From 3cf1e5457ee455054a5630b2cb8330e3c78dbd5d Mon Sep 17 00:00:00 2001
From: maagdenbehwvanden1 <h.w.van.den.maagdenberg@lacdr.leidenuniv.nl>
Date: Fri, 2 Feb 2024 15:00:38 +0100
Subject: [PATCH 01/18] add features list to docs

---
 docs/conf.py      |  4 +++-
 docs/features.rst | 56 +++++++++++++++++++++++++++++++++++++++++++++++
 docs/index.rst    |  6 ++++-
 3 files changed, 64 insertions(+), 2 deletions(-)
 create mode 100644 docs/features.rst

diff --git a/docs/conf.py b/docs/conf.py
index 5ccb066d..25d74926 100644
--- a/docs/conf.py
+++ b/docs/conf.py
@@ -27,7 +27,9 @@
     'sphinx.ext.viewcode',
     'sphinx.ext.intersphinx',
     'sphinx.ext.napoleon',
-    'sphinx.ext.viewcode'
+    'sphinx.ext.viewcode',
+    "sphinx_design",
+    "sphinx_design_elements",
 ]
 
 intersphinx_mapping = {'python': ('https://docs.python.org/3.10', None)}
diff --git a/docs/features.rst b/docs/features.rst
new file mode 100644
index 00000000..0195b760
--- /dev/null
+++ b/docs/features.rst
@@ -0,0 +1,56 @@
+.. _features:
+
+Overview of available features
+==============================
+
+.. div:: dropdown-group
+
+    .. dropdown:: Descriptor Sets
+
+        :class:`~qsprpred.data.descriptors.sets.DescriptorSet`: Base class for descriptor sets. 
+        Descriptor sets are used to calculate molecular descriptors for a set of molecules.
+
+        .. tab-set::
+
+            .. tab-item:: Core
+
+                * :class:`~qsprpred.data.descriptors.sets.DrugExPhyschem`: DrugExPhyschem 
+                * :class:`~qsprpred.data.descriptors.sets.PredictorDesc`:PredictorDesc 
+                * :class:`~qsprpred.data.descriptors.sets.RDKitDescs`: RDKitDescs
+                * :class:`~qsprpred.data.descriptors.sets.SmilesDescs`: SmilesDescs
+                * :class:`~qsprpred.data.descriptors.sets.TanimotoDistances`: TanimotoDistances
+                * :class:`~qsprpred.data.descriptors.sets.DataFrameDescriptorSet`: DataFrameDescriptorSet
+                * :class:`~qsprpred.data.descriptors.fingerprints.Fingerprint`: Fingerprint
+                    * :class:`~qsprpred.data.descriptors.fingerprints.AtomPairFP`: AtomPairFP
+                    * :class:`~qsprpred.data.descriptors.fingerprints.AvalonFP`: AvalonFP
+                    * :class:`~qsprpred.data.descriptors.fingerprints.LayeredFP`: LayeredFP
+                    * :class:`~qsprpred.data.descriptors.fingerprints.MACCsFP`: MACCsFP
+                    * :class:`~qsprpred.data.descriptors.fingerprints.MorganFP`: MorganFP
+                    * :class:`~qsprpred.data.descriptors.fingerprints.PatternFP`: PatternFP
+                    * :class:`~qsprpred.data.descriptors.fingerprints.RDKitFP`: RDKitFP
+                    * :class:`~qsprpred.data.descriptors.fingerprints.RDKitMACCSFP`: RDKitMACCSFP
+                    * :class:`~qsprpred.data.descriptors.fingerprints.TopologicalFP`: TopologicalFP
+
+            .. tab-item:: Extra
+
+                * :class:`~qsprpred.extra.data.descriptors.sets.ExtendedValenceSignature`: ExtendedValenceSignature
+                * :class:`~qsprpred.extra.data.descriptors.sets.Mold2`: Mold2
+                * :class:`~qsprpred.extra.data.descriptors.sets.Mordred`: Mordred
+                * :class:`~qsprpred.extra.data.descriptors.sets.PaDEL`: PaDEL
+                * :class:`~qsprpred.extra.data.descriptors.sets.ProteinDescriptorSet`: ProteinDescriptorSet
+                    * :class:`~qsprpred.extra.data.descriptors.sets.ProDec`: ProDec
+                * :class:`~qsprpred.data.descriptors.fingerprints.Fingerprint`: Fingerprint
+                    * :class:`~qsprpred.extra.data.descriptors.fingerprints.CDKAtomPairs2DFP`: CDKAtomPairs2DFP
+                    * :class:`~qsprpred.extra.data.descriptors.fingerprints.CDKEStateFP`: CDKEStateFP
+                    * :class:`~qsprpred.extra.data.descriptors.fingerprints.CDKExtendedFP`: CDKExtendedFP
+                    * :class:`~qsprpred.extra.data.descriptors.fingerprints.CDKFP`: CDKFP
+                    * :class:`~qsprpred.extra.data.descriptors.fingerprints.CDKGraphOnlyFP`: CDKGraphOnlyFP
+                    * :class:`~qsprpred.extra.data.descriptors.fingerprints.CDKKlekotaRothFP`: CDKKlekotaRothFP
+                    * :class:`~qsprpred.extra.data.descriptors.fingerprints.CDKMACCSFP`: CDKMACCSFP
+                    * :class:`~qsprpred.extra.data.descriptors.fingerprints.CDKPubchemFP`: CDKPubchemFP
+                    * :class:`~qsprpred.extra.data.descriptors.fingerprints.CDKSubstructureFP`: CDKSubstructureFP
+
+    .. dropdown:: Dropdown B
+
+        Dropdown content B
+
diff --git a/docs/index.rst b/docs/index.rst
index a7c4ef7a..d0b03ab2 100644
--- a/docs/index.rst
+++ b/docs/index.rst
@@ -6,7 +6,10 @@
 Welcome to QSPRpred's documentation!
 ====================================
 
-QSPRpred provides functionality to assist with building Quantitative Structure Property relationship models. Model built with QSPRpred can also be used as the enviroment reward function in DrugEx. Here you will find the installation guide (:ref:`installation`), usage examples (:ref:`usage`) and API documentation.
+QSPRpred provides functionality to assist with building Quantitative Structure Property relationship models.
+Models built with QSPRpred can also be used as the enviroment reward function in DrugEx.
+Here you will find the installation guide (:ref:`installation`), an overview of available features (:ref:`features`), usage examples (:ref:`usage`) and API documentation (:ref:`api/modules:Python API`).
+
 
 .. toctree::
    :maxdepth: 2
@@ -15,6 +18,7 @@ QSPRpred provides functionality to assist with building Quantitative Structure P
    self
    install
    use
+   features
    api/modules
 
 

From 27604b06e7c697549ec50556dbad550be02aad06 Mon Sep 17 00:00:00 2001
From: maagdenbehwvanden1 <h.w.van.den.maagdenberg@lacdr.leidenuniv.nl>
Date: Fri, 2 Feb 2024 15:06:00 +0100
Subject: [PATCH 02/18] add all categories

---
 docs/features.rst | 35 +++++++++++++++++++++++++++++++++--
 1 file changed, 33 insertions(+), 2 deletions(-)

diff --git a/docs/features.rst b/docs/features.rst
index 0195b760..dbaf8287 100644
--- a/docs/features.rst
+++ b/docs/features.rst
@@ -5,6 +5,14 @@ Overview of available features
 
 .. div:: dropdown-group
 
+    .. dropdown:: Data Sources
+
+        Dropdown content
+
+    .. dropdown:: Data Filters
+
+        Dropdown content
+
     .. dropdown:: Descriptor Sets
 
         :class:`~qsprpred.data.descriptors.sets.DescriptorSet`: Base class for descriptor sets. 
@@ -50,7 +58,30 @@ Overview of available features
                     * :class:`~qsprpred.extra.data.descriptors.fingerprints.CDKPubchemFP`: CDKPubchemFP
                     * :class:`~qsprpred.extra.data.descriptors.fingerprints.CDKSubstructureFP`: CDKSubstructureFP
 
-    .. dropdown:: Dropdown B
+    .. dropdown:: Data Splitters
+
+        Dropdown content
+
+    .. dropdown:: Feature Filters
+
+        Dropdown content
+
+    .. dropdown:: Models
+
+        Dropdown content
+
+    .. dropdown:: Metrics
+
+        Dropdown content
+
+    .. dropdown:: Model Assessors
+
+        Dropdown content 
+
+    .. dropdown:: Hyperparameter Optimizers
+
+        Dropdown content
 
-        Dropdown content B
+    .. dropdown:: Model Plots
 
+        Dropdown content
\ No newline at end of file

From c0d4c29b7b0757722d5a0c8e053d592fc2f32a17 Mon Sep 17 00:00:00 2001
From: maagdenbehwvanden1 <h.w.van.den.maagdenberg@lacdr.leidenuniv.nl>
Date: Fri, 2 Feb 2024 16:09:14 +0100
Subject: [PATCH 03/18] add sources to featurlist

---
 docs/features.rst | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/docs/features.rst b/docs/features.rst
index dbaf8287..5e34af58 100644
--- a/docs/features.rst
+++ b/docs/features.rst
@@ -7,7 +7,14 @@ Overview of available features
 
     .. dropdown:: Data Sources
 
-        Dropdown content
+        :class:`~qsprpred.data.sources.data_source.DataSource`: Base class for data sources.
+        Data sources are used to load data from a source programmatically.
+
+        .. tab-set::
+
+            .. tab-item:: Core
+
+                * :class:`~qsprpred.data.sources.papyrus.papyrus_class.Papyrus`: Papyrus
 
     .. dropdown:: Data Filters
 

From 6ea668b81e815b2a82bf63c0a7438874d0115da8 Mon Sep 17 00:00:00 2001
From: maagdenbehwvanden1 <h.w.van.den.maagdenberg@lacdr.leidenuniv.nl>
Date: Fri, 2 Feb 2024 16:16:48 +0100
Subject: [PATCH 04/18] add data filters to feature list

---
 docs/features.rst | 15 +++++++++++++--
 1 file changed, 13 insertions(+), 2 deletions(-)

diff --git a/docs/features.rst b/docs/features.rst
index 5e34af58..ce505914 100644
--- a/docs/features.rst
+++ b/docs/features.rst
@@ -8,6 +8,7 @@ Overview of available features
     .. dropdown:: Data Sources
 
         :class:`~qsprpred.data.sources.data_source.DataSource`: Base class for data sources.
+
         Data sources are used to load data from a source programmatically.
 
         .. tab-set::
@@ -18,11 +19,21 @@ Overview of available features
 
     .. dropdown:: Data Filters
 
-        Dropdown content
+        :class:`~qsprpred.data.processing.data_filters.DataFilter`: Base class for data filters.
+
+        Data filters are used to filter data based on some criteria.
+
+        .. tab-set::
+
+            .. tab-item:: Core
+                
+                * :class:`~qsprpred.data.processing.data_filters.CategoryFilter`: CategoryFilter
+                * :class:`~qsprpred.data.processing.data_filters.RepeatsFilter`: RepeatsFilter
 
     .. dropdown:: Descriptor Sets
 
-        :class:`~qsprpred.data.descriptors.sets.DescriptorSet`: Base class for descriptor sets. 
+        :class:`~qsprpred.data.descriptors.sets.DescriptorSet`: Base class for descriptor sets.
+        
         Descriptor sets are used to calculate molecular descriptors for a set of molecules.
 
         .. tab-set::

From e393281be1a85a210f44f9063712e641b1942c93 Mon Sep 17 00:00:00 2001
From: maagdenbehwvanden1 <h.w.van.den.maagdenberg@lacdr.leidenuniv.nl>
Date: Fri, 2 Feb 2024 16:38:33 +0100
Subject: [PATCH 05/18] add data splitters to feature list

---
 docs/features.rst | 26 ++++++++++++++++++++++++--
 1 file changed, 24 insertions(+), 2 deletions(-)

diff --git a/docs/features.rst b/docs/features.rst
index ce505914..f90f0e77 100644
--- a/docs/features.rst
+++ b/docs/features.rst
@@ -33,7 +33,7 @@ Overview of available features
     .. dropdown:: Descriptor Sets
 
         :class:`~qsprpred.data.descriptors.sets.DescriptorSet`: Base class for descriptor sets.
-        
+
         Descriptor sets are used to calculate molecular descriptors for a set of molecules.
 
