 ]
 
@@ -311,6 +311,7 @@ def __init__(self, pattern, flags):
                 flags_str.append(char)
         self.__flags_str = "".join(flags_str)
         self.__compiled_regexes = {}
+        self.__cached_flags = None
         compiled_regex = self.__tregex_compile()
         self.groups = compiled_regex.groupCount - 1
         groups = compiled_regex.groups
@@ -331,6 +332,8 @@ def __init__(self, pattern, flags):
     @property
     def flags(self):
         # Flags can be spcified both in the flag argument or inline in the regex. Extract them back from the regex
+        if self.__cached_flags != None:
+            return self.__cached_flags
         flags = self.__input_flags
         regex_flags = self.__tregex_compile().flags
         for flag, name in FLAG_NAMES:
@@ -339,6 +342,7 @@ def flags(self):
                     flags |= flag
             except AttributeError:
                 pass
+        self.__cached_flags = flags
         return flags
 
     def __check_input_type(self, input):