[nit.git] / src / metrics / codesmells_metrics.nit

# This file is part of NIT ( http://www.nitlanguage.org ).
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# Detect the code smells and antipatterns in the code.

module codesmells_metrics

import frontend
import nitsmell_toolcontext
import method_analyze_metrics
import mclassdef_collect

redef class ToolContext
        var codesmells_metrics_phase = new CodeSmellsMetricsPhase(self, null)
end

class CodeSmellsMetricsPhase
        super Phase
        var average_number_of_lines = 0.0
        var average_number_of_parameter = 0.0
        var average_number_of_method = 0.0
        var average_number_of_attribute = 0.0

        redef fun process_mainmodule(mainmodule, given_mmodules) do
                print toolcontext.format_h1("--- Code Smells Metrics ---")

                var view = new ModelView(toolcontext.modelbuilder.model, mainmodule)
                self.set_all_average_metrics(view)
                var mclass_codesmell = new BadConceptonController(view)
                var collect = new Counter[MClassDef]
                var mclassdefs = new Array[MClassDef]

                for mclass in mainmodule.flatten_mclass_hierarchy do
                        mclass_codesmell.collect(mclass.mclassdefs,self)
                end
                if toolcontext.opt_get_all.value then
                        mclass_codesmell.print_all
                else
                        mclass_codesmell.print_top(10)
                end
        end

        fun set_all_average_metrics(view: ModelView) do
                var model_builder = toolcontext.modelbuilder
                self.average_number_of_lines = view.get_avg_linenumber(model_builder)
                self.average_number_of_parameter = view.get_avg_parameter
                self.average_number_of_method = view.get_avg_method
                self.average_number_of_attribute = view.get_avg_attribut
        end
end

class BadConceptonController

        var view: ModelView

        # Code smell list
        var bad_conception_elements = new Array[BadConceptionFinder]

        # Print all collected code smell sort in decroissant order
        fun print_all do
                for bad_conception in self.sort do
                        bad_conception.print_collected_data
                end
        end

        # Print the n top element
        fun print_top(number: Int) do
                for bad_conception in self.get_numbers_of_elements(number) do
                        bad_conception.print_collected_data
                end
        end

        # Collect method take Array of mclassdef to find the code smells for every class
        fun collect(mclassdefs: Array[MClassDef],phase: CodeSmellsMetricsPhase) do
                for mclassdef in mclassdefs do
                        var bad_conception_class = new BadConceptionFinder(mclassdef, phase, view)
                        bad_conception_class.collect
                        bad_conception_elements.add(bad_conception_class)
                end
        end

        # Sort the bad_conception_elements array
        fun sort: Array[BadConceptionFinder]
        do
                var res = bad_conception_elements
                var sorter = new BadConceptionComparator
                sorter.sort(res)
                return res
        end

        # Return an array with n elements
        fun get_numbers_of_elements(number : Int) : Array[BadConceptionFinder]do
                var return_values = new Array[BadConceptionFinder]
                var list = self.sort
                var min = number
                if list.length <= number*2 then min = list.length
                for i in [0..min[ do
                        var t = list[list.length-i-1]
                        return_values.add(t)
                end
                return return_values
        end
end

class BadConceptionFinder
        var mclassdef: MClassDef
        var array_badconception = new Array[BadConception]
        var phase: CodeSmellsMetricsPhase
        var view: ModelView
        var score = 0.0

        # Collect code smell with selected toolcontext option
        fun collect do
                var bad_conception_elements = new Array[BadConception]
                # Check toolcontext option
                if phase.toolcontext.opt_feature_envy.value or phase.toolcontext.opt_all.value then bad_conception_elements.add(new FeatureEnvy(phase, view))
                if phase.toolcontext.opt_long_method.value or phase.toolcontext.opt_all.value then bad_conception_elements.add(new LongMethod(phase, view))
                if phase.toolcontext.opt_long_params.value or phase.toolcontext.opt_all.value then bad_conception_elements.add(new LongParameterList(phase, view))
                if phase.toolcontext.opt_no_abstract_implementation.value or phase.toolcontext.opt_all.value then bad_conception_elements.add(new NoAbstractImplementation(phase, view))
                if phase.toolcontext.opt_large_class.value or phase.toolcontext.opt_all.value then bad_conception_elements.add(new LargeClass(phase, view))
                # Collected all code smell if their state is true
                for bad_conception_element in bad_conception_elements do
                        if bad_conception_element.collect(self.mclassdef,phase.toolcontext.modelbuilder) then array_badconception.add(bad_conception_element)
                end
                # Compute global score
                collect_global_score
        end

