From 2894acc8e98da342b283c801feef0c9c4b4b1077 Mon Sep 17 00:00:00 2001
From: Thomas <purcell@fhi-berlin.mpg.de>
Date: Sun, 15 Aug 2021 20:35:51 +0200
Subject: [PATCH] Add the task name and sample ID column to model files
Also backwards compatiable
---
src/descriptor_identifier/model/Model.cpp | 89 ++++++++++++++-----
src/descriptor_identifier/model/Model.hpp | 14 ++-
.../model/ModelClassifier.cpp | 30 +++++--
.../model/ModelClassifier.hpp | 10 ++-
.../model/ModelLogRegressor.cpp | 7 +-
.../model/ModelLogRegressor.hpp | 8 +-
.../model/ModelRegressor.cpp | 28 +++++-
.../model/ModelRegressor.hpp | 10 ++-
.../solver/SISSOClassifier.cpp | 9 +-
.../solver/SISSOClassifier.hpp | 12 +--
.../solver/SISSOLogRegressor.cpp | 9 +-
.../solver/SISSOLogRegressor.hpp | 12 +--
.../solver/SISSORegressor.cpp | 9 +-
.../solver/SISSORegressor.hpp | 12 +--
.../solver/SISSOSolver.cpp | 4 +-
.../solver/SISSOSolver.hpp | 14 +--
src/inputs/InputParser.cpp | 11 ++-
src/inputs/InputParser.hpp | 2 +-
src/main.cpp | 6 +-
.../bindings_docstring_keyed.cpp | 12 +--
.../descriptor_identifier/SISSOClassifier.cpp | 8 +-
.../SISSOLogRegressor.cpp | 8 +-
.../descriptor_identifier/SISSORegressor.cpp | 8 +-
.../descriptor_identifier/SISSOSolver.cpp | 8 +-
src/python/py_interface/get_solver.py | 10 +--
src/python/py_interface/import_dataframe.py | 18 ++--
.../model/test_model_classifier.cc | 23 +++--
.../model/test_model_log_regressor.cc | 18 +++-
.../model/test_model_regressor.cc | 18 +++-
.../test_sisso_log_regressor.cc | 10 +++
.../sisso_regressor/test_sisso_regressor.cc | 11 +++
31 files changed, 321 insertions(+), 127 deletions(-)
diff --git a/src/descriptor_identifier/model/Model.cpp b/src/descriptor_identifier/model/Model.cpp
index 7ebc8e61..a0ff6139 100644
--- a/src/descriptor_identifier/model/Model.cpp
+++ b/src/descriptor_identifier/model/Model.cpp
@@ -26,8 +26,14 @@ Model::Model(
const Unit prop_unit,
const std::shared_ptr<LossFunction> loss,
const std::vector<model_node_ptr> feats,
- const std::vector<int> leave_out_inds
+ const std::vector<int> leave_out_inds,
+ const std::vector<std::string> sample_ids_train,
+ const std::vector<std::string> sample_ids_test,
+ const std::vector<std::string> task_names
) :
+ _sample_ids_train(sample_ids_train),
+ _sample_ids_test(sample_ids_test),
+ _task_names(task_names),
_n_samp_train(feats[0]->n_samp()),
_n_samp_test(feats[0]->n_samp_test()),
_n_dim(feats.size()),
@@ -161,17 +167,7 @@ void Model::to_file(const std::string filename, const bool train) const
out_file_stream << "# Property Label: $" << str_utils::latexify(_prop_label) << "$; Unit of the Property: " << _prop_unit.toString() << std::endl;
out_file_stream << error_summary_string(train);
- out_file_stream << coefs_header();
-
- for(int cc = 0; cc < _coefs.size(); ++cc)
- {
- out_file_stream << std::setw(10) << std::left << "# " + std::to_string(cc) + ", ";
- for(auto& coeff : _coefs[cc])
- {
- out_file_stream << std::setprecision(15) << std::scientific << std::right << std::setw(22) << coeff << std::setw(2) << ", ";
- }
- out_file_stream << "\n";
- }
+ out_file_stream << write_coefs();
out_file_stream << "# Feature Rung, Units, and Expressions" << std::endl;
for(int ff = 0; ff < _feats.size(); ++ff)
@@ -185,13 +181,17 @@ void Model::to_file(const std::string filename, const bool train) const
out_file_stream << boost::algorithm::join(_feats[ff]->get_x_in_expr_list(), ",") << std::endl;
}
+ int task_header_w = std::max(
+ 6,
+ static_cast<int>(std::max_element(_task_names.begin(), _task_names.end(), [](std::string s1, std::string s2){return s1.size() <= s2.size();})->size())
+ );
out_file_stream << "# Number of Samples Per Task" << std::endl;
if(train)
{
- out_file_stream << std::setw(10) << std::left << "# Task," << std::setw(24) << "n_mats_train" << std::endl;
+ out_file_stream << std::setw(task_header_w) << std::left << "# Task" << std::setw(2) << ", " << std::setw(24) << "n_mats_train" << std::endl;
for(int tt = 0; tt < task_sizes_train_vec.size(); ++tt)
{
- out_file_stream << std::left << std::setw(10) << "# " + std::to_string(tt) + ", ";
+ out_file_stream << std::left << std::setw(task_header_w) << "# " + _task_names[tt] << std::setw(2) << ", ";
out_file_stream << std::left << std::setw(22) << task_sizes_train_vec[tt] << std::endl;
}
}
@@ -200,7 +200,7 @@ void Model::to_file(const std::string filename, const bool train) const
out_file_stream << std::setw(10) << std::left << "# Task," << std::setw(24) << "n_mats_test" << std::endl;
for(int tt = 0; tt < task_sizes_test_vec.size(); ++tt)
{
- out_file_stream << std::left << std::setw(10) << "# " + std::to_string(tt) + ", ";
+ out_file_stream << std::left << std::setw(10) << "# " + _task_names[tt] + ", ";
out_file_stream << std::left << std::setw(22) << task_sizes_test_vec[tt] << std::endl;
}
@@ -212,7 +212,23 @@ void Model::to_file(const std::string filename, const bool train) const
out_file_stream << " ]" << std::endl;
}
- out_file_stream << "\n" << std::setw(22) << std::left << "# Property Value" << std::setw(2) << ", " << std::setw(22) << " Property Value (EST)";
+ int max_sample_id_len = 12;
+ if(train)
+ {
+ max_sample_id_len = std::max(
+ max_sample_id_len,
+ static_cast<int>(std::max_element(_sample_ids_train.begin(), _sample_ids_train.end(), [](std::string s1, std::string s2){return s1.size() <= s2.size();})->size())
+ );
+ }
+ else
+ {
+ max_sample_id_len = std::max(
+ max_sample_id_len,
+ static_cast<int>(std::max_element(_sample_ids_test.begin(), _sample_ids_test.end(), [](std::string s1, std::string s2){return s1.size() <= s2.size();})->size())
+ );
+ }
+ out_file_stream << "\n" << std::setw(max_sample_id_len) << std::left << "# Sample ID" << std::setw(2) << ", ";
+ out_file_stream << std::setw(22) << std::left << "Property Value" << std::setw(2) << ", " << std::setw(22) << " Property Value (EST)";
for(int ff = 0; ff < _feats.size(); ++ff)
{
out_file_stream << std::setw(2) << ", " << std::setw(22) << " Feature " + std::to_string(ff) + " Value";
@@ -223,6 +239,7 @@ void Model::to_file(const std::string filename, const bool train) const
{
for(int ss = 0; ss < _n_samp_train; ++ss)
{
+ out_file_stream << std::left << std::setw(max_sample_id_len) << _sample_ids_train[ss] << std::setw(2) << ", ";
out_file_stream << std::right << std::setw(22) << std::setprecision(15) << std::scientific << prop_train_vec[ss] << std::setw(2) << ", ";
out_file_stream << std::setw(22) << prop_train_est_vec[ss];
for(int ff = 0; ff < _n_dim; ++ff)
@@ -236,6 +253,7 @@ void Model::to_file(const std::string filename, const bool train) const
{
for(int ss = 0; ss < _n_samp_test; ++ss)
{
+ out_file_stream << std::left << std::setw(max_sample_id_len) << _sample_ids_test[ss] << std::setw(2) << ", ";
out_file_stream << std::right << std::setw(22) << std::setprecision(15) << std::scientific << prop_test_vec[ss] << std::setw(2) << ", ";
out_file_stream << std::setw(22) << prop_test_est_vec[ss];