         .. tab-set::
@@ -78,7 +78,29 @@ Overview of available features
 
     .. dropdown:: Data Splitters
 
-        Dropdown content
+        :class:`~qsprpred.data.sampling.splits.DataSplit`: Base class for data splitters.
+
+        Data splitters are used to split data into training and test sets.
+
+        .. tab-set::
+
+            .. tab-item:: Core
+
+                * :class:`~qsprpred.data.sampling.splits.RandomSplit`: RandomSplit
+                * :class:`~qsprpred.data.sampling.splits.ScaffoldSplit`: ScaffoldSplitter
+                * :class:`~qsprpred.data.sampling.splits.TemporalSplit`: StratifiedSplitter
+                * :class:`~qsprpred.data.sampling.splits.ManualSplit`: ManualSplit
+                * :class:`~qsprpred.data.sampling.splits.BootstrapSplit`: BootstrapSplit
+                * :class:`~qsprpred.data.sampling.splits.GBMTDataSplit`: GBMTDataSplit
+                    * :class:`~qsprpred.data.sampling.splits.GBMTRandomSplit`: GBMTRandomSplit
+                    * :class:`~qsprpred.data.sampling.splits.ClusterSplit`: ClusterSplit
+
+            .. tab-item:: Extra
+
+                * :class:`~qsprpred.extra.data.sampling.splits.LeaveTargetsOut`: LeaveTargetsOut
+                * :class:`~qsprpred.extra.data.sampling.splits.PCMSplit`: PCMSplit
+                    * :class:`~qsprpred.extra.data.sampling.splits.TemporalPerTarget`: TemporalPerTarget
+
 
     .. dropdown:: Feature Filters
 

From b1f90e14c6ef94a6f90450ac07aa85691151b455 Mon Sep 17 00:00:00 2001
From: maagdenbehwvanden1 <h.w.van.den.maagdenberg@lacdr.leidenuniv.nl>
Date: Fri, 2 Feb 2024 17:05:20 +0100
Subject: [PATCH 06/18] add feature filters to feature listt

---
 docs/features.rst | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)

diff --git a/docs/features.rst b/docs/features.rst
index f90f0e77..cd8e7005 100644
--- a/docs/features.rst
+++ b/docs/features.rst
@@ -104,7 +104,17 @@ Overview of available features
 
     .. dropdown:: Feature Filters
 
-        Dropdown content
+        :class:`~qsprpred.data.processing.feature_filters.FeatureFilter`: Base class for feature filters.
+
+        Feature filters are used to filter features based on some criteria.
+
+        .. tab-set::
+
+            .. tab-item:: Core
+
+                * :class:`~qsprpred.data.processing.feature_filters.HighCorrelationFilter`: HighCorrelationFilter
+                * :class:`~qsprpred.data.processing.feature_filters.LowVarianceFilter`: LowVarianceFilter
+                * :class:`~qsprpred.data.processing.feature_filters.BorutaFilter`: BorutaFilter
 
     .. dropdown:: Models
 

From 081061585a8b6aebd745c52d64112ef232f12f48 Mon Sep 17 00:00:00 2001
From: maagdenbehwvanden1 <h.w.van.den.maagdenberg@lacdr.leidenuniv.nl>
Date: Fri, 2 Feb 2024 17:54:01 +0100
Subject: [PATCH 07/18] add models to featurelist

---
 docs/features.rst | 19 ++++++++++++++++++-
 1 file changed, 18 insertions(+), 1 deletion(-)

diff --git a/docs/features.rst b/docs/features.rst
index cd8e7005..9686d85b 100644
--- a/docs/features.rst
+++ b/docs/features.rst
@@ -118,7 +118,24 @@ Overview of available features
 
     .. dropdown:: Models
 
-        Dropdown content
+        :class:`~qsprpred.models.models.QSPRModel`: Base class for models.
+
+        Models are used to predict properties of molecules.
+
+        .. tab-set::
+
+            .. tab-item:: Core
+
+                * :class:`~qsprpred.models.scikit_learn.SklearnModel`: SklearnModel
+
+            .. tab-item:: Extra
+
+                * :class:`~qsprpred.extra.models.pcm.PCMModel`: PCMModel
+
+            .. tab-item:: GPU
+                    
+                * :class:`~qsprpred.extra.gpu.models.chemprop.ChempropModel`: ChempropModel
+                * :class:`~qsprpred.extra.gpu.models.pyboost.PyBoostModel`: PyBoostModel
 
     .. dropdown:: Metrics
 

From e878653c19262a3cabef4c3f399bb3d2f6879aab Mon Sep 17 00:00:00 2001
From: maagdenbehwvanden1 <h.w.van.den.maagdenberg@lacdr.leidenuniv.nl>
Date: Fri, 2 Feb 2024 18:03:08 +0100
Subject: [PATCH 08/18] add metrics to featurelist

---
 docs/features.rst | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/docs/features.rst b/docs/features.rst
index 9686d85b..a8cff6c5 100644
--- a/docs/features.rst
+++ b/docs/features.rst
@@ -139,7 +139,15 @@ Overview of available features
 
     .. dropdown:: Metrics
 
-        Dropdown content
+        :class:`~qsprpred.models.metrics.Metric`: Base class for metrics
+
+        Metrics are used to evaluate the performance of models.
+
+        .. tab-set::
+
+            .. tab-item:: Core
+
+                * :class:`~qsprpred.models.metrics.SklearnMetrics`: SklearnMetrics
 
     .. dropdown:: Model Assessors
 

From 164922693b9765af5c587ddd0d2058a0b0c468a4 Mon Sep 17 00:00:00 2001
From: maagdenbehwvanden1 <h.w.van.den.maagdenberg@lacdr.leidenuniv.nl>
Date: Fri, 2 Feb 2024 18:06:37 +0100
Subject: [PATCH 09/18] add model assessors to feature list

---
 docs/features.rst | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/docs/features.rst b/docs/features.rst
index a8cff6c5..40091b28 100644
--- a/docs/features.rst
+++ b/docs/features.rst
@@ -151,7 +151,16 @@ Overview of available features
 
     .. dropdown:: Model Assessors
 
-        Dropdown content 
+        :class:`~qsprpred.models.assessment_methods.ModelAssessor`: Base class for model assessors.
+
+        Model assessors are used to assess the performance of models.
+
+        .. tab-set::
+
+            .. tab-item:: Core
+
+                * :class:`~qsprpred.models.assessment_methods.CrossValAssessor`: CrossValAssessor
+                * :class:`~qsprpred.models.assessment_methods.TestSetAssessor`: TestSetAssessor
 
     .. dropdown:: Hyperparameter Optimizers
 

From b753c18ec8988433db2a8d58c5a048f04177c725 Mon Sep 17 00:00:00 2001
From: maagdenbehwvanden1 <h.w.van.den.maagdenberg@lacdr.leidenuniv.nl>
Date: Fri, 2 Feb 2024 18:37:50 +0100
Subject: [PATCH 10/18] add hyperparam optimizers to feature list

---
 docs/features.rst | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)

diff --git a/docs/features.rst b/docs/features.rst
index 40091b28..3bc8d2fe 100644
--- a/docs/features.rst
+++ b/docs/features.rst
@@ -164,7 +164,17 @@ Overview of available features
 
     .. dropdown:: Hyperparameter Optimizers
 
-        Dropdown content
+        :class:`~qsprpred.models.hyperparam_optimization.HyperparameterOptimizion`: Base class for hyperparameter optimizers.
+
+        Hyperparameter optimizers are used to optimize the hyperparameters of models.
+
+        .. tab-set::
+
+            .. tab-item:: Core
+
+                * :class:`~qsprpred.models.hyperparam_optimization.GridSearchOptimization`: GridSearchOptimization
+                * :class:`~qsprpred.models.hyperparam_optimization.OptunaOptimization`: OptunaOptimization
+
 
     .. dropdown:: Model Plots
 

From 8e65c07aedf38f5b5129c8465d631b7f2e237be4 Mon Sep 17 00:00:00 2001
From: maagdenbehwvanden1 <h.w.van.den.maagdenberg@lacdr.leidenuniv.nl>
Date: Fri, 2 Feb 2024 18:50:55 +0100
Subject: [PATCH 11/18] add model plot

---
 docs/features.rst | 21 +++++++++++++++++++--
 1 file changed, 19 insertions(+), 2 deletions(-)

diff --git a/docs/features.rst b/docs/features.rst
index 3bc8d2fe..4ba90f7e 100644
--- a/docs/features.rst
+++ b/docs/features.rst
@@ -164,7 +164,7 @@ Overview of available features
 
     .. dropdown:: Hyperparameter Optimizers
 
-        :class:`~qsprpred.models.hyperparam_optimization.HyperparameterOptimizion`: Base class for hyperparameter optimizers.
+        :class:`~qsprpred.models.hyperparam_optimization.HyperparameterOptimization`: Base class for hyperparameter optimizers.
 
         Hyperparameter optimizers are used to optimize the hyperparameters of models.
 
@@ -178,4 +178,21 @@ Overview of available features
 
     .. dropdown:: Model Plots
 
-        Dropdown content
\ No newline at end of file
+        :class:`~qsprpred.plotting.base_plot.ModelPlot`: Base class for model plots.
+
+        Model plots are used to visualize the performance of models.
+
+        .. tab-set::
+
+            .. tab-item:: Core
+
+                * :class:`~qsprpred.plotting.regression.RegressionPlot`: RegressionPlot
+                    * :class:`~qsprpred.plotting.regression.CorrelationPlot`: CorrelationPlot
+                    * :class:`~qsprpred.plotting.regression.WilliamsPlot`: WilliamsPlot
+                * :class:`~qsprpred.plotting.classification.ClassifierPlot`: ClassifierPlot
+                    * :class:`~qsprpred.plotting.classification.ROCPlot`: ROCPlot
+                    * :class:`~qsprpred.plotting.classification.PRCPlot`: PRCPlot
+                    * :class:`~qsprpred.plotting.classification.CalibrationPlot`: CalibrationPlot
+                    * :class:`~qsprpred.plotting.classification.MetricsPlot`: MetricsPlot
+                    * :class:`~qsprpred.plotting.classification.ConfusionMatrixPlot`: ConfusionMatrixPlot
+

From 729140b0e74d53bbe6db8ea628c42189fe48f78f Mon Sep 17 00:00:00 2001
From: maagdenbehwvanden1 <h.w.van.den.maagdenberg@lacdr.leidenuniv.nl>
Date: Mon, 5 Feb 2024 12:49:31 +0100
Subject: [PATCH 12/18] add tutorial link to features

---
 docs/features.rst | 1 +
 1 file changed, 1 insertion(+)

diff --git a/docs/features.rst b/docs/features.rst
index 4ba90f7e..9bdfc019 100644
--- a/docs/features.rst
+++ b/docs/features.rst
@@ -10,6 +10,7 @@ Overview of available features
         :class:`~qsprpred.data.sources.data_source.DataSource`: Base class for data sources.
 
         Data sources are used to load data from a source programmatically.
+        An example may be found in the `tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/data/data_collection_with_papyrus.ipynb>`_.
 
         .. tab-set::
 

From b1b036af3c1c48821b8d935298b5a29b0ee9d6f1 Mon Sep 17 00:00:00 2001
From: maagdenbehwvanden1 <h.w.van.den.maagdenberg@lacdr.leidenuniv.nl>
Date: Mon, 5 Feb 2024 14:06:57 +0100
Subject: [PATCH 13/18] add tutorial links

---
 docs/features.rst | 19 +++++++++++++++++--
 1 file changed, 17 insertions(+), 2 deletions(-)

diff --git a/docs/features.rst b/docs/features.rst
index 9bdfc019..dcb0109f 100644
--- a/docs/features.rst
+++ b/docs/features.rst
@@ -10,7 +10,7 @@ Overview of available features
         :class:`~qsprpred.data.sources.data_source.DataSource`: Base class for data sources.
 
         Data sources are used to load data from a source programmatically.
-        An example may be found in the `tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/data/data_collection_with_papyrus.ipynb>`_.
+        An example may be found in the `data collection with Papyrus tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/data/data_collection_with_papyrus.ipynb>`_.
 
         .. tab-set::
 
@@ -23,6 +23,7 @@ Overview of available features
         :class:`~qsprpred.data.processing.data_filters.DataFilter`: Base class for data filters.
 
         Data filters are used to filter data based on some criteria.
+        Examples can be found in the `data preparation tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/data/data_preparation.ipynb>`_.
 
         .. tab-set::
 
@@ -36,6 +37,7 @@ Overview of available features
         :class:`~qsprpred.data.descriptors.sets.DescriptorSet`: Base class for descriptor sets.
 
         Descriptor sets are used to calculate molecular descriptors for a set of molecules.
+        Examples can be found in the `descriptor calculation tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/data/descriptors.ipynb>`_.
 
