From a19773873a08de694411024a21871bdf02332bae Mon Sep 17 00:00:00 2001
From: Tomas Babej <tomas@tbabej.com>
Date: Fri, 23 Apr 2021 22:35:04 -0400
Subject: [PATCH] performance: Cache CLI2::canonicalize lookups

Canonicalization of attributes or attribute prefixes happens once per
every single task considered in the filtering process. It maps short,
prefixed names to their full-length versions, for example:

"pro" -> "project" in the expression "project:Home"

It follows that such process is not task-dependant and as such the
lookups can be cached.
---
 src/CLI2.cpp | 15 ++++++++++++++-
 src/CLI2.h   |  4 +++-
 2 files changed, 17 insertions(+), 2 deletions(-)

diff --git a/src/CLI2.cpp b/src/CLI2.cpp
index 53d8366e2..cd57e4dbd 100644
--- a/src/CLI2.cpp
+++ b/src/CLI2.cpp
@@ -699,8 +699,19 @@ const std::vector <std::string> CLI2::getWords ()
 bool CLI2::canonicalize (
   std::string& canonicalized,
   const std::string& category,
-  const std::string& value) const
+  const std::string& value)
 {
+  // Utilize a cache mapping of (category, value) -> canonicalized value.
+  // This cache does not need to be invalidated, because entities are defined
+  // only once per initialization of the Context object.
+  int cache_key = 31 * std::hash<std::string>{} (category) + std::hash<std::string>{} (value);
+  auto cache_result = _canonical_cache.find (cache_key);
+  if (cache_result != _canonical_cache.end())
+  {
+    canonicalized = cache_result->second;
+    return true;
+  }
+
   // Extract a list of entities for category.
   std::vector <std::string> options;
   auto c = _entities.equal_range (category);
@@ -710,6 +721,7 @@ bool CLI2::canonicalize (
     if (value == e->second)
     {
       canonicalized = value;
+      _canonical_cache[cache_key] = value;
       return true;
     }
 
@@ -721,6 +733,7 @@ bool CLI2::canonicalize (
   if (autoComplete (value, options, matches, minimumMatchLength) == 1)
   {
     canonicalized = matches[0];
+    _canonical_cache[cache_key] = matches[0];
     return true;
   }
 
diff --git a/src/CLI2.h b/src/CLI2.h
index 6a276a758..b785c9f04 100644
--- a/src/CLI2.h
+++ b/src/CLI2.h
@@ -29,6 +29,7 @@
 #include <string>
 #include <vector>
 #include <map>
+#include <unordered_map>
 #include <Lexer.h>
 #include <FS.h>
 
@@ -77,7 +78,7 @@ public:
   void addContext (bool readable, bool writeable);
   void prepareFilter ();
   const std::vector <std::string> getWords ();
-  bool canonicalize (std::string&, const std::string&, const std::string&) const;
+  bool canonicalize (std::string&, const std::string&, const std::string&);
   std::string getBinary () const;
   std::string getCommand (bool canonical = true) const;
   const std::string dump (const std::string& title = "CLI2 Parser") const;
@@ -109,6 +110,7 @@ private:
 public:
   std::multimap <std::string, std::string>           _entities             {};
   std::map <std::string, std::string>                _aliases              {};
+  std::unordered_map <int, std::string>              _canonical_cache      {};
   std::vector <A2>                                   _original_args        {};
   std::vector <A2>                                   _args                 {};