for(int ff = 0; ff < _n_dim; ++ff)
@@ -435,11 +453,16 @@ void Model::populate_model(const std::string train_filename, const std::string t
{
++n_task;
split_line = str_utils::split_string_trim(line);
+ _task_names.push_back(split_line[0]);
_n_samp_train += std::stoi(split_line[1]);
task_sizes_train.push_back(std::stoi(split_line[1]));
if(with_test)
{
split_line = str_utils::split_string_trim(test_line);
+ if(split_line[0].compare(_task_names.back()) != 0)
+ {
+ throw std::logic_error("The task names for the test and train files are not in the same order.");
+ }
_n_samp_test += std::stoi(split_line[1]);
task_sizes_test.push_back(std::stoi(split_line[1]));
std::getline(test_file_stream, test_line);
@@ -454,6 +477,9 @@ void Model::populate_model(const std::string train_filename, const std::string t
std::vector<double> prop_train(_n_samp_train);
std::vector<double> prop_test(_n_samp_test);
+ _sample_ids_train.resize(_n_samp_train);
+ _sample_ids_test.resize(_n_samp_test);
+
if(with_test)
{
split_line = str_utils::split_string_trim(test_line, "[]");
@@ -479,16 +505,28 @@ void Model::populate_model(const std::string train_filename, const std::string t
std::vector<std::vector<double>> feat_vals(n_dim, std::vector<double>(_n_samp_train, 0.0));
std::vector<std::vector<double>> feat_test_vals(n_dim, std::vector<double>(_n_samp_test, 0.0));
+
+ bool with_samp_id = false;
for(int ns = 0; ns < _n_samp_train; ++ns)
{
std::getline(train_file_stream, line);
split_line = str_utils::split_string_trim(line);
+ if((split_line.size() > _n_dim + 2))
+ {
+ with_samp_id = true;
+ _sample_ids_train[ns] = split_line[0];
+ }
+ else
+ {
+ with_samp_id = false;
+ _sample_ids_train[ns] = std::to_string(ns);
+ }
- prop_train[ns] = std::stod(split_line[0]);
+ prop_train[ns] = std::stod(split_line[with_samp_id]);
for(int nf = 0; nf < n_dim; ++nf)
{
- feat_vals[nf][ns] = std::stod(split_line[2 + nf]);
+ feat_vals[nf][ns] = std::stod(split_line[2 + nf + with_samp_id]);
}
}
for(int ns = 0; ns < _n_samp_test; ++ns)
@@ -496,11 +534,22 @@ void Model::populate_model(const std::string train_filename, const std::string t
std::getline(test_file_stream, test_line);
split_line = str_utils::split_string_trim(test_line);
- prop_test[ns] = std::stod(split_line[0]);
+ if((split_line.size() > _n_dim + 2))
+ {
+ with_samp_id = true;
+ _sample_ids_test[ns] = split_line[0];
+ }
+ else
+ {
+ with_samp_id = false;
+ _sample_ids_test[ns] = std::to_string(ns);
+ }
+
+ prop_test[ns] = std::stod(split_line[with_samp_id]);
for(int nf = 0; nf < n_dim; ++nf)
{
- feat_test_vals[nf][ns] = std::stod(split_line[2 + nf]);
+ feat_test_vals[nf][ns] = std::stod(split_line[2 + nf + with_samp_id]);
}
}
train_file_stream.close();
diff --git a/src/descriptor_identifier/model/Model.hpp b/src/descriptor_identifier/model/Model.hpp
index 61a1e0c1..3940ba76 100644
--- a/src/descriptor_identifier/model/Model.hpp
+++ b/src/descriptor_identifier/model/Model.hpp
@@ -42,6 +42,10 @@
class Model
{
protected:
+ std::vector<std::string> _sample_ids_train; //!< Vector storing all sample ids for the training samples
+ std::vector<std::string> _sample_ids_test; //!< Vector storing all sample ids for the test samples
+ std::vector<std::string> _task_names; //!< Vector storing the ID of the task names
+
int _n_samp_train; //!< The number of samples per feature
int _n_samp_test; //!< The number of test samples per feature
int _n_dim; //!< The number of dimensions of the model
@@ -72,13 +76,19 @@ public:
* @param loss The LossFunction used to calculate the model
* @param feats The features of the model
* @param leave_out_inds The indexes of the samples for the test set
+ * @param sample_ids_train A vector storing all sample ids for the training samples
+ * @param sample_ids_test A vector storing all sample ids for the test samples
+ * @param task_names A vector storing the ID of the task names
*/
Model(
const std::string prop_label,
const Unit prop_unit,
const std::shared_ptr<LossFunction> loss,
const std::vector<model_node_ptr> feats,
- const std::vector<int> leave_out_inds
+ const std::vector<int> leave_out_inds,
+ const std::vector<std::string> sample_ids_train,
+ const std::vector<std::string> sample_ids_test,
+ const std::vector<std::string> task_names
);
/**
@@ -289,7 +299,7 @@ public:
/**
* @brief Get the coefficients list header for output file
*/
- virtual std::string coefs_header() const = 0;
+ virtual std::string write_coefs() const = 0;
// DocString: model_fix_intercept
/**
diff --git a/src/descriptor_identifier/model/ModelClassifier.cpp b/src/descriptor_identifier/model/ModelClassifier.cpp
index 480c04fc..74e75873 100644
--- a/src/descriptor_identifier/model/ModelClassifier.cpp
+++ b/src/descriptor_identifier/model/ModelClassifier.cpp
@@ -26,9 +26,12 @@ ModelClassifier::ModelClassifier(
const Unit prop_unit,
const std::shared_ptr<LossFunction> loss,
const std::vector<model_node_ptr> feats,
- const std::vector<int> leave_out_inds
+ const std::vector<int> leave_out_inds,
+ const std::vector<std::string> sample_ids_train,
+ const std::vector<std::string> sample_ids_test,
+ const std::vector<std::string> task_names
) :
- Model(prop_label, prop_unit, loss, feats, leave_out_inds),
+ Model(prop_label, prop_unit, loss, feats, leave_out_inds, sample_ids_train, sample_ids_test, task_names),
_train_n_convex_overlap(0),
_test_n_convex_overlap(0),
_n_class(loss->n_class())
@@ -209,11 +212,16 @@ std::string ModelClassifier::error_summary_string(bool train) const
return error_stream.str();
}
-std::string ModelClassifier::coefs_header() const
+std::string ModelClassifier::write_coefs() const
{
std::stringstream coef_head_stream;
- coef_head_stream << "# Plane Divider" << std::endl;
- coef_head_stream << std::setw(10) << std::left << "# Task";
+ coef_head_stream << "# Decision Boundaries" << std::endl;
+ int n_db = _n_class * (_n_class - 1) / 2;
+ int task_header_w = 1 + static_cast<int>(std::floor(std::log10(n_db))) + std::max(
+ 6,
+ static_cast<int>(std::max_element(_task_names.begin(), _task_names.end(), [](std::string s1, std::string s2){return s1.size() <= s2.size();})->size())
+ );
+ coef_head_stream << std::setw(task_header_w + 2) << std::left << "# Task";
for(int cc = 0; cc < _coefs[0].size() - 1; ++cc)
{
@@ -221,5 +229,17 @@ std::string ModelClassifier::coefs_header() const
}
coef_head_stream << " b" << std::endl;
+ for(int tt = 0; tt < _task_names.size(); ++tt)
+ {
+ for(int db = 0; db < n_db; ++db)
+ {
+ coef_head_stream << std::setw(task_header_w) << std::left << "# " + _task_names[tt] + "_" + std::to_string(db) << std::setw(2) << ", ";
+ for(auto& coeff : _coefs[tt * n_db + db])
+ {
+ coef_head_stream << std::setprecision(15) << std::scientific << std::right << std::setw(22) << coeff << std::setw(2) << ", ";
+ }
+ coef_head_stream << "\n";
+ }
+ }
return coef_head_stream.str();
}
diff --git a/src/descriptor_identifier/model/ModelClassifier.hpp b/src/descriptor_identifier/model/ModelClassifier.hpp