         .. tab-set::
 
@@ -82,6 +84,7 @@ Overview of available features
         :class:`~qsprpred.data.sampling.splits.DataSplit`: Base class for data splitters.
 
         Data splitters are used to split data into training and test sets.
+        Examples may be found in the `data splitting tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/data/data_splitting.ipynb>`_.
 
         .. tab-set::
 
@@ -108,6 +111,7 @@ Overview of available features
         :class:`~qsprpred.data.processing.feature_filters.FeatureFilter`: Base class for feature filters.
 
         Feature filters are used to filter features based on some criteria.
+        Examples can be found in the `data preparation tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/data/data_preparation.ipynb>`_.
 
         .. tab-set::
 
@@ -122,20 +126,27 @@ Overview of available features
         :class:`~qsprpred.models.models.QSPRModel`: Base class for models.
 
         Models are used to predict properties of molecules.
+        A general example may be found in the `quick start tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/quick_start.ipynb>`_.
+        More detailed information can be found throughouth the basic and advanced modelling tutorials.
 
         .. tab-set::
 
             .. tab-item:: Core
 
                 * :class:`~qsprpred.models.scikit_learn.SklearnModel`: SklearnModel
+                An example may be found in the `quick start tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/quick_start.ipynb>`_.
 
             .. tab-item:: Extra
 
                 * :class:`~qsprpred.extra.models.pcm.PCMModel`: PCMModel
+                More information may be found in the `PCM tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/advanced/modelling/PCM_modelling.ipynb>`_.
 
             .. tab-item:: GPU
-                    
+                More information may be found in the `deep learning tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/advanced/modelling/deep_learning_models.ipynb>`_.
+                
+                * :class:`~qsprpred.extra.gpu.models.dnn.DNNModel`: DNNModel
                 * :class:`~qsprpred.extra.gpu.models.chemprop.ChempropModel`: ChempropModel
+                More information may be found in the `Chemprop tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/advanced/modelling/chemprop_models.ipynb>`_.
                 * :class:`~qsprpred.extra.gpu.models.pyboost.PyBoostModel`: PyBoostModel
 
     .. dropdown:: Metrics
@@ -143,6 +154,7 @@ Overview of available features
         :class:`~qsprpred.models.metrics.Metric`: Base class for metrics
 
         Metrics are used to evaluate the performance of models.
+        More information may be found in the `model assessment tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/modelling/model_assessment.ipynb>`_.
 
         .. tab-set::
 
@@ -155,6 +167,7 @@ Overview of available features
         :class:`~qsprpred.models.assessment_methods.ModelAssessor`: Base class for model assessors.
 
         Model assessors are used to assess the performance of models.
+        Examples may be found in the `model assessment tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/modelling/model_assessment.ipynb>`_.
 
         .. tab-set::
 
@@ -168,6 +181,7 @@ Overview of available features
         :class:`~qsprpred.models.hyperparam_optimization.HyperparameterOptimization`: Base class for hyperparameter optimizers.
 
         Hyperparameter optimizers are used to optimize the hyperparameters of models.
+        More information may be found in the `hyperparameter optimization tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/advanced/modelling/hyperparameter_optimization.ipynb>`_.
 
         .. tab-set::
 
@@ -182,6 +196,7 @@ Overview of available features
         :class:`~qsprpred.plotting.base_plot.ModelPlot`: Base class for model plots.
 
         Model plots are used to visualize the performance of models.
+        Examples can be found throughout the basic and advanced modelling tutorials.
 
         .. tab-set::
 

From 0750af44bb240c10ae2adf45d12eb41aa93dbd0b Mon Sep 17 00:00:00 2001
From: maagdenbehwvanden1 <h.w.van.den.maagdenberg@lacdr.leidenuniv.nl>
Date: Mon, 5 Feb 2024 14:58:12 +0100
Subject: [PATCH 14/18] fix tutorial references

---
 docs/features.rst | 25 +++++++++++--------------
 1 file changed, 11 insertions(+), 14 deletions(-)

diff --git a/docs/features.rst b/docs/features.rst
index dcb0109f..cebe3cde 100644
--- a/docs/features.rst
+++ b/docs/features.rst
@@ -10,13 +10,12 @@ Overview of available features
         :class:`~qsprpred.data.sources.data_source.DataSource`: Base class for data sources.
 
         Data sources are used to load data from a source programmatically.
-        An example may be found in the `data collection with Papyrus tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/data/data_collection_with_papyrus.ipynb>`_.
 
         .. tab-set::
 
             .. tab-item:: Core
 
-                * :class:`~qsprpred.data.sources.papyrus.papyrus_class.Papyrus`: Papyrus
+                * :class:`~qsprpred.data.sources.papyrus.papyrus_class.Papyrus`: Papyrus (See `data collection with Papyrus tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/data/data_collection_with_papyrus.ipynb>`_.)
 
     .. dropdown:: Data Filters
 
@@ -84,7 +83,7 @@ Overview of available features
         :class:`~qsprpred.data.sampling.splits.DataSplit`: Base class for data splitters.
 
         Data splitters are used to split data into training and test sets.
-        Examples may be found in the `data splitting tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/data/data_splitting.ipynb>`_.
+        Examples can be found in the `data splitting tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/data/data_splitting.ipynb>`_.
 
         .. tab-set::
 
@@ -126,27 +125,25 @@ Overview of available features
         :class:`~qsprpred.models.models.QSPRModel`: Base class for models.
 
         Models are used to predict properties of molecules.
-        A general example may be found in the `quick start tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/quick_start.ipynb>`_.
-        More detailed information can be found throughouth the basic and advanced modelling tutorials.
+        A general example can be found in the `quick start tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/quick_start.ipynb>`_.
+        More detailed information can be found throughout the basic and advanced modelling tutorials.
 
         .. tab-set::
 
             .. tab-item:: Core
 
                 * :class:`~qsprpred.models.scikit_learn.SklearnModel`: SklearnModel
-                An example may be found in the `quick start tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/quick_start.ipynb>`_.
 
             .. tab-item:: Extra
 
-                * :class:`~qsprpred.extra.models.pcm.PCMModel`: PCMModel
-                More information may be found in the `PCM tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/advanced/modelling/PCM_modelling.ipynb>`_.
+                * :class:`~qsprpred.extra.models.pcm.PCMModel`: PCMModel (See `PCM tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/advanced/modelling/PCM_modelling.ipynb>`_.)
 
             .. tab-item:: GPU
-                More information may be found in the `deep learning tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/advanced/modelling/deep_learning_models.ipynb>`_.
+                
+                More information can be found in the `deep learning tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/advanced/modelling/deep_learning_models.ipynb>`_.
                 
                 * :class:`~qsprpred.extra.gpu.models.dnn.DNNModel`: DNNModel
-                * :class:`~qsprpred.extra.gpu.models.chemprop.ChempropModel`: ChempropModel
-                More information may be found in the `Chemprop tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/advanced/modelling/chemprop_models.ipynb>`_.
+                * :class:`~qsprpred.extra.gpu.models.chemprop.ChempropModel`: ChempropModel (See `Chemprop tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/advanced/modelling/chemprop_models.ipynb>`_.)
                 * :class:`~qsprpred.extra.gpu.models.pyboost.PyBoostModel`: PyBoostModel
 
     .. dropdown:: Metrics
@@ -154,7 +151,7 @@ Overview of available features
         :class:`~qsprpred.models.metrics.Metric`: Base class for metrics
 
         Metrics are used to evaluate the performance of models.
-        More information may be found in the `model assessment tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/modelling/model_assessment.ipynb>`_.
+        More information can be found in the `model assessment tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/modelling/model_assessment.ipynb>`_.
 
         .. tab-set::
 
@@ -167,7 +164,7 @@ Overview of available features
         :class:`~qsprpred.models.assessment_methods.ModelAssessor`: Base class for model assessors.
 
         Model assessors are used to assess the performance of models.
-        Examples may be found in the `model assessment tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/modelling/model_assessment.ipynb>`_.
+        More information be found in the `model assessment tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/modelling/model_assessment.ipynb>`_.
 
         .. tab-set::
 
@@ -181,7 +178,7 @@ Overview of available features
         :class:`~qsprpred.models.hyperparam_optimization.HyperparameterOptimization`: Base class for hyperparameter optimizers.
 
         Hyperparameter optimizers are used to optimize the hyperparameters of models.
-        More information may be found in the `hyperparameter optimization tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/advanced/modelling/hyperparameter_optimization.ipynb>`_.
+        More information can be found in the `hyperparameter optimization tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/advanced/modelling/hyperparameter_optimization.ipynb>`_.
 
         .. tab-set::
 

From 9e1074ea2172779bf20262d33aee9298fd96fe12 Mon Sep 17 00:00:00 2001
From: maagdenbehwvanden1 <h.w.van.den.maagdenberg@lacdr.leidenuniv.nl>
Date: Mon, 5 Feb 2024 15:37:22 +0100
Subject: [PATCH 15/18] update documentation installation and index

---
 docs/{use.rst => cli_usage.rst} |  9 ++++-----
 docs/conf.py                    |  1 +
 docs/index.rst                  | 10 +++++-----
 docs/install.rst                | 21 ++++++++++++++++++++-
 4 files changed, 30 insertions(+), 11 deletions(-)
 rename docs/{use.rst => cli_usage.rst} (99%)

diff --git a/docs/use.rst b/docs/cli_usage.rst
similarity index 99%
rename from docs/use.rst
rename to docs/cli_usage.rst
index 79f4ea33..b5442eb5 100644
--- a/docs/use.rst
+++ b/docs/cli_usage.rst
@@ -1,7 +1,7 @@
-..  _usage:
+..  _cli-usage:
 
-Usage
-=====
+Command Line Interface Usage
+============================
 
 You can use the command-line interface to preprocess data and build models.
 The description of the functionality can be displayed with the :code:`--help` argument,
@@ -18,10 +18,9 @@ you can also use the Python API directly (see `source code <https://github.com/C
 Here you can find a tutorial with a Jupyter notebook illustrating some common use cases in the project source code.
 Make sure to download the tutorial folder to follow the examples in this CLI tutorial.
 
-..  _cli-example:
 
 CLI Example
-===========
+***********
 
 In this example, we will use the command line utilities of QSPRpred to train a QSAR model for the GABAA receptor and the 
 Glutamate receptor NMDA. We will use the parkinsons dataset from the API tutorial. 
diff --git a/docs/conf.py b/docs/conf.py
index 25d74926..2d28931e 100644
--- a/docs/conf.py
+++ b/docs/conf.py
@@ -28,6 +28,7 @@
     'sphinx.ext.intersphinx',
     'sphinx.ext.napoleon',
     'sphinx.ext.viewcode',
+    'sphinx.ext.autosectionlabel',
     "sphinx_design",
     "sphinx_design_elements",
 ]
diff --git a/docs/index.rst b/docs/index.rst
index d0b03ab2..933b1f02 100644
--- a/docs/index.rst
+++ b/docs/index.rst
@@ -6,10 +6,10 @@
 Welcome to QSPRpred's documentation!
 ====================================
 
-QSPRpred provides functionality to assist with building Quantitative Structure Property relationship models.
-Models built with QSPRpred can also be used as the enviroment reward function in DrugEx.
-Here you will find the installation guide (:ref:`installation`), an overview of available features (:ref:`features`), usage examples (:ref:`usage`) and API documentation (:ref:`api/modules:Python API`).
-
+QSPRpred is open-source software libary for building **Quantitative Structure Property Relationship (QSPR)** models developed by Gerard van Westen's Computational Drug
+Discovery group. It provides a unified interface for building QSPR models based on different types of descriptors and machine learning algorithms.
+Here you will find the installation guide (:ref:`installation`), an overview of available features (:ref:`features`), usage examples (:ref:`cli_usage`) and API documentation (:ref:`Python API`).
+For tutorials and examples of the Python API, please visit the `QSPRpred GitHub repository <https://github.com/CDDLeiden/QSPRpred>`_.
 
 .. toctree::
    :maxdepth: 2
@@ -17,7 +17,7 @@ Here you will find the installation guide (:ref:`installation`), an overview of
 
    self
    install
-   use
+   cli_usage
    features
    api/modules
 
diff --git a/docs/install.rst b/docs/install.rst
index b4136b5f..9b4d107c 100644
--- a/docs/install.rst
+++ b/docs/install.rst
@@ -3,7 +3,7 @@
 Installation
 ============
 
-You do not need anything special to install the package. Just run the following to get the latest version and all dependencies:
+You do not need anything special to install the package . Just run the following (with python >= 3.10) to get the latest version and all basic dependencies:
 
 ..  code-block::
 
@@ -11,6 +11,25 @@ You do not need anything special to install the package. Just run the following
 
 You can also get tags and development snapshots by varying the :code:`@main` part (i.e. :code:`@1.0.0`). After that you can start building models (see :ref:`usage`).
 
