Make PlainTreeFiller more user friendly (#141)

* mv VectorConfig::Print() implementation to .cpp

* enable non-prepending output leaves names with branch name

* propagate examples .cpp into install/share

* enable cherry-picking fields in PlainTreeFiller (instead of ignoring)

* remove virtual from VectorConfig::Print()

* remove BranchReader infra-1.0 remnants

* remove rootlogon.C

* apply clang-format

Author: lubynets

core/BranchConfig.cpp

@@ -136,20 +136,49 @@ BranchConfig BranchConfig::CloneAndMerge(const BranchConfig& attached) const {
 }
 
 template<typename T>
-std::vector<std::string> VectorConfig<T>::SplitString(const std::string& input) {
-  std::vector<std::string> result;
-  std::vector<int> newlinepositions{-1};
-  int it{0};
-  while (it < std::string::npos) {
-    it = input.find("\n", it + 1);
-    newlinepositions.emplace_back(it);
+void VectorConfig<T>::Print() const {
+  if (map_.empty()) return;
+
+  auto print_row = [](const std::vector<std::pair<std::string, int>>& elements) {
+    for (const auto& el : elements) {
+      std::cout << std::left << std::setw(el.second) << std::setfill(' ') << el.first;
+    }
+    std::cout << std::endl;
+  };
+
+  int name_strlen{0};
+  for (const auto& entry : map_) {
+    name_strlen = std::max(name_strlen, (int) entry.first.length());
   }
-  newlinepositions.back() = input.size();
-  for (int ip = 0; ip < newlinepositions.size() - 1; ++ip) {
-    result.emplace_back(input.substr(newlinepositions.at(ip) + 1, newlinepositions.at(ip + 1) - newlinepositions.at(ip) - 1));
+  name_strlen += 4;
+
+  auto SplitString = [](const std::string& input) {
+    std::vector<std::string> result;
+    std::vector<int> newlinepositions{-1};
+    int it{0};
+    while (it < std::string::npos) {
+      it = input.find("\n", it + 1);
+      newlinepositions.emplace_back(it);
+    }
+    newlinepositions.back() = input.size();
+    for (int ip = 0; ip < newlinepositions.size() - 1; ++ip) {
+      result.emplace_back(input.substr(newlinepositions.at(ip) + 1, newlinepositions.at(ip + 1) - newlinepositions.at(ip) - 1));
+    }
+    return result;
+  };
+
+  print_row({{"Id", 10}, {"Name", name_strlen}, {"Info", 50}});
+  for (const auto& entry : map_) {
+    if (entry.second.title_.find("\n") == std::string::npos) {
+      print_row({{std::to_string(entry.second.id_), 10}, {entry.first, name_strlen}, {entry.second.title_, 50}});
+    } else {
+      auto est = SplitString(entry.second.title_);
+      print_row({{std::to_string(entry.second.id_), 10}, {entry.first, name_strlen}, {est.at(0), 50}});
+      for (int iest = 1; iest < est.size(); ++iest) {
+        print_row({{"", 10}, {"", name_strlen}, {est.at(iest), 50}});
+      }
+    }
   }
-
-  return result;
 }
 
 template<typename T>

core/BranchConfig.hpp

@@ -73,38 +73,9 @@ class VectorConfig {
   ANALYSISTREE_ATTR_NODISCARD virtual const MapType& GetMap() const { return map_; }
   ANALYSISTREE_ATTR_NODISCARD virtual ShortInt_t GetSize() const { return size_; }
 
-  virtual void Print() const {
-    if (map_.empty()) return;
-
-    auto print_row = [](const std::vector<std::pair<std::string, int>>& elements) {
-      for (const auto& el : elements) {
-        std::cout << std::left << std::setw(el.second) << std::setfill(' ') << el.first;
-      }
-      std::cout << std::endl;
-    };
-
-    int name_strlen{0};
-    for (const auto& entry : map_) {
-      name_strlen = std::max(name_strlen, (int) entry.first.length());
-    }
-    name_strlen += 4;
-
-    print_row({{"Id", 10}, {"Name", name_strlen}, {"Info", 50}});
-    for (const auto& entry : map_) {
-      if (entry.second.title_.find("\n") == std::string::npos) {
-        print_row({{std::to_string(entry.second.id_), 10}, {entry.first, name_strlen}, {entry.second.title_, 50}});
-      } else {
-        auto est = SplitString(entry.second.title_);
-        print_row({{std::to_string(entry.second.id_), 10}, {entry.first, name_strlen}, {est.at(0), 50}});
-        for (int iest = 1; iest < est.size(); ++iest) {
-          print_row({{"", 10}, {"", name_strlen}, {est.at(iest), 50}});
-        }
-      }
-    }
-  }
+  void Print() const;
 
  protected:
-  static std::vector<std::string> SplitString(const std::string& input);
   void RemoveField(const std::string& name, int id);
   MapType map_{};
   ShortInt_t size_{0};
@@ -127,7 +98,7 @@ class BranchConfig : public VectorConfig<int>, public VectorConfig<float>, publi
 