index fe29e171..28ef2d60 100644
--- a/src/descriptor_identifier/model/ModelClassifier.hpp
+++ b/src/descriptor_identifier/model/ModelClassifier.hpp
@@ -53,13 +53,19 @@ public:
* @param loss The LossFunction used to calculate the model
* @param feats The features of the model
* @param leave_out_inds The indexes of the samples for the test set
+ * @param sample_ids_train A vector storing all sample ids for the training samples
+ * @param sample_ids_test A vector storing all sample ids for the test samples
+ * @param task_names A vector storing the ID of the task names
*/
ModelClassifier(
const std::string prop_label,
const Unit prop_unit,
const std::shared_ptr<LossFunction> loss,
const std::vector<model_node_ptr> feats,
- const std::vector<int> leave_out_inds
+ const std::vector<int> leave_out_inds,
+ const std::vector<std::string> sample_ids_train,
+ const std::vector<std::string> sample_ids_test,
+ const std::vector<std::string> task_names
);
// DocString: model_class_init_train
@@ -195,7 +201,7 @@ public:
/**
* @brief Get the coefficients list header for output file
*/
- std::string coefs_header() const;
+ std::string write_coefs() const;
/**
* @brief Copy the training error into a different vector
diff --git a/src/descriptor_identifier/model/ModelLogRegressor.cpp b/src/descriptor_identifier/model/ModelLogRegressor.cpp
index 6bec56f6..a7519adc 100644
--- a/src/descriptor_identifier/model/ModelLogRegressor.cpp
+++ b/src/descriptor_identifier/model/ModelLogRegressor.cpp
@@ -29,9 +29,12 @@ ModelLogRegressor::ModelLogRegressor(
const Unit prop_unit,
const std::shared_ptr<LossFunction> loss,
const std::vector<model_node_ptr> feats,
- const std::vector<int> leave_out_inds
+ const std::vector<int> leave_out_inds,
+ const std::vector<std::string> sample_ids_train,
+ const std::vector<std::string> sample_ids_test,
+ const std::vector<std::string> task_names
) :
- ModelRegressor(prop_label, prop_unit, loss, feats, leave_out_inds)
+ ModelRegressor(prop_label, prop_unit, loss, feats, leave_out_inds, sample_ids_train, sample_ids_test, task_names)
{}
ModelLogRegressor::ModelLogRegressor(const std::string train_file)
diff --git a/src/descriptor_identifier/model/ModelLogRegressor.hpp b/src/descriptor_identifier/model/ModelLogRegressor.hpp
index 1f6bce56..da7f5e09 100644
--- a/src/descriptor_identifier/model/ModelLogRegressor.hpp
+++ b/src/descriptor_identifier/model/ModelLogRegressor.hpp
@@ -51,13 +51,19 @@ public:
* @param loss The LossFunction used to calculate the model
* @param feats The features of the model
* @param leave_out_inds The indexes of the samples for the test set
+ * @param sample_ids_train A vector storing all sample ids for the training samples
+ * @param sample_ids_test A vector storing all sample ids for the test samples
+ * @param task_names A vector storing the ID of the task names
*/
ModelLogRegressor(
const std::string prop_label,
const Unit prop_unit,
const std::shared_ptr<LossFunction> loss,
const std::vector<model_node_ptr> feats,
- const std::vector<int> leave_out_inds
+ const std::vector<int> leave_out_inds,
+ const std::vector<std::string> sample_ids_train,
+ const std::vector<std::string> sample_ids_test,
+ const std::vector<std::string> task_names
);
// DocString: model_log_reg_init_train
diff --git a/src/descriptor_identifier/model/ModelRegressor.cpp b/src/descriptor_identifier/model/ModelRegressor.cpp
index fd7de70a..4d809f64 100644
--- a/src/descriptor_identifier/model/ModelRegressor.cpp
+++ b/src/descriptor_identifier/model/ModelRegressor.cpp
@@ -29,9 +29,12 @@ ModelRegressor::ModelRegressor(
const Unit prop_unit,
const std::shared_ptr<LossFunction> loss,
const std::vector<model_node_ptr> feats,
- const std::vector<int> leave_out_inds
+ const std::vector<int> leave_out_inds,
+ const std::vector<std::string> sample_ids_train,
+ const std::vector<std::string> sample_ids_test,
+ const std::vector<std::string> task_names
) :
- Model(prop_label, prop_unit, loss, feats, leave_out_inds)
+ Model(prop_label, prop_unit, loss, feats, leave_out_inds, sample_ids_train, sample_ids_test, task_names)
{
double rmse = (*_loss)(_feats);
}
@@ -168,15 +171,22 @@ std::string ModelRegressor::error_summary_string(bool train) const
return error_stream.str();
}
-std::string ModelRegressor::coefs_header() const
+std::string ModelRegressor::write_coefs() const
{
std::stringstream coef_head_stream;
+ int task_header_w = std::max(
+ 6,
+ static_cast<int>(std::max_element(_task_names.begin(), _task_names.end(), [](std::string s1, std::string s2){return s1.size() <= s2.size();})->size())
+ );
+
coef_head_stream << "# Coefficients" << std::endl;
- coef_head_stream << std::setw(10) << std::left << "# Task";
+ coef_head_stream << std::setw(task_header_w + 2) << std::left << "# Task";
for(int cc = 0; cc < _coefs[0].size() - (!_fix_intercept); ++cc)
+ {
coef_head_stream << std::setw(24) << " a" + std::to_string(cc);
+ }
if(!_fix_intercept)
{
@@ -187,6 +197,16 @@ std::string ModelRegressor::coefs_header() const
coef_head_stream << std::endl;
}
+ for(int cc = 0; cc < _coefs.size(); ++cc)
+ {
+ coef_head_stream << std::setw(task_header_w) << std::left << "# " + _task_names[cc] << std::setw(2) << ", ";
+ for(auto& coeff : _coefs[cc])
+ {
+ coef_head_stream << std::setprecision(15) << std::scientific << std::right << std::setw(22) << coeff << std::setw(2) << ", ";
+ }
+ coef_head_stream << "\n";
+ }
+
return coef_head_stream.str();
}
diff --git a/src/descriptor_identifier/model/ModelRegressor.hpp b/src/descriptor_identifier/model/ModelRegressor.hpp
index af386356..80484280 100644
--- a/src/descriptor_identifier/model/ModelRegressor.hpp
+++ b/src/descriptor_identifier/model/ModelRegressor.hpp
@@ -50,13 +50,19 @@ public:
* @param loss The LossFunction used to calculate the model
* @param feats The features of the model
* @param leave_out_inds The indexes of the samples for the test set
+ * @param sample_ids_train A vector storing all sample ids for the training samples
+ * @param sample_ids_test A vector storing all sample ids for the test samples
+ * @param task_names A vector storing the ID of the task names
*/
ModelRegressor(
const std::string prop_label,
const Unit prop_unit,
const std::shared_ptr<LossFunction> loss,
const std::vector<model_node_ptr> feats,
- const std::vector<int> leave_out_inds
+ const std::vector<int> leave_out_inds,
+ const std::vector<std::string> sample_ids_train,
+ const std::vector<std::string> sample_ids_test,
+ const std::vector<std::string> task_names
);
// DocString: model_reg_init_train
@@ -195,7 +201,7 @@ public:
/**
* @brief Get the coefficients list header for output file
*/
- std::string coefs_header() const;
+ std::string write_coefs() const;
/**
* @brief Copy the training error into a different vector
diff --git a/src/descriptor_identifier/solver/SISSOClassifier.cpp b/src/descriptor_identifier/solver/SISSOClassifier.cpp
index 91f7dc92..28726599 100644
--- a/src/descriptor_identifier/solver/SISSOClassifier.cpp