+Note that this will install the basic dependencies, but not the optional dependencies.
+If you want to use the optional dependencies, you can install the package with an
+option:
+
+..  code-block::
+    
+    pip install git+https://github.com/CDDLeiden/QSPRpred.git@main#egg=qsprpred[<option>]
+
+The following options are available:
+
+- extra : include extra dependencies for PCM models and extra descriptor sets from
+  packages other than RDKit
+- deep : include deep learning models (torch and chemprop)
+- pyboost : include pyboost model (requires cupy, ``pip install cupy-cudaX``, replace X
+  with your `cuda version <https://docs.cupy.dev/en/stable/install.html>`_, you can obtain
+  cude toolkit from Anaconda as well: ``conda install cudatoolkit``)
+- full : include all optional dependecies (requires cupy, ``pip install cupy-cudaX``,
+  replace X with your `cuda version <https://docs.cupy.dev/en/stable/install.html>`_)
+
 You can test the installation by running the unit test suite:
 
 ..  code-block::

From d4784dbdcb2133fd8dac9cb69770adb9d231cd96 Mon Sep 17 00:00:00 2001
From: maagdenbehwvanden1 <h.w.van.den.maagdenberg@lacdr.leidenuniv.nl>
Date: Mon, 5 Feb 2024 19:09:45 +0100
Subject: [PATCH 16/18] update CLI documentation and CLI

---
 .gitignore                                    |   1 +
 CHANGELOG.md                                  |   1 +
 README.md                                     |   8 -
 docs/cli_usage.rst                            | 214 ++++++++++++------
 docs/features.rst                             |  24 +-
 qsprpred/data_CLI.py                          |  37 +--
 .../tutorial_data/create_tutorial_data.py     |  25 +-
 7 files changed, 200 insertions(+), 110 deletions(-)

diff --git a/.gitignore b/.gitignore
index c883d882..85514872 100644
--- a/.gitignore
+++ b/.gitignore
@@ -17,6 +17,7 @@
 /**/tutorials/tutorial_data/papyrus/
 /**/tutorials/tutorial_data/A2A_LIGANDS.tsv
 /**/tutorials/tutorial_data/AR_LIGANDS.tsv
+/**/tutorials/tutorial_data/AR_LIGANDS_pivot.tsv
 /**/tutorials/tutorial_output
 
 ### Python template
diff --git a/CHANGELOG.md b/CHANGELOG.md
index ad36c7a3..b8b7d20d 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -25,6 +25,7 @@ From v2.1.1 to v3.0.0
   or equal to `v2.1.0`.
 - Fixes to the  `fromMolTable` method in various data set implementations, in particular
   in copying of the feature standardizer and other settings.
+- Fixed not working `cluster` split and `--imputation` from `data_CLI.py`.
 
 ## Changes
 
diff --git a/README.md b/README.md
index 50e2a595..4f9c503b 100644
--- a/README.md
+++ b/README.md
@@ -95,14 +95,6 @@ document the use of the Python API to build different types of models. The tutor
 well as the documentation are still work in progress, and we will be happy for any
 contributions where it is still lacking.
 
-To use the commandline to train the same QSAR model as in the tutorial use (run from
-tutorial folder):
-
-```bash
-python -m qsprpred.data_CLI -i ./data/parkinsons_pivot.tsv -o qspr/data -pr GABAAalpha -pr NMDA -r true -sp random -sf 0.15 -fe Morgan
-python -m qsprpred.model_CLI -dp ./qspr/data/GABAAalpha_REGRESSION_df.pkl -o ./qspr/models -m PLS -o bayes -nt 5 -me -s
-```
-
 Contributions
 =============
 
diff --git a/docs/cli_usage.rst b/docs/cli_usage.rst
index b5442eb5..f8039eae 100644
--- a/docs/cli_usage.rst
+++ b/docs/cli_usage.rst
@@ -14,7 +14,7 @@ e.g. the help message for the :code:`QSPRpred.data_CLI` script can be shown as f
 A simple command-line workflow to prepare your dataset and train QSPR models is given below (see :ref:`cli-example`).
 
 If you want more control over the inputs and outputs or want to customize QSPRpred for your purpose,
-you can also use the Python API directly (see `source code <https://github.com/CDDLeiden/QSPRpred/tree/main/tutorial>`).
+you can also use the Python API directly (see `source code <https://github.com/CDDLeiden/QSPRpred/tree/main/tutorials>`).
 Here you can find a tutorial with a Jupyter notebook illustrating some common use cases in the project source code.
 Make sure to download the tutorial folder to follow the examples in this CLI tutorial.
 
@@ -22,13 +22,32 @@ Make sure to download the tutorial folder to follow the examples in this CLI tut
 CLI Example
 ***********
 
-In this example, we will use the command line utilities of QSPRpred to train a QSAR model for the GABAA receptor and the 
-Glutamate receptor NMDA. We will use the parkinsons dataset from the API tutorial. 
-Make sure to first run the Parkinsons function in the tutorial/datasets.py script (see dataset preparation in the API tutorial)
-to create the tutorial/data/parkinsons_pivot.tsv file.
-QSPRpred assumes that all input data are saved in a subfolder data of the base directory..
+In this example, we will use the command line utilities of QSPRpred to train a QSAR model for the A2A and A2B receptor.
+We will use the Adenosine dataset from the API tutorial. The data is available through [OneDrive](https://1drv.ms/u/s!AtzWqu0inkjX3QRxXOkTFNv7IV7u?e=PPj0O2)
+(just unzip and place the two datasets `A2A_LIGANDS.tsv` and `AR_LIGANDS.tsv` in the `tutorial_data` folder) or recreate the dataset yourself by running `tutorial_data/create_tutorial_data.py`.
 Input data should contain a column with SMILES sequences and at least one column with a property for modelling.
-Therefore, our example dataset 'parkinsons_pivot.tsv' is located in a subfolder 'data' of the tutorial directory.
+The `AR_LIGANDS.tsv` file contains a SMILES column, a column with the property 'pchembl_value_mean' and a column with the property 'Accession` (uniprot accession numbers).
+However, to create models for the A2A and A2B receptor, we need to have the data in a pivot table format, where the properties are in the columns.
+To create this pivot table, we can use the `pandas` library in Python as follows:
+
+..  code-block::
+
+    import pandas as pd
+    df = pd.read_csv('tutorial_data/AR_LIGANDS.tsv', sep='\t')
+    df = df.pivot(index="SMILES", columns="accession", values="pchembl_value_Mean")
+    df.columns.name = None
+    df.reset_index(inplace=True)
+    df.to_csv('tutorial_data/AR_LIGANDS_pivot.tsv', sep='\t')
+
+It is also possible to simply run the `create_tutorial_data.py` script in the `tutorial_data` folder with the following command:
+
+..  code-block::
+
+    python create_tutorial_data.py -m AR_LIGANDS.tsv
+
+This will create a pivot table with the name `AR_LIGANDS_pivot.tsv` in the `tutorial_data` folder.
+Our example dataset now contains a SMILES column and two columns with the properties **P29274** (A2AR) and **P29275** (A2BR) 
+(as well as a columns for A1 and A3, which we will not use in this example).
 
 Preparing Data
 --------------
@@ -36,73 +55,87 @@ Preparing Data
 Basics
 ^^^^^^
 We will now use the QSPRpred :code:`QSPRpred.data_CLI` script for data preparation.
-In the CLI we need to indicate which property/ies we are interested in predicting (here GABAAalpha and NMDA),
+In the CLI we need to indicate which property/ies we are interested in predicting (here P29274 and P29275, the A2A and A2B receptor respectively),
 this should be equal to the column names (should not contain spaces) containing the values to be predicted. 
 For regression models these columns should contain numerical datapoints.
 For categorical models either categorical data or numerical data can be used (the latter will be categorized based on the activity threshold).
 Furthermore, we should indicate how we wish to split the data to create a train and test set.
 Here we will use a random split with a test fraction of 15%. We need to calculate features that describe the molecules,
-here we use morgan fingerprints.
+here we use Morgan fingerprints.
 
 ..  code-block::
 
-    # input is in ./data/parkinsons_pivot.tsv
-    python -m qsprpred.data_CLI -i ./data/parkinsons_pivot.tsv -o ./cli_tutorial/data -pr GABAAalpha -pr NMDA -r REG -sp random -sf 0.15 -fe Morgan
+    # input is in tutorial/data/AR_LIGANDS_pivot.tsv
+    python -m qsprpred.data_CLI -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/data -pr P29274 -pr P29275 -r REG -sp random -sf 0.15 -fe Morgan
+
+Running this command will create folders 'tutorial_output/data', with subfolders 'P29274_REGRESSION' and 'P29275_REGRESSION' containing the prepared data.
+Each subfolder, named by the property identifiers (i.e. P29274/P29275) and the model task (REGRESSION),
+will contain the following files:
 
-Running this command will create folders 'cli_tutorial/data'. The cli_tutorial/data folder should contain 5 files for
-each property (10 in total), prefixed by the property identifiers (i.e. GABAAalpha/NMDA), the model task (REGRESSION),
-e.g. GABAAalpha_REGRESSION. As well as an log file and a run settings file.
++----------------------------------------------------+--------------------------------------------------------+
+| File                                               | Function                                               |
++====================================================+========================================================+
+|| {prefixes}_df.pkl                                 || Dataframe                                             |
+|| {prefixes}_meta.json                              || Meta data, also used to instantiate a QSPRData object |
+|| {prefixes}_MorganFP                               || Descriptor set folder                                 |
+|| {prefixes}_MorganFP/{prefixes}_MorganFP_df.pkl    || Calculated descriptors                                |
+|| {prefixes}_MorganFP/{prefixes}_MorganFP_meta.json || Meta data of the descriptor set                       |
++----------------------------------------------------+--------------------------------------------------------+
+
+Furthermore, the command line interface will create a log file and settings file in the output folder.
 
 +--------------------------------------------------+-------------------------------------------------------+
 | File                                             | Function                                              |
 +==================================================+=======================================================+
-|| {prefixes}_df.pkl                               || Dataframe                                            |
-|| {prefixes}_descriptor_calculator_Descriptor.json|| re-instantiate descriptor calculator                 |
-|| {prefixes}_Descriptor_df.pkl                    || re-instantiate descriptor dataframe                  |
-|| {prefixes}_feature_standardizer.json            || re-instantiate feature standardizer                  |
-|| {prefixes}_meta.json                            || Meta data, also used to instantiate QSAR data object |
 || QSPRdata.json                                   || Command Line interface settings                      |
 || QSPRdata.log                                    || Log file                                             |
 +--------------------------------------------------+-------------------------------------------------------+
 
-
 More
 ^^^^
 Run settings arguments
 ^^^^^^^^^^^^^^^^^^^^^^^
 Apart from the the input file name, there are a few other base options that can be set.
-The base-directory can be specified using `-b`. Including `-d`, will print debug information to the log file. The random
-seed (-ran) can also be set manually, which should guarantee identical results while keeping the same random seed.
-Furthermore, the number of cpu's (-ncpu) used for model training. Finally, the name of the smilescolumn in your dataset
-can be indicated with `-sm` (default SMILES).
+Including ``-d``, will print debug information to the log file. The random
+seed (``-ran``) can also be set manually, which should guarantee identical results while keeping the same random seed.
+Furthermore, the number of cpu's (-ncpu) used for model training. Finally, the name of the SMILES column in your dataset
+can be indicated with ``-sm`` (default SMILES).
 
