Add state charts as depds

Project.toml

@@ -19,13 +19,16 @@ StructEquality = "6ec83bb0-ed9f-11e9-3b4c-2b04cb4e219c"
 [weakdeps]
 AlgebraicPetri = "4f99eebe-17bf-4e98-b6a1-2c4f205a959b"
 Makie = "ee78f7c6-11fb-53f2-987a-cfe4a2b5a57a"
+StateCharts = "bfc335c0-d564-425b-98aa-d5c699c8b7ab"
 
 [extensions]
 MakieExt = "Makie"
 PetriInterface = "AlgebraicPetri"
+StateChartsInterface = "StateCharts"
 
 [compat]
 AlgebraicPetri = "0.9"
+StateCharts = "0.0.1"
 AlgebraicRewriting = "^0.4.0"
 Catlab = "^0.16.16"
 CompetingClocks = "0.1"

docs/Project.toml

@@ -11,4 +11,6 @@ Documenter = "e30172f5-a6a5-5a46-863b-614d45cd2de4"
 Literate = "98b081ad-f1c9-55d3-8b20-4c87d4299306"
 LiveServer = "16fef848-5104-11e9-1b77-fb7a48bbb589"
 Makie = "ee78f7c6-11fb-53f2-987a-cfe4a2b5a57a"
+Random = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"
+StateCharts = "bfc335c0-d564-425b-98aa-d5c699c8b7ab"
 StructEquality = "6ec83bb0-ed9f-11e9-3b4c-2b04cb4e219c"