+++ b/src/descriptor_identifier/solver/SISSOClassifier.cpp
@@ -35,7 +35,7 @@ SISSOClassifier::SISSOClassifier(
const int n_models_store,
const std::vector<std::string> sample_ids_train,
const std::vector<std::string> sample_ids_test,
- const std::vector<std::string> task_keys
+ const std::vector<std::string> task_names
):
SISSOSolver(
"classification",
@@ -52,7 +52,7 @@ SISSOClassifier::SISSOClassifier(
n_models_store,
sample_ids_train,
sample_ids_test,
- task_keys,
+ task_names,
false
),
_c(1000.0),
@@ -310,7 +310,10 @@ void SISSOClassifier::l0_regularization(const int n_dim)
_prop_unit,
loss_function_util::copy(_loss),
min_nodes[rr],
- _leave_out_inds
+ _leave_out_inds,
+ _sample_ids_train,
+ _sample_ids_test,
+ _task_names
)
);
}
diff --git a/src/descriptor_identifier/solver/SISSOClassifier.hpp b/src/descriptor_identifier/solver/SISSOClassifier.hpp
index bd854553..16ab12d0 100644
--- a/src/descriptor_identifier/solver/SISSOClassifier.hpp
+++ b/src/descriptor_identifier/solver/SISSOClassifier.hpp
@@ -67,7 +67,7 @@ public:
* @param n_models_store The number of models to output to files
* @param sample_ids_train A list storing all sample ids for the training samples
* @param sample_ids_test A list storing all sample ids for the test samples
- * @param task_keys A list storing the ID of the task names
+ * @param task_names A list storing the ID of the task names
*/
SISSOClassifier(
const std::shared_ptr<FeatureSpace> feat_space,
@@ -82,7 +82,7 @@ public:
const int n_models_store,
const std::vector<std::string> sample_ids_train,
const std::vector<std::string> sample_ids_test,
- const std::vector<std::string> task_keys
+ const std::vector<std::string> task_names
);
/**
@@ -153,7 +153,7 @@ public:
* @param n_models_store (int) The number of models to output to files
* @param sample_ids_train A list storing all sample ids for the training samples
* @param sample_ids_test A list storing all sample ids for the test samples
- * @param task_keys A list storing the ID of the task names
+ * @param task_names A list storing the ID of the task names
*/
SISSOClassifier(
std::shared_ptr<FeatureSpace> feat_space,
@@ -169,7 +169,7 @@ public:
int n_models_store,
py::list sample_ids_train,
py::list sample_ids_test,
- py::list task_keys
+ py::list task_names
);
// DocString: sisso_class_init_list
@@ -189,7 +189,7 @@ public:
* @param n_models_store (int) The number of models to output to files
* @param sample_ids_train A list storing all sample ids for the training samples
* @param sample_ids_test A list storing all sample ids for the test samples
- * @param task_keys A list storing the ID of the task names
+ * @param task_names A list storing the ID of the task names
*/
SISSOClassifier(
std::shared_ptr<FeatureSpace> feat_space,
@@ -205,7 +205,7 @@ public:
int n_models_store,
py::list sample_ids_train,
py::list sample_ids_test,
- py::list task_keys
+ py::list task_names
);
// DocString: sisso_class_models_py
diff --git a/src/descriptor_identifier/solver/SISSOLogRegressor.cpp b/src/descriptor_identifier/solver/SISSOLogRegressor.cpp
index ab36f0ed..9d502966 100644
--- a/src/descriptor_identifier/solver/SISSOLogRegressor.cpp
+++ b/src/descriptor_identifier/solver/SISSOLogRegressor.cpp
@@ -35,7 +35,7 @@ SISSOLogRegressor::SISSOLogRegressor(
const int n_models_store,
const std::vector<std::string> sample_ids_train,
const std::vector<std::string> sample_ids_test,
- const std::vector<std::string> task_keys,
+ const std::vector<std::string> task_names,
const bool fix_intercept
):
SISSORegressor(
@@ -52,7 +52,7 @@ SISSOLogRegressor::SISSOLogRegressor(
n_models_store,
sample_ids_train,
sample_ids_test,
- task_keys,
+ task_names,
fix_intercept
)
{
@@ -106,7 +106,10 @@ void SISSOLogRegressor::add_models(const std::vector<std::vector<int>> indexes)
_prop_unit,
loss_function_util::copy(_loss),
min_nodes.back(),
- _leave_out_inds
+ _leave_out_inds,
+ _sample_ids_train,
+ _sample_ids_test,
+ _task_names
);
_models.back().push_back(model);
}
diff --git a/src/descriptor_identifier/solver/SISSOLogRegressor.hpp b/src/descriptor_identifier/solver/SISSOLogRegressor.hpp
index 2f8ce7d5..34c63511 100644
--- a/src/descriptor_identifier/solver/SISSOLogRegressor.hpp
+++ b/src/descriptor_identifier/solver/SISSOLogRegressor.hpp
@@ -57,7 +57,7 @@ public:
* @param n_models_store The number of models to output to files
* @param sample_ids_train A list storing all sample ids for the training samples
* @param sample_ids_test A list storing all sample ids for the test samples
- * @param task_keys A list storing the ID of the task names
+ * @param task_names A list storing the ID of the task names
* @param fix_intrecept If true the bias term is fixed at 0
*/
SISSOLogRegressor(
@@ -74,7 +74,7 @@ public:
const int n_models_store,
const std::vector<std::string> sample_ids_train,
const std::vector<std::string> sample_ids_test,
- const std::vector<std::string> task_keys,
+ const std::vector<std::string> task_names,
const bool fix_intercept=false
);
@@ -121,7 +121,7 @@ public:
* @param n_models_store (int) The number of models to output to files
* @param sample_ids_train A list storing all sample ids for the training samples
* @param sample_ids_test A list storing all sample ids for the test samples
- * @param task_keys A list storing the ID of the task names
+ * @param task_names A list storing the ID of the task names
* @param fix_intrecept (bool) If true the bias term is fixed at 0
*/
SISSOLogRegressor(
@@ -138,7 +138,7 @@ public:
int n_models_store,
py::list sample_ids_train,
py::list sample_ids_test,
- py::list task_keys,
+ py::list task_names,
bool fix_intercept=false
);
@@ -159,7 +159,7 @@ public:
* @param n_models_store (int) The number of models to output to files
* @param sample_ids_train A list storing all sample ids for the training samples
* @param sample_ids_test A list storing all sample ids for the test samples
- * @param task_keys A list storing the ID of the task names
+ * @param task_names A list storing the ID of the task names
* @param fix_intrecept (bool) If true the bias term is fixed at 0
*/
SISSOLogRegressor(
@@ -176,7 +176,7 @@ public:
int n_models_store,
py::list sample_ids_train,
py::list sample_ids_test,
- py::list task_keys,
+ py::list task_names,
bool fix_intercept=false
);
diff --git a/src/descriptor_identifier/solver/SISSORegressor.cpp b/src/descriptor_identifier/solver/SISSORegressor.cpp
index 7705db94..e6853bdc 100644
--- a/src/descriptor_identifier/solver/SISSORegressor.cpp
+++ b/src/descriptor_identifier/solver/SISSORegressor.cpp
@@ -35,7 +35,7 @@ SISSORegressor::SISSORegressor(
const int n_models_store,
const std::vector<std::string> sample_ids_train,
const std::vector<std::string> sample_ids_test,
- const std::vector<std::string> task_keys,
+ const std::vector<std::string> task_names,
const bool fix_intercept
):
SISSOSolver(
@@ -53,7 +53,7 @@ SISSORegressor::SISSORegressor(
n_models_store,
sample_ids_train,