 ..  code-block::
 
-    # input is in tutorial/data/parkinsons_pivot.tsv, setting debug flag, smiles column, random seed, number of cpu's
-        python -m qsprpred.data_CLI -b tutorial -i ./data/parkinsons_pivot.tsv -o ./cli_tutorial/data -sm SMILES -de -ran 42 -ncpu 5 -pr GABAAalpha -pr NMDA -r REG -sp random -sf 0.15 -fe Morgan
+    # input is in tutorial/data/AR_LIGANDS_pivot.tsv, setting debug flag, smiles column, random seed, number of cpu's
+        python -m qsprpred.data_CLI  -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/data -sm SMILES -de -ran 42 -ncpu 5 -pr P29274 -pr P29275 -r REG -sp random -sf 0.15 -fe Morgan
 
 
 Transform target property
 """""""""""""""""""""""""
-To apply (-tr) transformations to target properties, indicate this in the CLI as follows:
+To apply (``-tr``) transformations to target properties, indicate this in the CLI as follows:
 
 ..  code-block::
 
-    # Log transform data for GABAAalpha
-        python -m qsprpred.data_CLI -i ./data/parkinsons_pivot.tsv -o ./cli_tutorial/data -pr GABAAalpha -pr NMDA -tr '{"GABAAalpha":"log"}' -r REG -sp random -sf 0.15 -fe Morgan
+    # Log transform data for P29274
+        python -m qsprpred.data_CLI -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/data -pr P29274 -pr P29275 -tr '{"P29274":"log"}' -r REG -sp random -sf 0.15 -fe Morgan
 
-note. on windows remove the single quotes around the dictionary and add backslashes before the double quotes, e.g. -tt {\"GABAAalpha\":\"log\"}
+note. on windows remove the single quotes around the dictionary and add backslashes before the double quotes, e.g. ``-tr {\"P29274\":\"log\"}``
 
-Train test split
+Train-test split
 """"""""""""""""
-In the base example we use a random split to create the train and test set. There are two more options,
-namely a scaffold split, where the data is split into a test and train set randomly but keeping molecules with the same 
+In the base example we use a random split to create the train and test set. There are several more options,
+One is a scaffold split, where the data is split into a test and train set randomly but keeping molecules with the same 
 (Murcko) scaffold in the same set.
 
 ..  code-block::
 
     # Scaffold split
-        python -m qsprpred.data_CLI -i ./data/parkinsons_pivot.tsv -o ./cli_tutorial/data -pr GABAAalpha -pr NMDA -r REG -sp scaffold -sf 0.15 -fe Morgan
+        python -m qsprpred.data_CLI -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/data -pr P29274 -pr P29275 -r REG -sp scaffold -sf 0.15 -fe Morgan
+
+Another option is the cluster split, where the data is split into a test and train set randomly but keeping molecules with the same
+clusters in the same set. Here you can set the clustering method as well (``-scm``).
+
+..  code-block::
+
+    # Cluster split
+        python -m qsprpred.data_CLI -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/data -pr P29274 -pr P29275 -r REG -sp cluster -scm MaxMin -sf 0.15 -fe Morgan
 
 The third option is a temporal split, where a column needs to be indicated which holds information on the time each
 sample was observed and split based on threshold in a column. In this example, all samples after 2015 (in column 'year')
@@ -111,8 +144,16 @@ make up the test set. NOTE: this example will not work on the example set as it
 ..  code-block::
 
     # Time split
-        python -m qsprpred.data_CLI -i ./data/parkinsons_pivot.tsv -o ./cli_tutorial/data -pr GABAAalpha -pr NMDA -r REG  -sp time -st 2015 -stc year -fe Morgan
+        python -m qsprpred.data_CLI -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/data -pr P29274 -pr P29275 -r REG  -sp time -st 2015 -stc year -fe Morgan
 
+Lastly, the data can be split based on a specific column in the dataset. This column has to be named 'datasplit'
+where the value 'test' indicates the test set and the value 'train' indicates the train set.
+NOTE. this example will not work on the example set as it does not contain a 'datasplit' column.
+
+..  code-block::
+
+    # Split based on a specific column
+        python -m qsprpred.data_CLI -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/data -pr P29274 -pr P29275 -r REG -sp manual -sf 0.15 -fe Morgan
 
 Data for classification models
 """"""""""""""""""""""""""""""
@@ -127,22 +168,22 @@ the bins need to be given. For binary classification only give 1 threshold per p
 ..  code-block::
 
     # Classification and regression
-        python -m qsprpred.data_CLI -i ./data/parkinsons_pivot.tsv -o ./cli_tutorial/data -pr GABAAalpha -pr NMDA -r CLS -sp random -sf 0.15 -fe Morgan -th '{"GABAAalpha":[6.5],"NMDA":[0,4,6,10]}'
+        python -m qsprpred.data_CLI -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/data -pr P29274 -pr P29275 -r CLS -sp random -sf 0.15 -fe Morgan -th '{"P29274":[6.5],"P29275":[0,4,6,12]}'
 
-note. on windows remove the single quotes around the dictionary and add backslashes before the double quotes, e.g. -th {\"GABAAalpha\":[6.5],\"NMDA\":[0,4,6,10]}
+note. on windows remove the single quotes around the dictionary and add backslashes before the double quotes, e.g. -th {\"P29274\":[6.5],\"P29275\":[0,4,6,10]}
 
 Feature calculation
 """""""""""""""""""
-There are seven different descriptor sets that can be calculated at the moment,
-namely Morgan fingerprints, rdkit, Mordred, Mold2 and Padel descriptors, the
-physicochemical properties used in the QSAR models in the DrugEx papers from Liu et al., and
-the SMILES based signatures of extended valence. They can also
-be combined. For more control over the descriptorcalculator settings use the python API.
+There are many different descriptor sets that can be calculated from the CLI,
+such as Morgan fingerprints, rdkit, Mordred, Mold2 and Padel descriptors.
+Check the help message for the full list of available descriptor sets.
+The different descriptor sets can also be combined.
+For more control over the descriptor settings use the python API.
 
 ..  code-block::
 
     # With Morgan, RDkit, Mordred, Mold2, PaDEL and DrugEx descriptors
-        python -m qsprpred.data_CLI -i ./data/parkinsons_pivot.tsv -o ./cli_tutorial/data -pr GABAAalpha -pr NMDA -r REG -sp random -sf 0.15 -fe Morgan RDkit Mordred Mold2 PaDEL DrugEx
+        python -m qsprpred.data_CLI -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/data -pr P29274 -pr P29275 -r REG -sp random -sf 0.15 -fe Morgan RDkit Mordred Mold2 PaDEL DrugEx
 
 Feature filtering
 """""""""""""""""
@@ -155,27 +196,29 @@ and real features.
 ..  code-block::
 
     # input is in ./data/LIGAND_RAW_small.tsv
-       python -m qsprpred.data_CLI -i ./data/parkinsons_pivot.tsv -o ./cli_tutorial/data -pr GABAAalpha -pr NMDA -r REG -sp random -sf 0.15 -fe Morgan -lv 0.1 -hc 0.9 -bf 90
+       python -m qsprpred.data_CLI -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/data -pr P29274 -pr P29275 -r REG -sp random -sf 0.15 -fe Morgan -lv 0.1 -hc 0.9 -bf 90
 
 Papyrus Low quality filter
 """"""""""""""""""""""""""
 Specifically for use with a dataset from the `Papyrus dataset <https://chemrxiv.org/engage/chemrxiv/article-details/617aa2467a002162403d71f0>`,
 an option is included for filtering low quality data from the dataset (All data is removed with value 'Low' in column 'Quality').
-To apply this filter include `-lq` or `--low_quality` in your command.
+To apply this filter include ``-lq`` or ``--low_quality`` in your command.
 
 Multitask data
 """"""""""""""
-Multitask modelling is possible by passing multiple properties to the `-pr` argument. Furthermore, missing data can be
-imputed using the `-im` argument. You can combine any number of targets and combination of regression and classification
+Multitask modelling is possible by passing multiple properties to the ``-pr`` argument. Furthermore, missing data can be
+imputed using the ``-im`` argument. You can combine any number of targets and combination of regression and classification
 tasks for the data preparation, however currently the DNN models do not support multitask modelling and only the random
-forest models and knn sklearn models are supported for multitask. The multitask sklearn modelling is only possible for 
+forest models and KNN sklearn models are supported for multitask. The multitask sklearn modelling is only possible for 
 multiple regression task or multiple single class classification tasks. For multiple multi-class classification tasks or
 a combination of regression and classification tasks, the multitask modelling is not supported at the moment.
 
 ..  code-block::
 
     # input is in ./data/parkinsons_pivot.tsv
-        python -m qsprpred.data_CLI -i ./data/parkinsons_pivot.tsv -o ./cli_tutorial/data -pr GABAAalpha NMDA -r REG -sp random -sf 0.15 -fe Morgan -im mean
+        python -m qsprpred.data_CLI -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/data -pr P29274 P29275 -r REG -sp random -sf 0.15 -fe Morgan -im '{"P29274":"mean", "P29275":"median"}'
+
+Note. on windows remove the single quotes around the dictionary and add backslashes before the double quotes, e.g. -im {\"P29274\":\"mean\",\"P29275\":\"median\"}
 
 Model Training
 --------------
@@ -185,24 +228,47 @@ Basics
 
 Finally, we need to indicate what models we want to train and which steps to take in the training.
 In this example, we will build regression random forest models through passing the prepared regression datasets files
-`GABAAalpha_REGRESSION` and `NMDA_REGRESSION` to the `-dp` argument. If you wish to train classification models, you
-can pass the classification datasets `GABAAalpha_CLASSIFICATION` and `NMDA_CLASSIFICATION` to the `-dp` argument
-(or any combination thereof). The model type is set with `-m`. 
-We will also evaluate the model through cross-validation (-me) and train the model on all data to save for further use (-s).
+'P29274_REGRESSION' and 'P29275_REGRESSION' to the ``-dp`` argument. If you wish to train classification models, you
+can pass the classification datasets 'P29274_SINGLECLASS' and 'P29275_MULTICLASS' to the ``-dp`` argument
+(or any combination thereof). The model type is set with ``-m``. 
+We will also evaluate the model through cross-validation (``-me``) and train the model on all data to save for further use (``-s``).
 
 ..  code-block::
 
-    # Using the prepared datasets GABAAalpha_REGRESSION and NMDA_REGRESSION
-        python -m qsprpred.model_CLI -dp ./cli_tutorial/data/GABAAalpha_REGRESSION_df.pkl ./cli_tutorial/data/NMDA_REGRESSION_df.pkl -o ./cli_tutorial/models -mt RF -me -s
+    # Using the prepared datasets P29274_REGRESSION and P29275_REGRESSION
+        python -m qsprpred.model_CLI -dp ./tutorial_output/data/P29274_REGRESSION/P29274_REGRESSION_meta.json ./tutorial_output/data/P29275_REGRESSION/P29275_REGRESSION_meta.json -o ./tutorial_output/models -mt RF -me -s
+
+This will create a folder 'tutorial_output/models' containing the trained models.
+Each subfolder, named by the model type (RF) and the dataset name (P29274_REGRESSION/P29275_REGRESSION),
+will contain the following files:
+
++--------------------------------------------------+---------------------------------------------------------+
+| File                                             | Function                                                |
++==================================================+=========================================================+
+|| {prefixes}.json                                 || Model file                                             |
+|| {prefixes}_meta.json                            || Meta data, also used to instantiate a QSPRModel object |
+|| {prefixes}_cv.tsv                               || Cross-validation predictions                           |
+|| {prefixes}_ind.tsv                              || Test set predictions                                   |
++--------------------------------------------------+---------------------------------------------------------+
+
+Furthermore, the command line interface will create a log file and settings file in the output folder.
+
++--------------------------------------------------+-------------------------------------------------------+
+| File                                             | Function                                              |
++==================================================+=======================================================+
+|| QSPRmodel.json                                  || Command Line interface settings                      |
+|| QSPRmodel.log                                   || Log file                                             |
++--------------------------------------------------+-------------------------------------------------------+
 
 More
 ^^^^
 The model training can be further customized with several CLI arguments.
+For more control over the model training settings use the python API.
 Here you can find a short overview.
 
 run settings arguments
 ^^^^^^^^^^^^^^^^^^^^^^^
-As with the data preparation including `-de`, will print debug information to the log file. The random 
+As with the data preparation including ``-de``, will print debug information to the log file. The random 
 seed can also be set manually (although identical results are not guaranteed while keeping
 the same random seed). Furthermore, the number of cpu's used for model training and the
 gpu number for training pytorch models can be set.
@@ -210,7 +276,7 @@ gpu number for training pytorch models can be set.
 ..  code-block::
 
     # Setting debug flag, random seed, number of cpu's and a specific gpu (for now multiple gpu's not possible)
-        python -m qsprpred.model_CLI -de -ran 42 -ncpu 5 -gpus [3] -dp ./cli_tutorial/data/GABAAalpha_REGRESSION_df.pkl ./cli_tutorial/data/NMDA_REGRESSION_df.pkl -o ./cli_tutorial/models -mt RF -me -s
+        python -m qsprpred.model_CLI -de -ran 42 -ncpu 5 -gpus [3] -dp ./tutorial_output/data/P29274_REGRESSION/P29274_REGRESSION_meta.json ./tutorial_output/data/P29275_REGRESSION/P29275_REGRESSION_meta.json -o ./tutorial_output/models -mt RF -me -s
 
 model types
 """""""""""
@@ -222,7 +288,7 @@ The default is to run all the different model types.
 ..  code-block::
 