        fun print_collected_data do
                if array_badconception.length != 0 then
                        print "--------------------"
                        print phase.toolcontext.format_h1("Full name: {mclassdef.full_name} Location: {mclassdef.location}")
                        for bad_conception in array_badconception do
                                bad_conception.print_result
                        end
                end
        end

        fun collect_global_score do
                if array_badconception.not_empty then
                        for bad_conception in array_badconception do
                                self.score += bad_conception.score
                        end
                end
        end
end

abstract class BadConception
        var phase: CodeSmellsMetricsPhase

        var view: ModelView

        var score = 0.0

        # Name
        fun name: String is abstract

        # Description
        fun desc: String is abstract

        # Collection method
        fun collect(mclassdef: MClassDef, model_builder: ModelBuilder): Bool is abstract

        # Show results in console
        fun print_result is abstract

        # Compute code smell score to sort
        fun score_rate do
                score = 1.0
        end
end

class LargeClass
        super BadConception
        var number_attribut = 0

        var number_method = 0

        redef fun name do return "LARGC"

        redef fun desc do return "Large class"

        redef fun collect(mclassdef, model_builder): Bool do
                self.number_attribut = mclassdef.collect_intro_and_redef_mattributes(view).length
                # Get the number of methods (Accessor include) (subtract the get and set of attibutes with (numberAtribut*2))
                self.number_method = mclassdef.collect_intro_and_redef_methods(view).length
                self.score_rate
                return self.number_method.to_f > phase.average_number_of_method and self.number_attribut.to_f > phase.average_number_of_attribute
        end

        redef fun print_result do
                print phase.toolcontext.format_h2("{desc}: {number_attribut} attributes and {number_method} methods ({phase.average_number_of_attribute}A {phase.average_number_of_method}M Average)")
        end

        redef fun score_rate do
                score = (number_method.to_f + number_attribut.to_f) / (phase.average_number_of_method + phase.average_number_of_attribute)
        end
end

class LongParameterList
        super BadConception
        var bad_methods = new Array[MMethodDef]

        redef fun name do return "LONGPL"

        redef fun desc do return "Long parameter list"

        redef fun collect(mclassdef, model_builder): Bool do
                for meth in mclassdef.collect_intro_and_redef_mpropdefs(view) do
                        var threshold_value = 4
                        # Get the threshold value from the toolcontext command
                        if phase.toolcontext.opt_long_params_threshold.value != 0 then threshold_value = phase.toolcontext.opt_long_params_threshold.value
                        # Check if the property is a method definition
                        if not meth isa MMethodDef then continue
                        # Check if method has a signature
                        if meth.msignature == null then continue
                        if meth.msignature.mparameters.length <= threshold_value then continue
                        self.bad_methods.add(meth)
                end
                self.score_rate
                return self.bad_methods.not_empty
        end

        redef fun print_result do
                print phase.toolcontext.format_h2("{desc}:")
                if self.bad_methods.not_empty then
                        print " Affected method(s):"
                        for method in self.bad_methods do
                                print "         -{method.name} has {method.msignature.mparameters.length} parameters"
                        end
                end
        end

        redef fun score_rate do
                if self.bad_methods.not_empty then
                        self.score = self.bad_methods.length.to_f/ phase.average_number_of_method
                end
        end
end

class FeatureEnvy
        super BadConception
        var bad_methods = new Array[MMethodDef]

        redef fun name do return "FEM"

        redef fun desc do return "Feature envy"

        redef fun collect(mclassdef, model_builder): Bool do
                var mmethoddefs = call_analyze_methods(mclassdef,model_builder, view)
                for mmethoddef in mmethoddefs do
                        var max_class_call = mmethoddef.class_call.max
                        # Check if the class with the maximum call is >= auto-call and the maximum call class is != of this class
                        if mmethoddef.class_call[max_class_call] <= mmethoddef.total_self_call or max_class_call.mclass.full_name == mclassdef.mclass.full_name then continue
                        self.bad_methods.add(mmethoddef)
                end
                self.score_rate
                return self.bad_methods.not_empty
        end

        redef fun print_result do
                print phase.toolcontext.format_h2("{desc}:")
                if self.bad_methods.not_empty then
                        print " Affected method(s):"
                        for method in self.bad_methods do
                                var max_class_call = method.class_call.max
                                if max_class_call != null then
                                        # Check if the type of max call class is generique
                                        if max_class_call.mclass.mclass_type isa MGenericType and not phase.toolcontext.opt_move_generics.value then
                                                print "         -{method.name}({method.msignature.mparameters.join(", ")}) {method.total_self_call}/{method.class_call[max_class_call]}"
                                        else
                                                print "         -{method.name}({method.msignature.mparameters.join(", ")}) {method.total_self_call}/{method.class_call[max_class_call]} move to {max_class_call}"
                                        end
                                end
                        end
                end
        end

