Optimize headers even more (#1785)

Follow-up to #1781, eliminates some dead code and optimizes includes again by moving the very expensive include of `RuntimeParameters.h` into the cpp files.

src/engine/CMakeLists.txt

@@ -14,5 +14,6 @@ add_library(engine
         CartesianProductJoin.cpp TextIndexScanForWord.cpp TextIndexScanForEntity.cpp
         TextLimit.cpp LazyGroupBy.cpp GroupByHashMapOptimization.cpp SpatialJoin.cpp
         CountConnectedSubgraphs.cpp SpatialJoinAlgorithms.cpp PathSearch.cpp ExecuteUpdate.cpp
-        Describe.cpp GraphStoreProtocol.cpp)
+        Describe.cpp GraphStoreProtocol.cpp
+        QueryExecutionContext.cpp)
 qlever_target_link_libraries(engine util index parser sparqlExpressions http SortPerformanceEstimator Boost::iostreams s2)

src/engine/OptionalJoin.h

@@ -58,14 +58,8 @@ class OptionalJoin : public Operation {
     return {_left.get(), _right.get()};
   }
 
-  /**
-   * @brief Joins two result tables on any number of columns, inserting the
-   *        special value ID_NO_VALUE for any entries marked as optional.
-   * @param a
-   * @param b
-   * @param joinColumns
-   * @param result
-   */
+  // Joins two result tables on any number of columns, inserting the special
+  // value `Id::makeUndefined()` for any entries marked as optional.
   void optionalJoin(
       const IdTable& left, const IdTable& right,
       const std::vector<std::array<ColumnIndex, 2>>& joinColumns,

src/engine/QueryExecutionContext.cpp

@@ -0,0 +1,11 @@
+//   Copyright 2025, University of Freiburg,
+//   Chair of Algorithms and Data Structures.
+//   Author: Robin Textor-Falconi <textorr@informatik.uni-freiburg.de>
+
+#include "engine/QueryExecutionContext.h"
+
+#include "global/RuntimeParameters.h"
+
+bool QueryExecutionContext::areWebSocketUpdatesEnabled() {
+  return RuntimeParameters().get<"websocket-updates-enabled">();
+}

src/engine/QueryExecutionContext.h

@@ -5,8 +5,6 @@
 
 #pragma once
 
-#include <global/RuntimeParameters.h>
-
 #include <memory>
 #include <string>
 
@@ -19,7 +17,6 @@
 #include "index/Index.h"
 #include "util/Cache.h"
 #include "util/ConcurrentCache.h"
-#include "util/Synchronized.h"
 
 // The value of the `QueryResultCache` below. It consists of a `Result` together
 // with its `RuntimeInfo`.
@@ -151,6 +148,9 @@ class QueryExecutionContext {
     return areWebsocketUpdatesEnabled_;
   }
 
+ private:
+  static bool areWebSocketUpdatesEnabled();
+
  private:
   const Index& _index;
 
@@ -168,6 +168,5 @@ class QueryExecutionContext {
   std::function<void(std::string)> updateCallback_;
   // Cache the state of that runtime parameter to reduce the contention of the
   // mutex.
-  bool areWebsocketUpdatesEnabled_ =
-      RuntimeParameters().get<"websocket-updates-enabled">();
+  bool areWebsocketUpdatesEnabled_ = areWebSocketUpdatesEnabled();
 };

src/engine/QueryExecutionTree.cpp

@@ -12,7 +12,7 @@
 #include <vector>
 
 #include "engine/Sort.h"
-#include "parser/RdfEscaping.h"
+#include "global/RuntimeParameters.h"
 
 using std::string;
 

src/global/Id.h

@@ -3,16 +3,12 @@
 // Author: Björn Buchhold (buchhold@informatik.uni-freiburg.de)
 #pragma once
 
-#include <absl/strings/str_cat.h>
-
 #include <cstdint>
-#include <limits>
 
 #include "global/ValueId.h"
-#include "util/Exception.h"
 
 using Id = ValueId;
-typedef uint16_t Score;
+using Score = uint16_t;
 