     # Training a RF, SVM and PLS model
-        python -m qsprpred.model_CLI -dp ./cli_tutorial/data/GABAAalpha_REGRESSION_df.pkl ./cli_tutorial/data/NMDA_REGRESSION_df.pkl -o ./cli_tutorial/models -me -s -mt RF SVM PLS
+        python -m qsprpred.model_CLI -dp ./tutorial_output/data/P29274_REGRESSION/P29274_REGRESSION_meta.json ./tutorial_output/data/P29275_REGRESSION/P29275_REGRESSION_meta.json -o ./tutorial_output/models -me -s -mt RF SVM PLS
 
 Defining model parameters
 """""""""""""""""""""""""
@@ -240,7 +306,7 @@ Specific model parameters can be set with the parameters argument by passing a j
 ..  code-block::
 
     # Setting some parameter values for a Random Forest and k-nearest neighbours model
-        python -m qsprpred.model_CLI -dp ./cli_tutorial/data/GABAAalpha_REGRESSION_df.pkl ./cli_tutorial/data/NMDA_REGRESSION_df.pkl -o ./cli_tutorial/models -mt RF KNN -me -s -p myparams
+        python -m qsprpred.model_CLI -dp ./tutorial_output/data/P29274_REGRESSION/P29274_REGRESSION_meta.json ./tutorial_output/data/P29275_REGRESSION/P29275_REGRESSION_meta.json -o ./tutorial_output/models -mt RF KNN -me -s -p ./tutorial_output/myparams
 
 Specifically for the training of the DNN model, you can set the tolerance and the patience from the CLI.
 Tolerance gives the mimimum decrease in loss needed to count as an improvement and 
@@ -249,7 +315,7 @@ patience is the number of training epochs without improvement in loss to stop th
 ..  code-block::
 
     # Setting the tolerance and patience for training a DNN model
-        python -m qsprpred.model_CLI -dp ./cli_tutorial/data/GABAAalpha_REGRESSION_df.pkl ./cli_tutorial/data/NMDA_REGRESSION_df.pkl -o ./cli_tutorial/models -mt DNN -me -s -tol 0.02 -pat 100
+        python -m qsprpred.model_CLI -dp ./tutorial_output/data/P29274_REGRESSION/P29274_REGRESSION_meta.json ./tutorial_output/data/P29275_REGRESSION/P29275_REGRESSION_meta.json -o ./tutorial_output/models -mt DNN -me -s -tol 0.02 -pat 100
 
 Hyperparameter optimization
 """""""""""""""""""""""""""
@@ -280,30 +346,29 @@ as third list item. The search space file should always include all models to be
 ..  code-block::
 
     # Bayesian optimization
-        python -m qsprpred.model_CLI -dp ./cli_tutorial/data/GABAAalpha_REGRESSION_df.pkl ./cli_tutorial/data/NMDA_REGRESSION_df.pkl -o ./cli_tutorial/models -mt RF -me -s -o bayes -nt 50 -ss mysearchspace.json -me -s
+        python -m qsprpred.model_CLI -dp ./tutorial_output/data/P29274_REGRESSION/P29274_REGRESSION_meta.json ./tutorial_output/data/P29275_REGRESSION/P29275_REGRESSION_meta.json -o ./tutorial_output/models -mt RF -me -s -o bayes -nt 5 -ss ./tutorial_output/mysearchspace.json -me -s
 
 Multitask modelling
 """""""""""""""""""
 Multitask modelling is also possible. This means that the models are trained on multiple targets at once.
-The modelling arguments are the same as for single task modelling, you just need to specifiy the a multitask
+The modelling arguments are the same as for single task modelling, you just need to specifiy a multitask
 dataset data prefix (see multitask data preparation).
 
-
 Prediction
 -----------
 Furthermore, trained QSPRpred models can be used to predict values from SMILES from the command line interface :code:`predict_CLI.py`.
 
 Basics
 ^^^^^^
-Here we will predict activity values for the A1 (GABAAalpha) and A3 receptor (NMDA) on the SMILES in the 
-dataset used in the previous examples using the models from the previous examples. The input `-i` here is the 
-set of SMILES for which we want to predict activity values. The argument `-mp`, is the paths to the meta files of the 
+Here we will predict activity values for the A1 (P29274) and A3 receptor (P29275) on the SMILES in the 
+dataset used in the previous examples using the models from the previous examples. The input ``-i`` here is the 
+set of SMILES for which we want to predict activity values. The argument ``-mp``, is the paths to the meta files of the 
 models we want to use for prediction relative to the base-directory subfolder qspr/models.
 
 ..  code-block::
     
-    # input is in ./data/parkinsons_pivot.tsv
-    python -m qsprpred.predict_CLI -i ./data/parkinsons_pivot.tsv -o ./cli_tutorial/predictions/parkinsons_preds.tsv -mp ./cli_tutorial/models/RF_GABAAalpha_REGRESSION/RF_GABAAalpha_REGRESSION_meta.json ./cli_tutorial/models/RF_NMDA_REGRESSION/RF_NMDA_REGRESSION_meta.json
+    # Making predictions for the A2A and A2B receptor
+    python -m qsprpred.predict_CLI -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/predictions/AR_LIGANDS_preds.tsv -mp ./tutorial_output/models/RF_P29274_REGRESSION/RF_P29274_REGRESSION_meta.json ./tutorial_output/models/RF_P29275_REGRESSION/RF_P29275_REGRESSION_meta.json
 
 More
 ^^^^
@@ -312,23 +377,22 @@ Here you can find a short overview.
 
 run settings arguments
 ^^^^^^^^^^^^^^^^^^^^^^
-As with the data preparation including `-d`, will print debug information to the log file. The random 
-seed can also be set manually (although identical results are not guaranteed while keeping
-the same random seed). The output file name can be set. Furthermore, the number of cpu's used for model prediction and the
+As with the data preparation including ``-de``, will print debug information to the log file. The random 
+seed can also be set manually. Furthermore, the number of cpu's used for model prediction and the
 gpu number for prediction with pytorch models can be set.
 
 ..  code-block::
 
     # Setting debug flag, random seed, output file name, number of cpu's and a specific gpu (for now multiple gpu's not possible)
-        python -m qsprpred.predict_CLI -i ./data/parkinsons_pivot.tsv -o ./cli_tutorial/predictions/parkinsons_preds.tsv -mp ./cli_tutorial/models/RF_GABAAalpha_REGRESSION/RF_GABAAalpha_REGRESSION_meta.json ./cli_tutorial/models/RF_NMDA_REGRESSION/RF_NMDA_REGRESSION_meta.json -o mypredictions -de -ran 42 -ncpu 5 -gpus [3]
+        python -m qsprpred.predict_CLI -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/predictions/AR_LIGANDS_preds.tsv -mp ./tutorial_output/models/RF_P29274_REGRESSION/RF_P29274_REGRESSION_meta.json ./tutorial_output/models/RF_P29275_REGRESSION/RF_P29275_REGRESSION_meta.json -de -ran 42 -ncpu 5 -gpus [3]
 
     
 Adding probability predictions
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
-When using a classification model, the probability of the predicted class can be added to the output file using the `-pr` flag.
+When using a classification model, the probability of the predicted class can be added to the output file using the ``-pr`` flag.
 
 ..  code-block::
     
-    # Do not standardize and sanitize SMILES
-    python -m qsprpred.predict_CLI -i ./data/parkinsons_pivot.tsv -o ./cli_tutorial/predictions/parkinsons_preds.tsv -mp ./cli_tutorial/models/RF_GABAAalpha_SINGLECLASS/RF_GABAAalpha_SINGLECLASS_meta.json ./cli_tutorial/models/RF_NMDA_MULTICLASS/RF_NMDA_MULTICLASS_meta.json -pr
+    # Adding probability predictions
+    python -m qsprpred.predict_CLI -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/predictions/AR_LIGANDS_preds.tsv -mp ./tutorial_output/models/RF_P29274_SINGLECLASS/RF_P29274_SINGLECLASS_meta.json ./tutorial_output/models/RF_P29275_MULTICLASS/RF_P29275_MULTICLASS_meta.json -pr
 
diff --git a/docs/features.rst b/docs/features.rst
index cebe3cde..f5625b2c 100644
--- a/docs/features.rst
+++ b/docs/features.rst
@@ -15,14 +15,14 @@ Overview of available features
 
             .. tab-item:: Core
 
-                * :class:`~qsprpred.data.sources.papyrus.papyrus_class.Papyrus`: Papyrus (See `data collection with Papyrus tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/data/data_collection_with_papyrus.ipynb>`_.)
+                * :class:`~qsprpred.data.sources.papyrus.papyrus_class.Papyrus`: Papyrus (See `data collection with Papyrus tutorial <https://github.com/CDDLeiden/QSPRpred/blob/main/tutorials/basics/data/data_collection_with_papyrus.ipynb>`_.)
 
     .. dropdown:: Data Filters
 
         :class:`~qsprpred.data.processing.data_filters.DataFilter`: Base class for data filters.
 
         Data filters are used to filter data based on some criteria.
-        Examples can be found in the `data preparation tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/data/data_preparation.ipynb>`_.
+        Examples can be found in the `data preparation tutorial <https://github.com/CDDLeiden/QSPRpred/blob/main/tutorials/basics/data/data_preparation.ipynb>`_.
 
         .. tab-set::
 
@@ -36,7 +36,7 @@ Overview of available features
         :class:`~qsprpred.data.descriptors.sets.DescriptorSet`: Base class for descriptor sets.
 
         Descriptor sets are used to calculate molecular descriptors for a set of molecules.
-        Examples can be found in the `descriptor calculation tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/data/descriptors.ipynb>`_.
+        Examples can be found in the `descriptor calculation tutorial <https://github.com/CDDLeiden/QSPRpred/blob/main/tutorials/basics/data/descriptors.ipynb>`_.
 
         .. tab-set::
 
@@ -83,7 +83,7 @@ Overview of available features
         :class:`~qsprpred.data.sampling.splits.DataSplit`: Base class for data splitters.
 
         Data splitters are used to split data into training and test sets.
-        Examples can be found in the `data splitting tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/data/data_splitting.ipynb>`_.
+        Examples can be found in the `data splitting tutorial <https://github.com/CDDLeiden/QSPRpred/blob/main/tutorials/basics/data/data_splitting.ipynb>`_.
 
         .. tab-set::
 
@@ -110,7 +110,7 @@ Overview of available features
         :class:`~qsprpred.data.processing.feature_filters.FeatureFilter`: Base class for feature filters.
 
         Feature filters are used to filter features based on some criteria.
-        Examples can be found in the `data preparation tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/data/data_preparation.ipynb>`_.
+        Examples can be found in the `data preparation tutorial <https://github.com/CDDLeiden/QSPRpred/blob/main/tutorials/basics/data/data_preparation.ipynb>`_.
 
         .. tab-set::
 
@@ -125,7 +125,7 @@ Overview of available features
         :class:`~qsprpred.models.models.QSPRModel`: Base class for models.
 
         Models are used to predict properties of molecules.
-        A general example can be found in the `quick start tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/quick_start.ipynb>`_.
+        A general example can be found in the `quick start tutorial <https://github.com/CDDLeiden/QSPRpred/blob/main/tutorials/quick_start.ipynb>`_.
         More detailed information can be found throughout the basic and advanced modelling tutorials.
 
         .. tab-set::
@@ -136,14 +136,14 @@ Overview of available features
 
             .. tab-item:: Extra
 
-                * :class:`~qsprpred.extra.models.pcm.PCMModel`: PCMModel (See `PCM tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/advanced/modelling/PCM_modelling.ipynb>`_.)
+                * :class:`~qsprpred.extra.models.pcm.PCMModel`: PCMModel (See `PCM tutorial <https://github.com/CDDLeiden/QSPRpred/blob/main/tutorials/advanced/modelling/PCM_modelling.ipynb>`_.)
 
             .. tab-item:: GPU
                 
-                More information can be found in the `deep learning tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/advanced/modelling/deep_learning_models.ipynb>`_.
+                More information can be found in the `deep learning tutorial <https://github.com/CDDLeiden/QSPRpred/blob/main/tutorials/advanced/modelling/deep_learning_models.ipynb>`_.
                 