sample_ids_test,
- task_keys,
+ task_names,
fix_intercept
)
{}
@@ -76,7 +76,10 @@ void SISSORegressor::add_models(const std::vector<std::vector<int>> indexes)
_prop_unit,
loss_function_util::copy(_loss),
min_nodes.back(),
- _leave_out_inds
+ _leave_out_inds,
+ _sample_ids_train,
+ _sample_ids_test,
+ _task_names
);
_models.back().push_back(model);
}
diff --git a/src/descriptor_identifier/solver/SISSORegressor.hpp b/src/descriptor_identifier/solver/SISSORegressor.hpp
index 897298ef..c7f8ff71 100644
--- a/src/descriptor_identifier/solver/SISSORegressor.hpp
+++ b/src/descriptor_identifier/solver/SISSORegressor.hpp
@@ -57,7 +57,7 @@ public:
* @param n_models_store The number of models to output to files
* @param sample_ids_train A list storing all sample ids for the training samples
* @param sample_ids_test A list storing all sample ids for the test samples
- * @param task_keys A list storing the ID of the task names
+ * @param task_names A list storing the ID of the task names
* @param fix_intrecept If true the bias term is fixed at 0
*/
SISSORegressor(
@@ -74,7 +74,7 @@ public:
const int n_models_store,
const std::vector<std::string> sample_ids_train,
const std::vector<std::string> sample_ids_test,
- const std::vector<std::string> task_keys,
+ const std::vector<std::string> task_names,
const bool fix_intercept=false
);
@@ -128,7 +128,7 @@ public:
* @param n_models_store (int) The number of models to output to files
* @param sample_ids_train A list storing all sample ids for the training samples
* @param sample_ids_test A list storing all sample ids for the test samples
- * @param task_keys A list storing the ID of the task names
+ * @param task_names A list storing the ID of the task names
* @param fix_intrecept (bool) If true the bias term is fixed at 0
*/
SISSORegressor(
@@ -145,7 +145,7 @@ public:
int n_models_store,
py::list sample_ids_train,
py::list sample_ids_test,
- py::list task_keys,
+ py::list task_names,
bool fix_intercept=false
);
@@ -166,7 +166,7 @@ public:
* @param n_models_store (int) The number of models to output to files
* @param sample_ids_train A list storing all sample ids for the training samples
* @param sample_ids_test A list storing all sample ids for the test samples
- * @param task_keys A list storing the ID of the task names
+ * @param task_names A list storing the ID of the task names
* @param fix_intrecept (bool) If true the bias term is fixed at 0
*/
SISSORegressor(
@@ -183,7 +183,7 @@ public:
int n_models_store,
py::list sample_ids_train,
py::list sample_ids_test,
- py::list task_keys,
+ py::list task_names,
bool fix_intercept=false
);
diff --git a/src/descriptor_identifier/solver/SISSOSolver.cpp b/src/descriptor_identifier/solver/SISSOSolver.cpp
index 9039b0aa..33019c72 100644
--- a/src/descriptor_identifier/solver/SISSOSolver.cpp
+++ b/src/descriptor_identifier/solver/SISSOSolver.cpp
@@ -36,12 +36,12 @@ SISSOSolver::SISSOSolver(
const int n_models_store,
const std::vector<std::string> sample_ids_train,
const std::vector<std::string> sample_ids_test,
- const std::vector<std::string> task_keys,
+ const std::vector<std::string> task_names,
const bool fix_intercept
):
_sample_ids_train(sample_ids_train),
_sample_ids_test(sample_ids_test),
- _task_keys(task_keys),
+ _task_names(task_names),
_task_sizes_train(task_sizes_train),
_task_sizes_test(task_sizes_test),
_leave_out_inds(leave_out_inds),
diff --git a/src/descriptor_identifier/solver/SISSOSolver.hpp b/src/descriptor_identifier/solver/SISSOSolver.hpp
index a7b279d6..939b8285 100644
--- a/src/descriptor_identifier/solver/SISSOSolver.hpp
+++ b/src/descriptor_identifier/solver/SISSOSolver.hpp
@@ -40,7 +40,7 @@ class SISSOSolver
protected:
const std::vector<std::string> _sample_ids_train; //!< Vector storing all sample ids for the training samples
const std::vector<std::string> _sample_ids_test; //!< Vector storing all sample ids for the test samples
- const std::vector<std::string> _task_keys; //!< Vector storing the ID of the task names
+ const std::vector<std::string> _task_names; //!< Vector storing the ID of the task names
const std::vector<int> _task_sizes_train; //!< Number of training samples per task
const std::vector<int> _task_sizes_test; //!< Number of testing samples per task
@@ -77,7 +77,7 @@ public:
* @param n_models_store The number of models to output to files
* @param sample_ids_train A vector storing all sample ids for the training samples
* @param sample_ids_test A vector storing all sample ids for the test samples
- * @param task_keys A vector storing the ID of the task names
+ * @param task_names A vector storing the ID of the task names
* @param fix_intrecept If true the bias term is fixed at 0
*/
SISSOSolver(
@@ -95,7 +95,7 @@ public:
const int n_models_store,
const std::vector<std::string> sample_ids_train,
const std::vector<std::string> sample_ids_test,
- const std::vector<std::string> task_keys,
+ const std::vector<std::string> task_names,
const bool fix_intercept=false
);
@@ -175,7 +175,7 @@ public:
* @param n_models_store The number of models to output to files
* @param sample_ids_train A list storing all sample ids for the training samples
* @param sample_ids_test A list storing all sample ids for the test samples
- * @param task_keys A list storing the ID of the task names
+ * @param task_names A list storing the ID of the task names
* @param fix_intrecept If true the bias term is fixed at 0
*/
SISSOSolver(
@@ -193,7 +193,7 @@ public:
int n_models_store,
py::list sample_ids_train,
py::list sample_ids_test,
- py::list task_keys,
+ py::list task_names,
bool fix_intercept=false
);
@@ -214,7 +214,7 @@ public:
* @param n_models_store The number of models to output to files
* @param sample_ids_train A list storing all sample ids for the training samples
* @param sample_ids_test A list storing all sample ids for the test samples
- * @param task_keys A list storing the ID of the task names
+ * @param task_names A list storing the ID of the task names
* @param fix_intrecept If true the bias term is fixed at 0
*/
SISSOSolver(
@@ -232,7 +232,7 @@ public:
int n_models_store,
py::list sample_ids_train,
py::list sample_ids_test,
- py::list task_keys,
+ py::list task_names,
bool fix_intercept=false
);
diff --git a/src/inputs/InputParser.cpp b/src/inputs/InputParser.cpp
index 36565188..13f0183a 100644
--- a/src/inputs/InputParser.cpp
+++ b/src/inputs/InputParser.cpp
@@ -119,8 +119,8 @@ InputParser::InputParser(pt::ptree ip, std::string fn, std::shared_ptr<MPI_Inter
{
++_n_samp;
}
- tasks["none"] = std::vector<int>(_n_samp);
- std::iota(tasks["none"].begin(), tasks["none"].end(), 0);
+ tasks["all"] = std::vector<int>(_n_samp);
+ std::iota(tasks["all"].begin(), tasks["all"].end(), 0);
}
else
{
@@ -151,7 +151,6 @@ InputParser::InputParser(pt::ptree ip, std::string fn, std::shared_ptr<MPI_Inter
int start = 0;
for(auto& el : tasks)
{
- _task_keys.push_back(el.first);