docs/literate/sir_statechart_with_networks.jl

@@ -0,0 +1,128 @@
+# # Petri-net based ABMs
+# 
+# First we want to load our packages with `using
+
+using AlgebraicABMs, Catlab, DataMigrations
+using StateCharts
+using Makie, CairoMakie
+
+
+ENV["JULIA_DEBUG"] = "AlgebraicABMs";  # turn on @debug messages for this package
+
+# SIR Agent-based model
+# This model is the same  as the AnyLogic example model "SIR Agent-Based Networks.". 
+# This model currently support three types of networks:
+#   1. Random (if the parameter "p_random_connect" = 1.0)
+#   2. Ring lattice (if the parameter "p_random_connect" = 0.0)
+#   3. Small world (if the parameter "p_random_connect":  0.0 < p_random_connect < 1.0 )
+#       it indicates that there are "p_random_connect" percentage connections are radom connections, and 
+#       ( 1 - "p_random_connect" ) percentage connections are connections are Ring lattice connections.
+# This example model in Anylogic can be found: https://cloud.anylogic.com/model/7088e817-1dab-42b1-89fe-b19bc0a823e1?mode=SETTINGS
+
+# Step 1: define parameters
+# parameters of state transition
+total_population = 50 # Unit: persons
+frac_initial_infected = 0.02
+contact_rate = 5 # Unit: contacts per day
+infectivity = 0.01
+average_illness_duration = 15 # days
+# parameters define the network
+average_connections = 2
+p_random_connect = 0.9
+
+# Step 2: create the StateChart
+
+# function UnitStateChartF takes in 3 input argummapents:
+# states: tuple/array of all states
+# transitions: tuple/array of all transitions. The syntax of each transition:
+#              transition_name => (source_state => target_state, tranisiton_type => values)
+# alternative transitions: tuple/array of all alternative transitions. The syntax of each altanative transition:
+#              :source_transition_name=>((alternative_transition_name, probability_value) => target_state)
+states = [:S, :I, :R] # define the states
+transitions = [ :Infection => (:S => :I, :Pattern => 1), :Recovery => (:I => :R, :Rate => 1.0 / average_illness_duration)]
+alternatives = [] # this model does not include alternative transitions
+# 2.1 create the Statechart
+SIRStatechart = UnitStateChartF(states, transitions, alternatives)
+# 2.2 Visualization of the Statechart
+stateColors = Dict(:S => "green",:I => "red",:R => "gray"); # argument define colors of each state in StateChart
+StateCharts.Graph(SIRStatechart, stateColors = stateColors)
+
+# Step 3: create the model schema
+### generate the model schema by composing auto state chart schema and the network schema
+
+# 3.1 define the persons object named :V, because we plan to compose the schema_statechart with graph schema by identifying ":V"
+schema_statechart = StateChartABMSchema_MultipleObjects(SIRStatechart,:V) |> schemaACSet |> schemaPresent
+# 3.2 compose the state chart schema with the network schema (symmetric reflective graph) as the ABM model schema (without equations)
+schema_model_without_equations = compose(Open([:S],schemaACSet(schema_statechart),[:V]), Open([:V],schemaACSet(SchUndirectedReflectiveNetwork),[:E])) |> apex |> schemaPresent 
+# 3.3 add the composition equations of the model schema, since those equations disapper after composition. This would be fixed in the future using GATlab
+@present schema_model <: schema_model_without_equations begin
+    compose(inv,inv) == id(E)
+    compose(inv,src) == tgt
+    compose(inv,tgt) == src
+    compose(refl, src) == id(V)
+    compose(refl, tgt) == id(V)
+    compose(refl, inv) == refl 
+end
+
+@acset_type SIRModelStaticNet(schema_model, index=[:src,:tgt]){Symbol} <: AbstractSymmetricReflexiveGraph
+# show the model schema
+to_graphviz(schema_model; prog="dot")
+
+# Step 4: Data Migration from the ACSet with schema -- schema_statechart to the ACSet with schema -- schema_model
+# 4.1 define the data migration rule to automatically migrate ACSets from pure state chart schema to the composed model shema    
+const migrate_rule = @migration schema_model schema_statechart begin
+    V => V; 
+    ID=>ID; VID => VID
+    S => S; SV => SV
+    I => I; IV => IV
+    R => R; RV => RV
+
+    E => @product begin
+        p1::V
+        p2::V
+    end
+    src => p1
+    tgt => p2
+    inv => begin
+        p1 => p2
+        p2 => p1
+    end
+end
+
+# 4.2 define the user_defined rewrite rules: rewrite rules for contacts of Infectives
+# Note that each transitions_rule has a pair: timer=>rule
+transition_rules = [ ContinuousHazard( 1.0 / (contact_rate * infectivity)) => make_infectious_rule_MultipleObjects(SIRStatechart, [[:S],[:I]],[:I],[[:I],[:I]], :V; use_DataMigration = true, acset = SIRModelStaticNet, migration_rule = migrate_rule)]
+
+# Step 5: define the network
+network = smallworldNetWork(Int(total_population), average_connections, p_random_connect);
+
+# Step 6: Initialization
+# In the initialization, randomly assign "total_population * frac_initial_infected" persons as Infectives (state I), and the rest persons are all Susceptibles (state S)
+init = radomlyAssignInitialInfectives(SIRModelStaticNet(), Int(total_population), Int(frac_initial_infected * total_population); obn=:V, network = network)
+
+# Step 7: run the ABM model
+ts = 50.0 # the total running time
+abm = make_ABM(SIRStatechart, transition_rules, :V; is_schema_singObject=false, use_DataMigration=true, acset = SIRModelStaticNet, migration_rule = migrate_rule)
+res = run!(abm, init; maxtime=ts);
+
+# Step 8: Visualization of results
+# 8.1 plot out the time series of each state
+Makie.plot(res; Dict(o=>X->nparts(X,o) for o in states)...)
+# 8.2 show the networks at time t
+t0 = 0 # initial state
+StateCharts.Graph(res,t0,stateColors)
+
+t10 = 10 # initial state
+StateCharts.Graph(res,t10,stateColors)
+
+t20 = 20
+StateCharts.Graph(res,t20,stateColors)
+
+t30 = 30
+StateCharts.Graph(res,t30,stateColors)
+
+t40 = 40
+StateCharts.Graph(res,t40,stateColors)
+
+t50 = Int(ts) # final state
+StateCharts.Graph(res,t50,stateColors)