                 * :class:`~qsprpred.extra.gpu.models.dnn.DNNModel`: DNNModel
-                * :class:`~qsprpred.extra.gpu.models.chemprop.ChempropModel`: ChempropModel (See `Chemprop tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/advanced/modelling/chemprop_models.ipynb>`_.)
+                * :class:`~qsprpred.extra.gpu.models.chemprop.ChempropModel`: ChempropModel (See `Chemprop tutorial <https://github.com/CDDLeiden/QSPRpred/blob/main/tutorials/advanced/modelling/chemprop_models.ipynb>`_.)
                 * :class:`~qsprpred.extra.gpu.models.pyboost.PyBoostModel`: PyBoostModel
 
     .. dropdown:: Metrics
@@ -151,7 +151,7 @@ Overview of available features
         :class:`~qsprpred.models.metrics.Metric`: Base class for metrics
 
         Metrics are used to evaluate the performance of models.
-        More information can be found in the `model assessment tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/modelling/model_assessment.ipynb>`_.
+        More information can be found in the `model assessment tutorial <https://github.com/CDDLeiden/QSPRpred/blob/main/tutorials/basics/modelling/model_assessment.ipynb>`_.
 
         .. tab-set::
 
@@ -164,7 +164,7 @@ Overview of available features
         :class:`~qsprpred.models.assessment_methods.ModelAssessor`: Base class for model assessors.
 
         Model assessors are used to assess the performance of models.
-        More information be found in the `model assessment tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/basics/modelling/model_assessment.ipynb>`_.
+        More information be found in the `model assessment tutorial <https://github.com/CDDLeiden/QSPRpred/blob/main/tutorials/basics/modelling/model_assessment.ipynb>`_.
 
         .. tab-set::
 
@@ -178,7 +178,7 @@ Overview of available features
         :class:`~qsprpred.models.hyperparam_optimization.HyperparameterOptimization`: Base class for hyperparameter optimizers.
 
         Hyperparameter optimizers are used to optimize the hyperparameters of models.
-        More information can be found in the `hyperparameter optimization tutorial <https://github.com/CDDLeiden/QSPRpred/blob/dev/tutorials/advanced/modelling/hyperparameter_optimization.ipynb>`_.
+        More information can be found in the `hyperparameter optimization tutorial <https://github.com/CDDLeiden/QSPRpred/blob/main/tutorials/advanced/modelling/hyperparameter_optimization.ipynb>`_.
 
         .. tab-set::
 
diff --git a/qsprpred/data_CLI.py b/qsprpred/data_CLI.py
index d221c083..b9ed2a86 100644
--- a/qsprpred/data_CLI.py
+++ b/qsprpred/data_CLI.py
@@ -24,6 +24,10 @@
     RDKitFP,
     AvalonFP,
 )
+from qsprpred.data.chem.clustering import (
+    FPSimilarityMaxMinClusters,
+    FPSimilarityLeaderPickerClusters
+)
 from qsprpred.data.descriptors.sets import (
     DrugExPhyschem,
     PredictorDesc,
@@ -111,7 +115,16 @@ def QSPRArgParser(txt=None):
         ),
     )
     parser.add_argument(
-        "-im", "--imputation", type=str, choices=["mean", "median", "most_frequent"]
+        "-im",
+        "--imputation",
+        type=json.loads,
+        help=(
+            "Imputation method for missing values. Specify the imputation method as a "
+            "dictionary with the property name as key and the imputation method as "
+            "value, e.g. -im \"{'CL':'mean','fu':'median'}\". Note: no spaces and "
+            "surrounded by single quotes. Choose from 'mean', 'median', 'most_frequent'"
+        ),
+        default={},
     )
     # model type arguments
     parser.add_argument(
@@ -222,7 +235,8 @@ def QSPRArgParser(txt=None):
             "Mold2",
             "PaDEL",
             "DrugEx",
-            "Signature" "MaccsFP",
+            "Signature",
+            "MaccsFP",
             "AvalonFP",
             "TopologicalFP",
             "AtomPairFP",
@@ -347,18 +361,11 @@ def QSPR_dataprep(args):
                         "transformer": transform_dict[args.transform_data[prop]]
                         if prop in args.transform_data
                         else None,
+                        "imputer": SimpleImputer(strategy=args.imputation[prop])
+                        if prop in args.imputation
+                        else None
                     }
                 )
-            # missing value imputation
-            if args.imputation is not None:
-                if args.imputation == "mean":
-                    imputer = SimpleImputer(strategy="mean")
-                elif args.imputation == "median":
-                    imputer = SimpleImputer(strategy="median")
-                elif args.imputation == "most_frequent":
-                    imputer = SimpleImputer(strategy="most_frequent")
-                else:
-                    sys.exit("invalid impute arg given")
             dataset_name = (
                 f"{props_name}_{task}_{args.data_suffix}"
                 if args.data_suffix
@@ -404,9 +411,13 @@ def QSPR_dataprep(args):
                     splitcol=df["datasplit"], trainval="train", testval="test"
                 )
             elif args.split == "cluster":
+                if args.split_cluster_method == "MaxMin":
+                    clustering = FPSimilarityMaxMinClusters()
+                elif args.split_cluster_method == "LeaderPicker":
+                    clustering = FPSimilarityLeaderPickerClusters()
                 split = ClusterSplit(
                     test_fraction=args.split_fraction,
-                    clustering_algorithm=args.split_clustering_method,
+                    clustering=clustering,
                     dataset=mydataset,
                 )
             else:
diff --git a/tutorials/tutorial_data/create_tutorial_data.py b/tutorials/tutorial_data/create_tutorial_data.py
index 69371408..fd602e63 100644
--- a/tutorials/tutorial_data/create_tutorial_data.py
+++ b/tutorials/tutorial_data/create_tutorial_data.py
@@ -1,6 +1,7 @@
 import os
 
 from qsprpred.data.sources.papyrus import Papyrus
+import argparse
 
 
 def SingleTaskTutorialData(data_dir: str | None = None):
@@ -88,7 +89,27 @@ def MultiTaskTutorialData(data_dir: str | None = None):
 
     return dataset
 
+def prepare_multitask(file_path: str):
+    import pandas as pd
+    df = pd.read_csv(file_path, sep="\t")
+    df = df.pivot(index="SMILES", columns="accession", values="pchembl_value_Mean")
+    df.columns.name = None
+    df.reset_index(inplace=True)
+    df.to_csv(f"{file_path.split('.')[0]}_pivot.tsv", sep="\t", index=False)
+
 
 if __name__ == "__main__":
-    SingleTaskTutorialData()
-    MultiTaskTutorialData()
+    parser = argparse.ArgumentParser(description="Create tutorial data")
+    parser.add_argument(
+        "-p",
+        "--pivot",
+        type=str,
+        help="Pivot AR_ligands dataset, provide the file path")
+
+    args = parser.parse_args()
+
+    if args.pivot is not None:
+        prepare_multitask(args.pivot)
+    else:
+        SingleTaskTutorialData()
+        MultiTaskTutorialData()

From 79b5d973f26392691b0ec55753a4a284f45d3393 Mon Sep 17 00:00:00 2001
From: maagdenbehwvanden1 <h.w.van.den.maagdenberg@lacdr.leidenuniv.nl>
Date: Tue, 6 Feb 2024 10:38:18 +0100
Subject: [PATCH 17/18] some syntax fixes to docs

---
 docs/cli_usage.rst            | 70 ++++++++++++++++++++++-------------
 docs/index.rst                |  2 +-
 docs/install.rst              |  4 +-
 docs/requirements.txt         |  4 +-
 qsprpred/data/tables/tests.py |  4 +-
 tutorials/README.md           |  3 +-
 6 files changed, 55 insertions(+), 32 deletions(-)

diff --git a/docs/cli_usage.rst b/docs/cli_usage.rst
index f8039eae..f41c9ebf 100644
--- a/docs/cli_usage.rst
+++ b/docs/cli_usage.rst
@@ -11,10 +11,10 @@ e.g. the help message for the :code:`QSPRpred.data_CLI` script can be shown as f
 
     python -m qsprpred.data_CLI --help
 
-A simple command-line workflow to prepare your dataset and train QSPR models is given below (see :ref:`cli-example`).
+A simple command-line workflow to prepare your dataset and train QSPR models is given below (see :ref:`CLI Example`).
 
 If you want more control over the inputs and outputs or want to customize QSPRpred for your purpose,
-you can also use the Python API directly (see `source code <https://github.com/CDDLeiden/QSPRpred/tree/main/tutorials>`).
+you can also use the Python API directly (see `source code <https://github.com/CDDLeiden/QSPRpred/tree/main/tutorials>`_).
 Here you can find a tutorial with a Jupyter notebook illustrating some common use cases in the project source code.
 Make sure to download the tutorial folder to follow the examples in this CLI tutorial.
 
@@ -23,12 +23,13 @@ CLI Example
 ***********
 
 In this example, we will use the command line utilities of QSPRpred to train a QSAR model for the A2A and A2B receptor.
-We will use the Adenosine dataset from the API tutorial. The data is available through [OneDrive](https://1drv.ms/u/s!AtzWqu0inkjX3QRxXOkTFNv7IV7u?e=PPj0O2)
-(just unzip and place the two datasets `A2A_LIGANDS.tsv` and `AR_LIGANDS.tsv` in the `tutorial_data` folder) or recreate the dataset yourself by running `tutorial_data/create_tutorial_data.py`.
+We will use the Adenosine dataset from the API tutorial. The data is available through `OneDrive <https://1drv.ms/u/s!AtzWqu0inkjX3QRxXOkTFNv7IV7u?e=PPj0O2>`_
+(just unzip and place the two datasets **A2A_LIGANDS.tsv** and **AR_LIGANDS.tsv** in the 'tutorial_data' folder) or recreate the dataset yourself by running 'tutorial_data/create_tutorial_data.py'.
+
 Input data should contain a column with SMILES sequences and at least one column with a property for modelling.
-The `AR_LIGANDS.tsv` file contains a SMILES column, a column with the property 'pchembl_value_mean' and a column with the property 'Accession` (uniprot accession numbers).
+The **AR_LIGANDS.tsv** file contains a SMILES column, a column with the property **pchembl_value_mean** and a column with the property **accession** (uniprot accession numbers).
 However, to create models for the A2A and A2B receptor, we need to have the data in a pivot table format, where the properties are in the columns.
-To create this pivot table, we can use the `pandas` library in Python as follows:
+To create this pivot table, we can use the pandas library in Python as follows:
 
 ..  code-block::
 
@@ -39,19 +40,21 @@ To create this pivot table, we can use the `pandas` library in Python as follows
     df.reset_index(inplace=True)
     df.to_csv('tutorial_data/AR_LIGANDS_pivot.tsv', sep='\t')
 
-It is also possible to simply run the `create_tutorial_data.py` script in the `tutorial_data` folder with the following command:
+It is also possible to simply run the **create_tutorial_data.py** script in the **tutorial_data** folder with the following command:
 
 ..  code-block::
 
     python create_tutorial_data.py -m AR_LIGANDS.tsv
 
-This will create a pivot table with the name `AR_LIGANDS_pivot.tsv` in the `tutorial_data` folder.
+This will create a pivot table with the name **AR_LIGANDS_pivot.tsv** in the **tutorial_data** folder.
 Our example dataset now contains a SMILES column and two columns with the properties **P29274** (A2AR) and **P29275** (A2BR) 
 (as well as a columns for A1 and A3, which we will not use in this example).
 
 Preparing Data
 --------------
 
+.. _Data Basics:
+
 Basics
 ^^^^^^
 We will now use the QSPRpred :code:`QSPRpred.data_CLI` script for data preparation.
@@ -68,7 +71,7 @@ here we use Morgan fingerprints.
     # input is in tutorial/data/AR_LIGANDS_pivot.tsv
     python -m qsprpred.data_CLI -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/data -pr P29274 -pr P29275 -r REG -sp random -sf 0.15 -fe Morgan
 
-Running this command will create folders 'tutorial_output/data', with subfolders 'P29274_REGRESSION' and 'P29275_REGRESSION' containing the prepared data.
+Running this command will create folders **tutorial_output/data**, with subfolders **P29274_REGRESSION** and **P29275_REGRESSION** containing the prepared data.
 Each subfolder, named by the property identifiers (i.e. P29274/P29275) and the model task (REGRESSION),
 will contain the following files:
 
@@ -91,10 +94,15 @@ Furthermore, the command line interface will create a log file and settings file
 || QSPRdata.log                                    || Log file                                             |
 +--------------------------------------------------+-------------------------------------------------------+
 
+.. _Data More:
+
 More
 ^^^^
+
+.. _Data Run settings arguments:
+
 Run settings arguments
-^^^^^^^^^^^^^^^^^^^^^^^
+""""""""""""""""""""""
 Apart from the the input file name, there are a few other base options that can be set.
 Including ``-d``, will print debug information to the log file. The random
 seed (``-ran``) can also be set manually, which should guarantee identical results while keeping the same random seed.
@@ -138,17 +146,17 @@ clusters in the same set. Here you can set the clustering method as well (``-scm
         python -m qsprpred.data_CLI -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/data -pr P29274 -pr P29275 -r REG -sp cluster -scm MaxMin -sf 0.15 -fe Morgan
 
 The third option is a temporal split, where a column needs to be indicated which holds information on the time each
-sample was observed and split based on threshold in a column. In this example, all samples after 2015 (in column 'year')
-make up the test set. NOTE: this example will not work on the example set as it does not contain a 'year' column.
+sample was observed and split based on threshold in a column. In this example, all samples after 2015 (in column **year**)
+make up the test set. NOTE: this example will not work on the example set as it does not contain a **year** column.
 