_task_sizes_test.push_back(static_cast<int>(std::round(leave_out_frac * el.second.size())));
_task_sizes_train.push_back(el.second.size() - _task_sizes_test.back());
@@ -269,6 +268,11 @@ void InputParser::generate_feature_space(
int n_train_samp = 0;
int n_samp_test = 0;
+ for(auto& task : tasks)
+ {
+ _task_names.push_back(task.first);
+ }
+
while (std::getline(data_stream, line))
{
std::vector<std::string> split_line;
@@ -289,6 +293,7 @@ void InputParser::generate_feature_space(
n_samp_test = 0;
for(auto& task : tasks)
{
+
int task_ind = std::find(task.second.begin(), task.second.end(), cur_line) - task.second.begin();
for(int ii = 0; ii < task_ind; ++ii)
{
diff --git a/src/inputs/InputParser.hpp b/src/inputs/InputParser.hpp
index 6c72773a..21df1c8f 100644
--- a/src/inputs/InputParser.hpp
+++ b/src/inputs/InputParser.hpp
@@ -50,7 +50,7 @@ class InputParser
public:
std::vector<std::string> _sample_ids_train; //!< Vector storing all sample ids for the training samples
std::vector<std::string> _sample_ids_test; //!< Vector storing all sample ids for the test samples
- std::vector<std::string> _task_keys; //!< Vector storing the ID of the task names
+ std::vector<std::string> _task_names; //!< Vector storing the ID of the task names
std::vector<std::string> _param_opset; //!< Vector containing all allowed operators strings for operators with free parameters
std::vector<std::string> _opset; //!< Vector containing all allowed operators strings
diff --git a/src/main.cpp b/src/main.cpp
index bd1959f2..2f85c884 100644
--- a/src/main.cpp
+++ b/src/main.cpp
@@ -86,7 +86,7 @@ int main(int argc, char const *argv[])
ip._n_models_store,
ip._sample_ids_train,
ip._sample_ids_test,
- ip._task_keys,
+ ip._task_names,
ip._fix_intercept
);
sisso.fit();
@@ -124,7 +124,7 @@ int main(int argc, char const *argv[])
ip._n_models_store,
ip._sample_ids_train,
ip._sample_ids_test,
- ip._task_keys,
+ ip._task_names,
ip._fix_intercept
);
sisso.fit();
@@ -162,7 +162,7 @@ int main(int argc, char const *argv[])
ip._n_models_store,
ip._sample_ids_train,
ip._sample_ids_test,
- ip._task_keys
+ ip._task_names
);
sisso.fit();
diff --git a/src/python/py_binding_cpp_def/bindings_docstring_keyed.cpp b/src/python/py_binding_cpp_def/bindings_docstring_keyed.cpp
index 62907e4a..cbb996bc 100644
--- a/src/python/py_binding_cpp_def/bindings_docstring_keyed.cpp
+++ b/src/python/py_binding_cpp_def/bindings_docstring_keyed.cpp
@@ -1414,7 +1414,7 @@ void sisso::descriptor_identifier::registerSISSORegressor()
py::list,
optional<bool>
>(
- (arg("self"), arg("feat_space"), arg("prop_label"), arg("prop_unit"), arg("prop"), arg("prop_test"), arg("task_sizes_train"), arg("task_sizes_test"), arg("leave_out_inds"), arg("n_dim"), arg("n_residual"), arg("n_models_store"), arg("sample_ids_train"), arg("sample_ids_test"), arg("task_keys"), arg("fix_intercept")),
+ (arg("self"), arg("feat_space"), arg("prop_label"), arg("prop_unit"), arg("prop"), arg("prop_test"), arg("task_sizes_train"), arg("task_sizes_test"), arg("leave_out_inds"), arg("n_dim"), arg("n_residual"), arg("n_models_store"), arg("sample_ids_train"), arg("sample_ids_test"), arg("task_names"), arg("fix_intercept")),
"@DocString_sisso_reg_init_arr@"
)
)
@@ -1436,7 +1436,7 @@ void sisso::descriptor_identifier::registerSISSORegressor()
py::list,
optional<bool>
>(
- (arg("self"), arg("feat_space"), arg("prop_label"), arg("prop_unit"), arg("prop"), arg("prop_test"), arg("task_sizes_train"), arg("task_sizes_test"), arg("leave_out_inds"), arg("n_dim"), arg("n_residual"), arg("n_models_store"), arg("sample_ids_train"), arg("sample_ids_test"), arg("task_keys"), arg("fix_intercept")),
+ (arg("self"), arg("feat_space"), arg("prop_label"), arg("prop_unit"), arg("prop"), arg("prop_test"), arg("task_sizes_train"), arg("task_sizes_test"), arg("leave_out_inds"), arg("n_dim"), arg("n_residual"), arg("n_models_store"), arg("sample_ids_train"), arg("sample_ids_test"), arg("task_names"), arg("fix_intercept")),
"@DocString_sisso_reg_init_list@"
)
)
@@ -1467,7 +1467,7 @@ void sisso::descriptor_identifier::registerSISSOLogRegressor()
py::list,
optional<bool>
>(
- (arg("self"), arg("feat_space"), arg("prop_label"), arg("prop_unit"), arg("prop"), arg("prop_test"), arg("task_sizes_train"), arg("task_sizes_test"), arg("leave_out_inds"), arg("n_dim"), arg("n_residual"), arg("n_models_store"), arg("sample_ids_train"), arg("sample_ids_test"), arg("task_keys"), arg("fix_intercept")),
+ (arg("self"), arg("feat_space"), arg("prop_label"), arg("prop_unit"), arg("prop"), arg("prop_test"), arg("task_sizes_train"), arg("task_sizes_test"), arg("leave_out_inds"), arg("n_dim"), arg("n_residual"), arg("n_models_store"), arg("sample_ids_train"), arg("sample_ids_test"), arg("task_names"), arg("fix_intercept")),
"@DocString_sisso_log_reg_init_arr@"
)
)
@@ -1489,7 +1489,7 @@ void sisso::descriptor_identifier::registerSISSOLogRegressor()
py::list,
optional<bool>
>(
- (arg("self"), arg("feat_space"), arg("prop_label"), arg("prop_unit"), arg("prop"), arg("prop_test"), arg("task_sizes_train"), arg("task_sizes_test"), arg("leave_out_inds"), arg("n_dim"), arg("n_residual"), arg("n_models_store"), arg("sample_ids_train"), arg("sample_ids_test"), arg("task_keys"), arg("fix_intercept")),
+ (arg("self"), arg("feat_space"), arg("prop_label"), arg("prop_unit"), arg("prop"), arg("prop_test"), arg("task_sizes_train"), arg("task_sizes_test"), arg("leave_out_inds"), arg("n_dim"), arg("n_residual"), arg("n_models_store"), arg("sample_ids_train"), arg("sample_ids_test"), arg("task_names"), arg("fix_intercept")),
"@DocString_sisso_log_reg_init_list@"
)
)
@@ -1503,13 +1503,13 @@ void sisso::descriptor_identifier::registerSISSOClassifier()
"SISSOClassifier",
"@DocString_cls_sisso_class@",
init<std::shared_ptr<FeatureSpace>, std::string, Unit, np::ndarray, np::ndarray, py::list, py::list, py::list, int, int, int, py::list, py::list, py::list>(
- (arg("self"), arg("feat_space"), arg("prop_label"), arg("prop_unit"), arg("prop"), arg("prop_test"), arg("task_sizes_train"), arg("task_sizes_test"), arg("leave_out_inds"), arg("n_dim"), arg("n_residual"), arg("n_models_store"), arg("sample_ids_train"), arg("sample_ids_test"), arg("task_keys")),
+ (arg("self"), arg("feat_space"), arg("prop_label"), arg("prop_unit"), arg("prop"), arg("prop_test"), arg("task_sizes_train"), arg("task_sizes_test"), arg("leave_out_inds"), arg("n_dim"), arg("n_residual"), arg("n_models_store"), arg("sample_ids_train"), arg("sample_ids_test"), arg("task_names")),
"@DocString_sisso_class_init_arr@"
)
)
.def(
init<std::shared_ptr<FeatureSpace>, std::string, Unit, py::list, py::list, py::list, py::list, py::list, int, int, int, py::list, py::list, py::list>(
- (arg("self"), arg("feat_space"), arg("prop_label"), arg("prop_unit"), arg("prop"), arg("prop_test"), arg("task_sizes_train"), arg("task_sizes_test"), arg("leave_out_inds"), arg("n_dim"), arg("n_residual"), arg("n_models_store"), arg("sample_ids_train"), arg("sample_ids_test"), arg("task_keys")),