   BranchConfig(std::string name, DetType type, std::string title = "");
 
-  void Print() const override;
+  void Print() const;
 
   void PrintBranchId() const;
 

examples/CMakeLists.txt

@@ -69,4 +69,5 @@ foreach(EXE ${EXECUTABLES})
         add_executable(${EXE} ${EXE}.cpp)
         target_link_libraries(${EXE} AnalysisTreeBase AnalysisTreeInfra AnalysisTreeUser ${ROOT_LIBRARIES} EG)
         install (TARGETS ${EXE} RUNTIME DESTINATION bin)
+        install (FILES ${EXE}.cpp DESTINATION share)
 endforeach()

examples/rootlogon.C

@@ -17,7 +17,6 @@
 #pragma link C++ class AnalysisTree::AnalysisTask;
 #pragma link C++ class AnalysisTree::Chain+;
 
-#pragma link C++ class AnalysisTree::BranchReader+;
 #pragma link C++ class AnalysisTree::AnalysisEntry+;
 
 #endif

infra/AnalysisTreeInfraLinkDef.h

@@ -19,18 +19,18 @@ inline std::string ToStringWithPrecision(const T a_value, const int n) {
 
 template<typename T>
 inline std::string ToStringWithSignificantFigures(const T a_value, const int n) {
-  if(a_value == 0) return "0";
-
-  const double dMag = std::log10(std::abs(a_value)); // scale of the a_value (e.g 1.* for 1.2345, 2.* for 12.345 etc)
-  const int iMag = static_cast<int>(dMag-n+1 > 0 ? dMag-n+1 : dMag-n);
-  const T shifted_value = a_value/std::pow(10, iMag); // shift decimal point to have all required digits to l.h.s. from it
-  const T rounded_value = static_cast<T>(std::round(shifted_value)); // get rid of r.h.s. from decimal point
-  const T reshifted_value = rounded_value*std::pow(10, iMag); // return decimal point to its original place
-  const int precision = iMag < 0 ? -iMag : 0; // determine how many digits after decimal point one needs
+  if (a_value == 0) return "0";
+
+  const double dMag = std::log10(std::abs(a_value));// scale of the a_value (e.g 1.* for 1.2345, 2.* for 12.345 etc)
+  const int iMag = static_cast<int>(dMag - n + 1 > 0 ? dMag - n + 1 : dMag - n);
+  const T shifted_value = a_value / std::pow(10, iMag);             // shift decimal point to have all required digits to l.h.s. from it
+  const T rounded_value = static_cast<T>(std::round(shifted_value));// get rid of r.h.s. from decimal point
+  const T reshifted_value = rounded_value * std::pow(10, iMag);     // return decimal point to its original place
+  const int precision = iMag < 0 ? -iMag : 0;                       // determine how many digits after decimal point one needs
   return ToStringWithPrecision(reshifted_value, precision);
 }
 
-inline std::vector<AnalysisTree::SimpleCut> CreateRangeCuts(const std::vector<float>& ranges, const std::string& cutNamePrefix, const std::string& branchFieldName, int precision=2) {
+inline std::vector<AnalysisTree::SimpleCut> CreateRangeCuts(const std::vector<float>& ranges, const std::string& cutNamePrefix, const std::string& branchFieldName, int precision = 2) {
   std::vector<AnalysisTree::SimpleCut> sliceCuts;
   for (int iRange = 0; iRange < ranges.size() - 1; iRange++) {
     const std::string cutName = cutNamePrefix + ToStringWithPrecision(ranges.at(iRange), 2) + "_" + ToStringWithPrecision(ranges.at(iRange + 1), precision);
@@ -40,7 +40,7 @@ inline std::vector<AnalysisTree::SimpleCut> CreateRangeCuts(const std::vector<fl
   return sliceCuts;
 }
 
-inline std::vector<AnalysisTree::SimpleCut> CreateEqualCuts(const std::vector<float>& values, const std::string& cutNamePrefix, const std::string& branchFieldName, int precision=2) {
+inline std::vector<AnalysisTree::SimpleCut> CreateEqualCuts(const std::vector<float>& values, const std::string& cutNamePrefix, const std::string& branchFieldName, int precision = 2) {
   std::vector<AnalysisTree::SimpleCut> sliceCuts;
   for (int iValue = 0; iValue < values.size(); iValue++) {
     const std::string cutName = cutNamePrefix + ToStringWithPrecision(values.at(iValue), precision);

infra/HelperFunctions.hpp

@@ -15,7 +15,7 @@ void PlainTreeFiller::AddBranch(const std::string& branch_name) {
   in_branches_.emplace(branch_name);
 }
 
-void PlainTreeFiller::SetFieldsToIgnore(const std::vector<std::string>&& fields_to_ignore) {
+void PlainTreeFiller::SetFieldsToIgnore(const std::vector<std::string>& fields_to_ignore) {
   if (branch_name_ == "") {
     throw std::runtime_error("PlainTreeFiller::SetFieldsToIgnore() must be called after PlainTreeFiller::AddBranch()\n");
   }
@@ -24,6 +24,15 @@ void PlainTreeFiller::SetFieldsToIgnore(const std::vector<std::string>&& fields_
   }
 }
 