 // TODO<joka921> Make the following ID and index types strong.
 using ColumnIndex = uint64_t;
@@ -23,92 +19,3 @@ using WordIndex = uint64_t;
 using WordOrEntityIndex = uint64_t;
 using TextBlockIndex = uint64_t;
 using CompressionCode = uint64_t;
-
-// Integers, that are probably not integers but strong IDs or indices, but their
-// true nature is still to be discovered.
-using UnknownIndex = uint64_t;
-
-// A value to use when the result should be empty (e.g. due to an optional join)
-// The highest two values are used as sentinels.
-static const Id ID_NO_VALUE = Id::makeUndefined();
-
-namespace ad_utility {
-
-// An exception that is thrown when an integer overflow occurs in the
-// `MilestoneIdManager`
-class MilestoneIdOverflowException : public std::exception {
- private:
-  std::string _message;
-
- public:
-  explicit MilestoneIdOverflowException(std::string message)
-      : _message{std::move(message)} {}
-  [[nodiscard]] const char* what() const noexcept override {
-    return _message.c_str();
-  }
-};
-
-// Manages two kinds of IDs: plain IDs (unsigned 64-bit integers, just called
-// "IDs" in the following), and milestone IDs (unsigned 64-bit integers that are
-// multiples of `distanceBetweenMilestones`. This class has the functionality to
-// find the next milestone of plain ID, to check whether an ID is amilestone ID
-// and to convert milestone IDs from and to a local ID space.
-template <size_t distanceBetweenMilestones>
-class MilestoneIdManager {
- private:
-  // The next free ID;
-  uint64_t _nextId{0};
-  // The last ID that was assigned. Used for overflow detection.
-  uint64_t _previousId{0};
-
- public:
-  MilestoneIdManager() = default;
-
-  // The maximum number of milestone Ids.
-  constexpr static uint64_t numMilestoneIds =
-      std::numeric_limits<uint64_t>::max() / distanceBetweenMilestones;
-
-  // Get the smallest milestone ID that is larger than all (milestone and
-  // non-milestone) previously obtained IDs.
-  uint64_t getNextMilestoneId() {
-    if (!isMilestoneId(_nextId)) {
-      _nextId = (milestoneIdFromLocal(milestoneIdToLocal(_nextId) + 1));
-    }
-    return getNextId();
-  }
-
-  // Get the smallest ID that is larger than all previously obtained IDs.
-  uint64_t getNextId() {
-    if (_nextId < _previousId) {
-      throw MilestoneIdOverflowException{absl::StrCat(
-          "Overflow while assigning Ids from a MilestoneIdManager. The "
-          "previous "
-          "milestone Id was ",
-          _previousId, " the next id would be ", _nextId,
-          ". The maximum number of milestones is ", numMilestoneIds, ".")};
-    }
-    _previousId = _nextId;
-    _nextId++;
-    return _previousId;
-  }
-
-  // Is this ID a milestone id, equivalently: Is this ID a multiple of
-  // `distanceBetweenMilestones`?
-  constexpr static bool isMilestoneId(uint64_t id) {
-    return id % distanceBetweenMilestones == 0;
-  }
-
-  // Convert a milestone ID to its "local" ID by dividing it by
-  // `distanceBetweenMilestones` (the i-th milestone ID will become `i`).
-  constexpr static uint64_t milestoneIdToLocal(uint64_t id) {
-    return id / distanceBetweenMilestones;
-  }
-
-  // Convert "local" ID to milestone ID by multiplying it with
-  // `distanceBetweenMilestones`.
-  constexpr static uint64_t milestoneIdFromLocal(uint64_t id) {
-    return id * distanceBetweenMilestones;
-  }
-};
-
-}  // namespace ad_utility

src/index/VocabularyMerger.h

@@ -68,8 +68,8 @@ struct VocabularyMetaData {
     bool contains(Id id) const { return begin_ <= id && id < end_; }
 
    private:
-    Id begin_ = ID_NO_VALUE;
-    Id end_ = ID_NO_VALUE;
+    Id begin_ = Id::makeUndefined();
+    Id end_ = Id::makeUndefined();
     std::string prefix_;
     bool beginWasSeen_ = false;
   };

src/util/AllocatorWithLimit.h

@@ -7,7 +7,6 @@
 
 #include <absl/strings/str_cat.h>
 
-#include <atomic>
 #include <functional>
 #include <memory>
 

test/CMakeLists.txt

@@ -254,8 +254,6 @@ addLinkAndDiscoverTest(ContentEncodingHelperTest http)
 
 addLinkAndDiscoverTest(PrefixCompressorTest)
 
-addLinkAndDiscoverTest(MilestoneIdTest)
-
 addLinkAndDiscoverTest(VocabularyTest index)
 
 addLinkAndDiscoverTestNoLibs(IteratorTest)

test/GroupByTest.cpp

@@ -263,9 +263,9 @@ TEST_F(GroupByTest, doGroupBy) {
   ASSERT_EQ(123u, outTable._data[1][9]);
   ASSERT_EQ(0u, outTable._data[2][9]);
 
-  ASSERT_EQ(ID_NO_VALUE, outTable._data[0][10]);
-  ASSERT_EQ(ID_NO_VALUE, outTable._data[1][10]);
-  ASSERT_EQ(ID_NO_VALUE, outTable._data[2][10]);
+  ASSERT_EQ(Id::makeUndefined(), outTable._data[0][10]);
+  ASSERT_EQ(Id::makeUndefined(), outTable._data[1][10]);
+  ASSERT_EQ(Id::makeUndefined(), outTable._data[2][10]);
 
   std::memcpy(&buffer, &outTable._data[0][11], sizeof(float));
   ASSERT_FLOAT_EQ(-3, buffer);
@@ -283,9 +283,9 @@ TEST_F(GroupByTest, doGroupBy) {
   ASSERT_EQ(41223u, outTable._data[1][13]);
   ASSERT_EQ(41223u, outTable._data[2][13]);
 
-  ASSERT_EQ(ID_NO_VALUE, outTable._data[0][14]);
-  ASSERT_EQ(ID_NO_VALUE, outTable._data[1][14]);
-  ASSERT_EQ(ID_NO_VALUE, outTable._data[2][14]);
+  ASSERT_EQ(Id::makeUndefined(), outTable._data[0][14]);
+  ASSERT_EQ(Id::makeUndefined(), outTable._data[1][14]);
+  ASSERT_EQ(Id::makeUndefined(), outTable._data[2][14]);
 
   std::memcpy(&buffer, &outTable._data[0][15], sizeof(float));
   ASSERT_FLOAT_EQ(2, buffer);