+ (arg("self"), arg("feat_space"), arg("prop_label"), arg("prop_unit"), arg("prop"), arg("prop_test"), arg("task_sizes_train"), arg("task_sizes_test"), arg("leave_out_inds"), arg("n_dim"), arg("n_residual"), arg("n_models_store"), arg("sample_ids_train"), arg("sample_ids_test"), arg("task_names")),
"@DocString_sisso_class_init_list@"
)
)
diff --git a/src/python/py_binding_cpp_def/descriptor_identifier/SISSOClassifier.cpp b/src/python/py_binding_cpp_def/descriptor_identifier/SISSOClassifier.cpp
index 220a520a..4ae3b8ba 100644
--- a/src/python/py_binding_cpp_def/descriptor_identifier/SISSOClassifier.cpp
+++ b/src/python/py_binding_cpp_def/descriptor_identifier/SISSOClassifier.cpp
@@ -35,7 +35,7 @@ SISSOClassifier::SISSOClassifier(
int n_models_store,
py::list sample_ids_train,
py::list sample_ids_test,
- py::list task_keys
+ py::list task_names
) :
SISSOSolver(
"classification",
@@ -52,7 +52,7 @@ SISSOClassifier::SISSOClassifier(
n_models_store,
sample_ids_train,
sample_ids_test,
- task_keys,
+ task_names,
false
),
_c(100.0),
@@ -76,7 +76,7 @@ SISSOClassifier::SISSOClassifier(
int n_models_store,
py::list sample_ids_train,
py::list sample_ids_test,
- py::list task_keys
+ py::list task_names
) :
SISSOSolver(
"classification",
@@ -93,7 +93,7 @@ SISSOClassifier::SISSOClassifier(
n_models_store,
sample_ids_train,
sample_ids_test,
- task_keys,
+ task_names,
false
),
_c(100.0),
diff --git a/src/python/py_binding_cpp_def/descriptor_identifier/SISSOLogRegressor.cpp b/src/python/py_binding_cpp_def/descriptor_identifier/SISSOLogRegressor.cpp
index f910cde5..53ccf1ac 100644
--- a/src/python/py_binding_cpp_def/descriptor_identifier/SISSOLogRegressor.cpp
+++ b/src/python/py_binding_cpp_def/descriptor_identifier/SISSOLogRegressor.cpp
@@ -35,7 +35,7 @@ SISSOLogRegressor::SISSOLogRegressor(
int n_models_store,
py::list sample_ids_train,
py::list sample_ids_test,
- py::list task_keys,
+ py::list task_names,
bool fix_intercept
) :
SISSORegressor(
@@ -52,7 +52,7 @@ SISSOLogRegressor::SISSOLogRegressor(
n_models_store,
sample_ids_train,
sample_ids_test,
- task_keys,
+ task_names,
fix_intercept
)
{
@@ -86,7 +86,7 @@ SISSOLogRegressor::SISSOLogRegressor(
int n_models_store,
py::list sample_ids_train,
py::list sample_ids_test,
- py::list task_keys,
+ py::list task_names,
bool fix_intercept
) :
SISSORegressor(
@@ -103,7 +103,7 @@ SISSOLogRegressor::SISSOLogRegressor(
n_models_store,
sample_ids_train,
sample_ids_test,
- task_keys,
+ task_names,
fix_intercept
)
{
diff --git a/src/python/py_binding_cpp_def/descriptor_identifier/SISSORegressor.cpp b/src/python/py_binding_cpp_def/descriptor_identifier/SISSORegressor.cpp
index 6f9f8b03..4b023f75 100644
--- a/src/python/py_binding_cpp_def/descriptor_identifier/SISSORegressor.cpp
+++ b/src/python/py_binding_cpp_def/descriptor_identifier/SISSORegressor.cpp
@@ -35,7 +35,7 @@ SISSORegressor::SISSORegressor(
int n_models_store,
py::list sample_ids_train,
py::list sample_ids_test,
- py::list task_keys,
+ py::list task_names,
bool fix_intercept
) :
SISSOSolver(
@@ -53,7 +53,7 @@ SISSORegressor::SISSORegressor(
n_models_store,
sample_ids_train,
sample_ids_test,
- task_keys,
+ task_names,
fix_intercept
)
{}
@@ -72,7 +72,7 @@ SISSORegressor::SISSORegressor(
int n_models_store,
py::list sample_ids_train,
py::list sample_ids_test,
- py::list task_keys,
+ py::list task_names,
bool fix_intercept
) :
SISSOSolver(
@@ -90,7 +90,7 @@ SISSORegressor::SISSORegressor(
n_models_store,
sample_ids_train,
sample_ids_test,
- task_keys,
+ task_names,
fix_intercept
)
{}
diff --git a/src/python/py_binding_cpp_def/descriptor_identifier/SISSOSolver.cpp b/src/python/py_binding_cpp_def/descriptor_identifier/SISSOSolver.cpp
index 4c5cb8e4..3d6ef29a 100644
--- a/src/python/py_binding_cpp_def/descriptor_identifier/SISSOSolver.cpp
+++ b/src/python/py_binding_cpp_def/descriptor_identifier/SISSOSolver.cpp
@@ -36,12 +36,12 @@ SISSOSolver::SISSOSolver(
int n_models_store,
py::list sample_ids_train,
py::list sample_ids_test,
- py::list task_keys,
+ py::list task_names,
bool fix_intercept
) :
_sample_ids_train(python_conv_utils::from_list<std::string>(sample_ids_train)),
_sample_ids_test(python_conv_utils::from_list<std::string>(sample_ids_test)),
- _task_keys(python_conv_utils::from_list<std::string>(task_keys)),
+ _task_names(python_conv_utils::from_list<std::string>(task_names)),
_task_sizes_train(python_conv_utils::from_list<int>(task_sizes_train)),
_task_sizes_test(python_conv_utils::from_list<int>(task_sizes_test)),
_leave_out_inds(python_conv_utils::from_list<int>(leave_out_inds)),
@@ -83,12 +83,12 @@ SISSOSolver::SISSOSolver(
int n_models_store,
py::list sample_ids_train,
py::list sample_ids_test,
- py::list task_keys,
+ py::list task_names,
bool fix_intercept
) :
_sample_ids_train(python_conv_utils::from_list<std::string>(sample_ids_train)),
_sample_ids_test(python_conv_utils::from_list<std::string>(sample_ids_test)),
- _task_keys(python_conv_utils::from_list<std::string>(task_keys)),
+ _task_names(python_conv_utils::from_list<std::string>(task_names)),
_task_sizes_train(python_conv_utils::from_list<int>(task_sizes_train)),
_task_sizes_test(python_conv_utils::from_list<int>(task_sizes_test)),
_leave_out_inds(python_conv_utils::from_list<int>(leave_out_inds)),
diff --git a/src/python/py_interface/get_solver.py b/src/python/py_interface/get_solver.py
index 8e61c0a6..b5dea564 100644
--- a/src/python/py_interface/get_solver.py
+++ b/src/python/py_interface/get_solver.py
@@ -153,7 +153,7 @@ def get_fs_solver(
leave_out_inds,
sample_ids_train,
sample_ids_test,
- task_keys,
+ task_names,
) = read_csv(
df,
prop_key,
@@ -178,7 +178,7 @@ def get_fs_solver(
if loss_type.lower() == "regression":
print(sample_ids_train)
print(sample_ids_test)
- print(task_keys)
+ print(task_names)
solver = SISSORegressor(
fs,
prop_label,
@@ -193,7 +193,7 @@ def get_fs_solver(
n_model_store,
sample_ids_train,
sample_ids_test,
- task_keys,
+ task_names,
)
elif loss_type.lower() == "log_regression":
solver = SISSOLogRegressor(
@@ -210,7 +210,7 @@ def get_fs_solver(
n_model_store,
sample_ids_train,
sample_ids_test,
- task_keys,
+ task_names,
)
else:
solver = SISSOClassifier(
@@ -227,6 +227,6 @@ def get_fs_solver(
n_model_store,
sample_ids_train,
sample_ids_test,
- task_keys,
+ task_names,
)
return fs, solver
diff --git a/src/python/py_interface/import_dataframe.py b/src/python/py_interface/import_dataframe.py
index d4dbdd91..a80a1a71 100644
--- a/src/python/py_interface/import_dataframe.py
+++ b/src/python/py_interface/import_dataframe.py
@@ -138,7 +138,7 @@ def read_csv(
- leave_out_inds (list): Indices to use as the test set
- sample_ids_train (list): List of sample id's for the training data
- sample_ids_test (list): List of sample id's for the test data
- - task_keys (list): List of all task id names
+ - task_names (list): List of all task id names
"""
if not max_rung:
raise ValueError("Maximum rung for the calculation is not defined.")