+void PlainTreeFiller::SetFieldsToPreserve(const std::vector<std::string>& fields_to_preserve) {
+  if (branch_name_ == "") {
+    throw std::runtime_error("PlainTreeFiller::SetFieldsToPreserve() must be called after PlainTreeFiller::AddBranch()\n");
+  }
+  for (auto& fti : fields_to_preserve) {
+    fields_to_preserve_.emplace_back((branch_name_ + "." + fti).c_str());
+  }
+}
+
 void PlainTreeFiller::Init() {
   if (is_ignore_defual_fields_) {
     std::vector<std::string> defaultFieldsNames;
@@ -38,6 +47,10 @@ void PlainTreeFiller::Init() {
     SetFieldsToIgnore(std::move(defaultFieldsNames));
   }
 
+  if (!fields_to_ignore_.empty() && !fields_to_preserve_.empty()) {
+    throw std::runtime_error("PlainTreeFiller::Init() - only one of fields_to_ignore_ and fields_to_preserve_ can be set");
+  }
+
   if (!branch_name_.empty()) {
     const auto& branch_config = config_->GetBranchConfig(branch_name_);
     for (const auto& field : branch_config.GetMap<float>()) {
@@ -64,7 +77,9 @@ void PlainTreeFiller::Init() {
   plain_tree_->SetAutoSave(0);
   for (size_t i = 0; i < vars.size(); ++i) {
     std::string leaf_name = vars[i].GetName();
-    if (std::find(fields_to_ignore_.begin(), fields_to_ignore_.end(), leaf_name) != fields_to_ignore_.end()) continue;
+    if (!fields_to_ignore_.empty() && std::find(fields_to_ignore_.begin(), fields_to_ignore_.end(), leaf_name) != fields_to_ignore_.end()) continue;
+    if (!fields_to_preserve_.empty() && std::find(fields_to_preserve_.begin(), fields_to_preserve_.end(), leaf_name) == fields_to_preserve_.end()) continue;
+    if (!is_prepend_leaves_with_branchname_) leaf_name.erase(0, branch_name_.size() + 1);
     std::replace(leaf_name.begin(), leaf_name.end(), '.', '_');
     plain_tree_->Branch(leaf_name.c_str(), &(vars_.at(i)), Form("%s/F", leaf_name.c_str()));
   }

infra/PlainTreeFiller.cpp

@@ -27,9 +27,11 @@ class PlainTreeFiller : public AnalysisTask {
     tree_name_ = std::move(tree);
   }
 
-  void SetFieldsToIgnore(const std::vector<std::string>&& fields_to_ignore);
+  void SetFieldsToIgnore(const std::vector<std::string>& fields_to_ignore);
+  void SetFieldsToPreserve(const std::vector<std::string>& fields_to_preserve);
 
   void SetIsIgnoreDefaultFields(bool is = true) { is_ignore_defual_fields_ = is; }
+  void SetIsPrependLeavesWithBranchName(bool is = true) { is_prepend_leaves_with_branchname_ = is; }
 
  protected:
   TFile* file_{nullptr};
@@ -41,8 +43,10 @@ class PlainTreeFiller : public AnalysisTask {
 
   std::vector<float> vars_{};
   std::vector<std::string> fields_to_ignore_{};
+  std::vector<std::string> fields_to_preserve_{};
 
   bool is_ignore_defual_fields_{false};
+  bool is_prepend_leaves_with_branchname_{true};
 };
 
 }// namespace AnalysisTree

infra/PlainTreeFiller.hpp

@@ -25,7 +25,7 @@ bool operator==(const SimpleCut& that, const SimpleCut& other) {
     return true;
   }
   // if both SimpleCuts were defined via lambda, they're assumed not equal (unless they're in the same memory place)
-  if(that.hash_ == 1 && other.hash_ == 1) return false;
+  if (that.hash_ == 1 && other.hash_ == 1) return false;
   return that.vars_ == other.vars_ && that.title_ == other.title_ && that.hash_ == other.hash_;
 }
 

infra/SimpleCut.cpp

@@ -41,7 +41,7 @@ class SimpleCut {
                    [](const std::string& arg_name) { return Variable::FromString(arg_name); });
 
     FillBranchNames();
-    hash_ = 1; // Impossible to calculate for lambda_
+    hash_ = 1;// Impossible to calculate for lambda_
   }
 
   SimpleCut(const std::vector<Variable>& vars, std::function<bool(std::vector<double>&)> lambda, std::string title = "") : title_(std::move(title)),
@@ -50,7 +50,7 @@ class SimpleCut {
       vars_.emplace_back(var);
     }
     FillBranchNames();
-    hash_ = 1; // Impossible to calculate for lambda_
+    hash_ = 1;// Impossible to calculate for lambda_
   }
 
   /**