 ..  code-block::
 
     # Time split
         python -m qsprpred.data_CLI -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/data -pr P29274 -pr P29275 -r REG  -sp time -st 2015 -stc year -fe Morgan
 
-Lastly, the data can be split based on a specific column in the dataset. This column has to be named 'datasplit'
-where the value 'test' indicates the test set and the value 'train' indicates the train set.
-NOTE. this example will not work on the example set as it does not contain a 'datasplit' column.
+Lastly, the data can be split based on a specific column in the dataset. This column has to be named **datasplit**
+where the value **test** indicates the test set and the value **train** indicates the train set.
+NOTE. this example will not work on the example set as it does not contain a **datasplit** column.
 
 ..  code-block::
 
@@ -170,7 +178,7 @@ the bins need to be given. For binary classification only give 1 threshold per p
     # Classification and regression
         python -m qsprpred.data_CLI -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/data -pr P29274 -pr P29275 -r CLS -sp random -sf 0.15 -fe Morgan -th '{"P29274":[6.5],"P29275":[0,4,6,12]}'
 
-note. on windows remove the single quotes around the dictionary and add backslashes before the double quotes, e.g. -th {\"P29274\":[6.5],\"P29275\":[0,4,6,10]}
+note. on windows remove the single quotes around the dictionary and add backslashes before the double quotes, e.g. ``-th {\"P29274\":[6.5],\"P29275\":[0,4,6,10]}``.
 
 Feature calculation
 """""""""""""""""""
@@ -200,7 +208,7 @@ and real features.
 
 Papyrus Low quality filter
 """"""""""""""""""""""""""
-Specifically for use with a dataset from the `Papyrus dataset <https://chemrxiv.org/engage/chemrxiv/article-details/617aa2467a002162403d71f0>`,
+Specifically for use with a dataset from the `Papyrus dataset <https://chemrxiv.org/engage/chemrxiv/article-details/617aa2467a002162403d71f0>`_,
 an option is included for filtering low quality data from the dataset (All data is removed with value 'Low' in column 'Quality').
 To apply this filter include ``-lq`` or ``--low_quality`` in your command.
 
@@ -223,13 +231,15 @@ Note. on windows remove the single quotes around the dictionary and add backslas
 Model Training
 --------------
 
+.. _Model Basics:
+
 Basics
 ^^^^^^
 
 Finally, we need to indicate what models we want to train and which steps to take in the training.
 In this example, we will build regression random forest models through passing the prepared regression datasets files
-'P29274_REGRESSION' and 'P29275_REGRESSION' to the ``-dp`` argument. If you wish to train classification models, you
-can pass the classification datasets 'P29274_SINGLECLASS' and 'P29275_MULTICLASS' to the ``-dp`` argument
+**P29274_REGRESSION** and **P29275_REGRESSION** to the ``-dp`` argument. If you wish to train classification models, you
+can pass the classification datasets **P29274_SINGLECLASS** and **P29275_MULTICLASS** to the ``-dp`` argument
 (or any combination thereof). The model type is set with ``-m``. 
 We will also evaluate the model through cross-validation (``-me``) and train the model on all data to save for further use (``-s``).
 
@@ -238,7 +248,7 @@ We will also evaluate the model through cross-validation (``-me``) and train the
     # Using the prepared datasets P29274_REGRESSION and P29275_REGRESSION
         python -m qsprpred.model_CLI -dp ./tutorial_output/data/P29274_REGRESSION/P29274_REGRESSION_meta.json ./tutorial_output/data/P29275_REGRESSION/P29275_REGRESSION_meta.json -o ./tutorial_output/models -mt RF -me -s
 
-This will create a folder 'tutorial_output/models' containing the trained models.
+This will create a folder **tutorial_output/models** containing the trained models.
 Each subfolder, named by the model type (RF) and the dataset name (P29274_REGRESSION/P29275_REGRESSION),
 will contain the following files:
 
@@ -260,14 +270,18 @@ Furthermore, the command line interface will create a log file and settings file
 || QSPRmodel.log                                   || Log file                                             |
 +--------------------------------------------------+-------------------------------------------------------+
 
+.. _Model More:
+
 More
 ^^^^
 The model training can be further customized with several CLI arguments.
 For more control over the model training settings use the python API.
 Here you can find a short overview.
 
-run settings arguments
-^^^^^^^^^^^^^^^^^^^^^^^
+.. _Model Run settings arguments:
+
+Run settings arguments
+""""""""""""""""""""""
 As with the data preparation including ``-de``, will print debug information to the log file. The random 
 seed can also be set manually (although identical results are not guaranteed while keeping
 the same random seed). Furthermore, the number of cpu's used for model training and the
@@ -358,6 +372,8 @@ Prediction
 -----------
 Furthermore, trained QSPRpred models can be used to predict values from SMILES from the command line interface :code:`predict_CLI.py`.
 
+.. _Prediction Basics:
+
 Basics
 ^^^^^^
 Here we will predict activity values for the A1 (P29274) and A3 receptor (P29275) on the SMILES in the 
@@ -370,13 +386,17 @@ models we want to use for prediction relative to the base-directory subfolder qs
     # Making predictions for the A2A and A2B receptor
     python -m qsprpred.predict_CLI -i ./tutorial_data/AR_LIGANDS_pivot.tsv -o ./tutorial_output/predictions/AR_LIGANDS_preds.tsv -mp ./tutorial_output/models/RF_P29274_REGRESSION/RF_P29274_REGRESSION_meta.json ./tutorial_output/models/RF_P29275_REGRESSION/RF_P29275_REGRESSION_meta.json
 
+.. _Prediction More:
+
 More
 ^^^^
 The predictions can be further customized with several CLI arguments.
 Here you can find a short overview.
 
-run settings arguments
-^^^^^^^^^^^^^^^^^^^^^^
+.. _Prediction Run settings arguments:
+
+Run settings arguments
+""""""""""""""""""""""
 As with the data preparation including ``-de``, will print debug information to the log file. The random 
 seed can also be set manually. Furthermore, the number of cpu's used for model prediction and the
 gpu number for prediction with pytorch models can be set.
@@ -388,7 +408,7 @@ gpu number for prediction with pytorch models can be set.
 
     
 Adding probability predictions
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+""""""""""""""""""""""""""""""
 When using a classification model, the probability of the predicted class can be added to the output file using the ``-pr`` flag.
 
 ..  code-block::
diff --git a/docs/index.rst b/docs/index.rst
index 933b1f02..c97e9d1b 100644
--- a/docs/index.rst
+++ b/docs/index.rst
@@ -8,7 +8,7 @@ Welcome to QSPRpred's documentation!
 
 QSPRpred is open-source software libary for building **Quantitative Structure Property Relationship (QSPR)** models developed by Gerard van Westen's Computational Drug
 Discovery group. It provides a unified interface for building QSPR models based on different types of descriptors and machine learning algorithms.
-Here you will find the installation guide (:ref:`installation`), an overview of available features (:ref:`features`), usage examples (:ref:`cli_usage`) and API documentation (:ref:`Python API`).
+Here you will find the installation guide (:ref:`installation-guide`), an overview of available features (:ref:`features`), usage examples (:ref:`cli-usage`) and API documentation (:ref:`Python API`).
 For tutorials and examples of the Python API, please visit the `QSPRpred GitHub repository <https://github.com/CDDLeiden/QSPRpred>`_.
 
 .. toctree::
diff --git a/docs/install.rst b/docs/install.rst
index 9b4d107c..86415813 100644
--- a/docs/install.rst
+++ b/docs/install.rst
@@ -1,4 +1,4 @@
-..  _installation:
+..  _installation-guide:
 
 Installation
 ============
@@ -9,7 +9,7 @@ You do not need anything special to install the package . Just run the following
 
     pip install git+https://github.com/CDDLeiden/QSPRpred.git@main
 
-You can also get tags and development snapshots by varying the :code:`@main` part (i.e. :code:`@1.0.0`). After that you can start building models (see :ref:`usage`).
+You can also get tags and development snapshots by varying the :code:`@main` part (i.e. :code:`@1.0.0`). After that you can start building models (see :ref:`cli-usage`).
 
 Note that this will install the basic dependencies, but not the optional dependencies.
 If you want to use the optional dependencies, you can install the package with an
diff --git a/docs/requirements.txt b/docs/requirements.txt
index 856b3f4c..5642b7c2 100644
--- a/docs/requirements.txt
+++ b/docs/requirements.txt
@@ -1,2 +1,4 @@
 sphinx
-sphinx_rtd_theme
\ No newline at end of file
+sphinx_rtd_theme
+sphinx_design
+sphinx_design_elements
\ No newline at end of file
diff --git a/qsprpred/data/tables/tests.py b/qsprpred/data/tables/tests.py
index e0ca6138..184e6637 100644
--- a/qsprpred/data/tables/tests.py
+++ b/qsprpred/data/tables/tests.py
@@ -595,7 +595,7 @@ def testSerialization(self, task, name, th, transformer):
 
 class TestDataSetPreparation(DataSetsPathMixIn, DataPrepCheckMixIn, QSPRTestCase):
     """Test as many possible combinations of data sets and their preparation
-    settings. These can run potentially for a long time so use the `skip` decorator
+    settings. These can run potentially for a long time so use the ``skip`` decorator
     if you want to skip all these tests to speed things up during development."""
     def setUp(self):
         super().setUp()
@@ -615,7 +615,7 @@ def testPrepCombos(
     ):
         """Tests one combination of a data set and its preparation settings.
 
-        This generates a large number of parameterized tests. Use the `skip` decorator
+        This generates a large number of parameterized tests. Use the ``skip`` decorator
         if you want to skip all these tests. Note that the combinations are not
         exhaustive, but defined by `DataSetsPathMixIn.getPrepCombos()`."""
         np.random.seed(42)
diff --git a/tutorials/README.md b/tutorials/README.md
index 1c4c81bc..806901e3 100644
--- a/tutorials/README.md
+++ b/tutorials/README.md
@@ -14,7 +14,8 @@ The rest of the tutorials is divided into two categories: basic and advanced.
 The Basics tutorials cover all basic functionality of QSPRpred that you will need in
 most projects. The Advanced tutorials cover more advanced topics and are designed for
 users who are already familiar with the basics of QSPRpred.
-For detailed description of all QSPRpred classes and functions, see
+For detailed description of all QSPRpred classes and functions, as well as examples of
+how to use the command line interface, see
 the [documentation](https://cddleiden.github.io/QSPRpred/docs/).
 
 - **[Quick Start](quick_start.ipynb)**: A quick start guide to using QSPRpred.

From a0d06cab613fd0196002574fd67e52f5da4382dd Mon Sep 17 00:00:00 2001
From: maagdenbehwvanden1 <h.w.van.den.maagdenberg@lacdr.leidenuniv.nl>
Date: Tue, 6 Feb 2024 11:04:52 +0100
Subject: [PATCH 18/18] quick solution for backup data not working

---
 qsprpred/data_CLI.py | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

diff --git a/qsprpred/data_CLI.py b/qsprpred/data_CLI.py
index b9ed2a86..9de10a89 100644
--- a/qsprpred/data_CLI.py
+++ b/qsprpred/data_CLI.py
@@ -524,15 +524,17 @@ def QSPR_dataprep(args):
     if not os.path.exists(args.output_dir):
         os.makedirs(args.output_dir)
 
-    # Backup files
-    tasks = ["REG" if reg is True else "CLS" for reg in args.regression]
-    file_prefixes = [
-        f"{property}_{task}" for task in tasks for property in args.properties
-    ]
+    # get a list of all the folders in the output directory
+    folders = [f for f in os.listdir(args.output_dir) if os.path.isdir(f"{args.output_dir}/{f}")]
+
+    # remove folders that start with backup
+    folders = [f for f in folders if not f.startswith("backup")]
+
+
     if not args.skip_backup:
         backup_msg = backup_files(
             args.output_dir,
-            tuple(file_prefixes),
+            tuple(folders),
             cp_suffix=["calculators", "standardizer", "meta"],
         )