@@ -152,14 +152,14 @@ def read_csv(
if task_key:
task, _, _ = extract_col(df, task_key)
else:
- task = np.zeros(prop.shape, dtype=np.int64).astype(str)
+ task = np.array(["all"] * len(prop))
# Map out which index belongs to which task and get the size of each task
task_map = {}
- task_keys, task_sizes = np.unique(task, return_counts=True)
+ task_names, task_sizes = np.unique(task, return_counts=True)
task_sizes = task_sizes.astype(np.int32)
- for kk, key in enumerate(task_keys):
+ for kk, key in enumerate(task_names):
task_map[key] = np.where(task == key)[0].astype(np.int32)
assert task_sizes[kk] == len(task_map[key])
@@ -172,21 +172,21 @@ def read_csv(
if leave_out_frac > 0.0:
task_sizes_test = [int(math.ceil(ts * leave_out_frac)) for ts in task_sizes]
- for kk, key in enumerate(task_keys):
+ for kk, key in enumerate(task_names):
leave_out_inds += list(
np.random.choice(task_map[key], task_sizes_test[kk], False).astype(
np.int32
)
)
else:
- task_sizes_test = list(np.zeros(len(task_keys), dtype=np.int32))
+ task_sizes_test = list(np.zeros(len(task_names), dtype=np.int32))
else:
assert (leave_out_frac == 0.0) or (
int(round(len(df) * leave_out_frac)) == len(leave_out_inds)
)
- task_sizes_test = list(np.zeros(len(task_keys), dtype=np.int32))
- for kk, key in enumerate(task_keys):
+ task_sizes_test = list(np.zeros(len(task_names), dtype=np.int32))
+ for kk, key in enumerate(task_names):
left_out = [ind for ind in leave_out_inds if ind in task_map[key]]
task_sizes_test[kk] = len(left_out)
@@ -231,5 +231,5 @@ def read_csv(
leave_out_inds,
list(df.index[train_inds].to_numpy().astype(str)),
list(df.index[leave_out_inds].to_numpy().astype(str)),
- list(task_keys),
+ list(task_names),
)
diff --git a/tests/googletest/descriptor_identification/model/test_model_classifier.cc b/tests/googletest/descriptor_identification/model/test_model_classifier.cc
index 405b7bdf..71086ca7 100644
--- a/tests/googletest/descriptor_identification/model/test_model_classifier.cc
+++ b/tests/googletest/descriptor_identification/model/test_model_classifier.cc
@@ -42,16 +42,24 @@ namespace
_task_sizes_test,
1
);
+
+ _task_keys = {"all"};
+ _sample_ids_train = {"0", "1", "2", "3", "4", "5", "6", "7", "8", "9", "10", "11"};
+ _sample_ids_test = {"20", "21"};
}
- std::vector<int> _leave_out_inds;
- std::vector<int> _task_sizes_train;
- std::vector<int> _task_sizes_test;
+ std::vector<std::string> _sample_ids_train;
+ std::vector<std::string> _sample_ids_test;
+ std::vector<std::string> _task_keys;
std::vector<double> _prop;
std::vector<double> _prop_test;
std::vector<model_node_ptr> _features;
std::shared_ptr<LossFunction> _loss;
+
+ std::vector<int> _task_sizes_train;
+ std::vector<int> _task_sizes_test;
+ std::vector<int> _leave_out_inds;
};
TEST_F(ModelClassifierTests, NodesTest)
@@ -61,7 +69,10 @@ namespace
Unit("m"),
_loss,
_features,
- _leave_out_inds
+ _leave_out_inds,
+ _sample_ids_train,
+ _sample_ids_test,
+ _task_keys
);
EXPECT_STREQ(model.toString().c_str(), "[A]");
EXPECT_EQ(model.n_convex_overlap_train(), 0);
@@ -96,7 +107,7 @@ namespace
EXPECT_EQ(model.n_dim(), 1);
EXPECT_EQ(model.prop_unit(), Unit("m"));
- // boost::filesystem::remove("train_class_mods.dat");
- // boost::filesystem::remove("test_class_mods.dat");
+ boost::filesystem::remove("train_class_mods.dat");
+ boost::filesystem::remove("test_class_mods.dat");
}
}
diff --git a/tests/googletest/descriptor_identification/model/test_model_log_regressor.cc b/tests/googletest/descriptor_identification/model/test_model_log_regressor.cc
index 4c1e43ec..07f37869 100644
--- a/tests/googletest/descriptor_identification/model/test_model_log_regressor.cc
+++ b/tests/googletest/descriptor_identification/model/test_model_log_regressor.cc
@@ -42,7 +42,15 @@ namespace
std::transform(value_1.begin(), value_1.end(), value_2.begin(), _prop.begin(), [](double v1, double v2){return std::log(0.001 * std::pow(v1, 0.1) * std::pow(v2, -2.1));});
std::transform(test_value_1.begin(), test_value_1.end(), test_value_2.begin(), _prop_test.begin(), [](double v1, double v2){return std::log(0.001 * std::pow(v1, 0.1) * std::pow(v2, -2.1));});
+
+ _task_keys = {"all"};
+ _sample_ids_train = {"0", "1", "2", "3", "4", "6", "7", "8", "9", "10"};
+ _sample_ids_test = {"5", "11"};
}
+ std::vector<std::string> _sample_ids_train;
+ std::vector<std::string> _sample_ids_test;
+ std::vector<std::string> _task_keys;
+
std::vector<int> _leave_out_inds;
std::vector<int> _task_sizes_train;
std::vector<int> _task_sizes_test;
@@ -70,7 +78,10 @@ namespace
Unit("m"),
_loss,
_features,
- _leave_out_inds
+ _leave_out_inds,
+ _sample_ids_train,
+ _sample_ids_test,
+ _task_keys
);
EXPECT_STREQ(model.toString().c_str(), "exp(c0) * (A)^a0 * (B)^a1");
EXPECT_LT(model.rmse(), 1e-10);
@@ -156,7 +167,10 @@ namespace
Unit("m"),
_loss,
_features,
- _leave_out_inds
+ _leave_out_inds,
+ _sample_ids_train,
+ _sample_ids_test,
+ _task_keys
);
EXPECT_STREQ(model.toString().c_str(), "(A)^a0 * (B)^a1");
diff --git a/tests/googletest/descriptor_identification/model/test_model_regressor.cc b/tests/googletest/descriptor_identification/model/test_model_regressor.cc
index 031ddd4d..d4eb9012 100644
--- a/tests/googletest/descriptor_identification/model/test_model_regressor.cc
+++ b/tests/googletest/descriptor_identification/model/test_model_regressor.cc
@@ -45,7 +45,15 @@ namespace
std::transform(test_value_1.begin(), test_value_1.begin() + 1, test_value_2.begin(), _prop_test.begin(), [](double v1, double v2){return 0.001 + v1 + v2;});
std::transform(test_value_1.begin() + 1, test_value_1.end(), test_value_2.begin() + 1, _prop_test.begin() + 1, [](double v1, double v2){return -6.5 + 1.25 * v1 - 0.4 * v2;});
+
+ _task_keys = {"task_1", "task_2"};
+ _sample_ids_train = {"0", "1", "2", "3", "4", "6", "7", "8", "9", "10"};
+ _sample_ids_test = {"5", "11"};
}
+ std::vector<std::string> _sample_ids_train;
+ std::vector<std::string> _sample_ids_test;
+ std::vector<std::string> _task_keys;
+
std::vector<int> _leave_out_inds;
std::vector<int> _task_sizes_train;
std::vector<int> _task_sizes_test;
@@ -73,7 +81,10 @@ namespace
Unit("m"),
_loss,
_features,
- _leave_out_inds
+ _leave_out_inds,
+ _sample_ids_train,
+ _sample_ids_test,
+ _task_keys
);
EXPECT_STREQ(model.toString().c_str(), "c0 + a0 * A + a1 * B");
@@ -169,7 +180,10 @@ namespace
Unit("m"),
_loss,
_features,
- _leave_out_inds
+ _leave_out_inds,
+ _sample_ids_train,
+ _sample_ids_test,
+ _task_keys
);
EXPECT_STREQ(model.toString().c_str(), "a0 * A + a1 * B");
diff --git a/tests/googletest/descriptor_identification/sisso_regressor/test_sisso_log_regressor.cc b/tests/googletest/descriptor_identification/sisso_regressor/test_sisso_log_regressor.cc
index c44e494c..431e22aa 100644
--- a/tests/googletest/descriptor_identification/sisso_regressor/test_sisso_log_regressor.cc
+++ b/tests/googletest/descriptor_identification/sisso_regressor/test_sisso_log_regressor.cc
@@ -96,6 +96,16 @@ namespace
_allowed_ops = {"div", "add", "mult", "sub"};
_allowed_param_ops = {};
+ _task_keys = {"all"};
+ for(int ii = 10; ii < 100; ++ii)
+ {
+ _sample_ids_train.push_back(std::to_string(ii));
+ }
+
+ for(int ii = 0; ii < 10; ++ii)
+ {
+ _sample_ids_test.push_back(std::to_string(ii));
+ }
}
std::vector<std::string> _sample_ids_train;
std::vector<std::string> _sample_ids_test;
diff --git a/tests/googletest/descriptor_identification/sisso_regressor/test_sisso_regressor.cc b/tests/googletest/descriptor_identification/sisso_regressor/test_sisso_regressor.cc
index 48a02d57..788b73d1 100644
--- a/tests/googletest/descriptor_identification/sisso_regressor/test_sisso_regressor.cc
+++ b/tests/googletest/descriptor_identification/sisso_regressor/test_sisso_regressor.cc
@@ -116,6 +116,17 @@ namespace
_allowed_ops = {"div", "sq", "cb", "sub"};
_allowed_param_ops = {};
+
+ _task_keys = {"task_1", "task_2"};
+ for(int ii = 10; ii < 100; ++ii)
+ {
+ _sample_ids_train.push_back(std::to_string(ii));
+ }
+
+ for(int ii = 0; ii < 10; ++ii)
+ {
+ _sample_ids_test.push_back(std::to_string(ii));
+ }
}
std::vector<std::string> _sample_ids_train;
std::vector<std::string> _sample_ids_test;
--
GitLab