        redef fun score_rate do
                if self.bad_methods.not_empty then
                        self.score = self.bad_methods.length.to_f / phase.average_number_of_method
                end
        end
end

class LongMethod
        super BadConception
        var bad_methods = new Array[MMethodDef]

        redef fun name do return "LONGMETH"

        redef fun desc do return "Long method"

        redef fun collect(mclassdef, model_builder): Bool do
                var mmethoddefs = call_analyze_methods(mclassdef,model_builder, view)
                var threshold_value = phase.average_number_of_lines.to_i
                # Get the threshold value from the toolcontext command
                if phase.toolcontext.opt_long_method_threshold.value != 0 then threshold_value = phase.toolcontext.opt_long_method_threshold.value

                for mmethoddef in mmethoddefs do
                        if mmethoddef.line_number <= threshold_value then continue
                        self.bad_methods.add(mmethoddef)
                end
                self.score_rate
                return self.bad_methods.not_empty
        end

        redef fun print_result do
                print phase.toolcontext.format_h2("{desc}:  Average {phase.average_number_of_lines.to_i} lines")
                if self.bad_methods.not_empty then
                        print " Affected method(s):"
                        for method in self.bad_methods do
                                print "         -{method.name} has {method.line_number} lines"
                        end
                end
        end

        redef fun score_rate do
                if self.bad_methods.not_empty then
                        self.score = self.bad_methods.length.to_f / phase.average_number_of_method
                end
        end
end

class NoAbstractImplementation
        super BadConception
        var bad_methods = new Array[MMethodDef]

        redef fun name do return "LONGMETH"

        redef fun desc do return "No Implemented abstract property"

        redef fun collect(mclassdef, model_builder): Bool do
                if not mclassdef.mclass.is_abstract and not mclassdef.mclass.is_interface then
                        if mclassdef.collect_abstract_methods(view).not_empty then
                                bad_methods.add_all(mclassdef.collect_not_define_properties(view))
                        end
                end
                self.score_rate
                return bad_methods.not_empty
        end

        redef fun print_result do
                print phase.toolcontext.format_h2("{desc}:")
                if self.bad_methods.not_empty then
                        print " Affected method(s):"
                        for method in self.bad_methods do
                                print "         -{method.name}"
                        end
                end
        end

        redef fun score_rate do
                if self.bad_methods.not_empty then
                        self.score = self.bad_methods.length.to_f / phase.average_number_of_method
                end
        end
end

redef class ModelView
        fun get_avg_parameter: Float do
                var counter = new Counter[MMethodDef]
                for mclassdef in mclassdefs do
                        for method in mclassdef.collect_intro_and_redef_mpropdefs(self) do
                        # check if the property is a method definition
                                if not method isa MMethodDef then continue
                                #Check if method has a signature
                                if method.msignature == null then continue
                                if method.msignature.mparameters.length == 0 then continue
                                counter[method] = method.msignature.mparameters.length
                        end
                end
                return counter.avg + counter.std_dev
        end

        fun get_avg_attribut: Float do
                var counter = new Counter[MClassDef]
                for mclassdef in mclassdefs do
                        var number_attributs = mclassdef.collect_intro_and_redef_mattributes(self).length
                        if number_attributs != 0 then counter[mclassdef] = number_attributs
                end
                return counter.avg + counter.std_dev
        end

        fun get_avg_method: Float do
                var counter = new Counter[MClassDef]
                for mclassdef in mclassdefs do
                        var number_methodes = mclassdef.collect_intro_and_redef_methods(self).length
                        if number_methodes != 0 then counter[mclassdef] = number_methodes
                end
                return counter.avg + counter.std_dev
        end

        fun get_avg_linenumber(model_builder: ModelBuilder): Float do
                var methods_analyse_metrics = new Counter[MClassDef]
                for mclassdef in mclassdefs do
                        var result = 0
                        var count = 0
                        for mmethoddef in call_analyze_methods(mclassdef,model_builder, self) do
                                result += mmethoddef.line_number
                                if mmethoddef.line_number == 0 then continue
                                count += 1
                        end
                        if not mclassdef.collect_local_mproperties(self).length != 0 then continue
                        if count == 0 then continue
                        methods_analyse_metrics[mclassdef] = (result/count).to_i
                end
                return methods_analyse_metrics.avg + methods_analyse_metrics.std_dev
        end
end

class BadConceptionComparator
        super Comparator
        redef type COMPARED: BadConceptionFinder
        redef fun compare(a,b) do
                var test = a.array_badconception.length <=> b.array_badconception.length
                if test == 0 then
                        return a.score <=> b.score
                end
                return a.array_badconception.length <=> b.array_badconception.length
        end
end