Misc polish

.gitignore

@@ -1,11 +1,17 @@
 *.jl.*.cov
 *.jl.cov
 *.jl.mem
-/docs/build/
 .DS_Store
+.vscode
+
 testing_script*.jl
 settings.json
+
 Manifest.toml
 docs/Manifest.toml
 test/Manifest.toml
-/docs/src/generated_markdowns/*.md
+
+/docs/src/generated
+/docs/src/index.md
+/docs/build/
+/build

Project.toml

@@ -4,15 +4,15 @@ authors = [ "Peter Thestrup Waade [email protected]",
             "Anna Hedvig Møller [email protected]",
             "Jacopo Comoglio [email protected]",
             "Christoph Mathys [email protected]"]
-version = "0.3.3"
+version = "0.5.0"
 
 [deps]
 ActionModels = "320cf53b-cc3b-4b34-9a10-0ecb113566a3"
 Distributions = "31c24e10-a181-5473-b8eb-7969acd0382f"
 RecipesBase = "3cdcf5f2-1ef4-517c-9805-6587b60abb01"
 
 [compat]
-ActionModels = "0.4"
+ActionModels = "0.5"
 Distributions = "0.25"
 RecipesBase = "1"
 julia = "1.9"

README.md

@@ -38,7 +38,7 @@ premade_agent("help")
 ### Create agent
 
 ````@example index
-agent = premade_agent("hgf_binary_softmax_action")
+agent = premade_agent("hgf_binary_softmax")
 ````
 
 ### Get states and parameters
@@ -55,10 +55,10 @@ get_parameters(agent)
 
 ![Image1](docs/src/images/readme/get_parameters.png)
 
-Set a new parameter for initial precision of x2 and define some inputs
+Set a new parameter for initial precision of xprob and define some inputs
 
 ````@example index
-set_parameters!(agent, ("x2", "initial_precision"), 0.9)
+set_parameters!(agent, ("xprob", "initial_precision"), 0.9)
 inputs = [1, 0, 1, 1, 1, 0, 1, 0, 1, 0, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 0, 0, 0, 0];
 nothing #hide
 ````
@@ -75,23 +75,23 @@ actions = give_inputs!(agent, inputs)
 using StatsPlots
 using Plots
 plot_trajectory(agent, ("u", "input_value"))
-plot_trajectory!(agent, ("x1", "prediction"))
+plot_trajectory!(agent, ("x", "prediction"))
 ````
 ![Image1](docs/src/images/readme/plot_trajectory.png)
 
-Plot state trajectory of input value, action and prediction of x1
+Plot state trajectory of input value, action and prediction of x
 
 ````@example index
 plot_trajectory(agent, ("u", "input_value"))
 plot_trajectory!(agent, "action")
-plot_trajectory!(agent, ("x1", "prediction"))
+plot_trajectory!(agent, ("x", "prediction"))
 ````
 ![Image1](docs/src/images/readme/plot_trajectory_2.png)
 ### Fitting parameters
 
 ````@example index
 using Distributions
-prior = Dict(("x2", "volatility") => Normal(1, 0.5))
+prior = Dict(("xprob", "volatility") => Normal(1, 0.5))
 
 model = fit_model(agent, prior, inputs, actions, n_iterations = 20)
 ````

docs/Project.toml

@@ -5,10 +5,10 @@ DataFrames = "a93c6f00-e57d-5684-b7b6-d8193f3e46c0"
 Distributed = "8ba89e20-285c-5b6f-9357-94700520ee1b"
 Distributions = "31c24e10-a181-5473-b8eb-7969acd0382f"
 Documenter = "e30172f5-a6a5-5a46-863b-614d45cd2de4"
+Glob = "c27321d9-0574-5035-807b-f59d2c89b15c"
 HierarchicalGaussianFiltering = "63d42c3e-681c-42be-892f-a47f35336a79"
 JuliaFormatter = "98e50ef6-434e-11e9-1051-2b60c6c9e899"
 Literate = "98b081ad-f1c9-55d3-8b20-4c87d4299306"
 Plots = "91a5bcdd-55d7-5caf-9e0b-520d859cae80"
-RecipesBase = "3cdcf5f2-1ef4-517c-9805-6587b60abb01"
 StatsPlots = "f3b207a7-027a-5e70-b257-86293d7955fd"
 Turing = "fce5fe82-541a-59a6-adf8-730c64b5f9a0"

docs/src/Julia_src_files/index.jl → docs/julia_files/index.jl

@@ -24,15 +24,15 @@ using ActionModels
 premade_agent("help")
 
 # ### Create agent
-agent = premade_agent("hgf_binary_softmax_action")
+agent = premade_agent("hgf_binary_softmax")
 
 # ### Get states and parameters
 get_states(agent)
 #-
 get_parameters(agent)
 
-# Set a new parameter for initial precision of x2 and define some inputs
-set_parameters!(agent, ("x2", "initial_precision"), 0.9)
+# Set a new parameter for initial precision of xprob and define some inputs
+set_parameters!(agent, ("xprob", "initial_precision"), 0.9)
 inputs = [1, 0, 1, 1, 1, 0, 1, 0, 1, 0, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 0, 0, 0, 0];
 
 # ### Give inputs to the agent
@@ -42,18 +42,18 @@ actions = give_inputs!(agent, inputs)
 using StatsPlots
 using Plots
 plot_trajectory(agent, ("u", "input_value"))
-plot_trajectory!(agent, ("x1", "prediction"))
+plot_trajectory!(agent, ("xbin", "prediction"))
 
-# Plot state trajectory of input value, action and prediction of x1
+# Plot state trajectory of input value, action and prediction of xbin
 plot_trajectory(agent, ("u", "input_value"))
 plot_trajectory!(agent, "action")
-plot_trajectory!(agent, ("x1", "prediction"))
+plot_trajectory!(agent, ("xbin", "prediction"))
 
 
 # ### Fitting parameters
 
 using Distributions
-prior = Dict(("x2", "volatility") => Normal(1, 0.5))
+prior = Dict(("xprob", "volatility") => Normal(1, 0.5))
 
 model = fit_model(agent, prior, inputs, actions, n_iterations = 20)
 

docs/julia_files/tutorials/classic_JGET.jl

@@ -0,0 +1,43 @@
+using ActionModels, HierarchicalGaussianFiltering
+using CSV, DataFrames
+using Plots, StatsPlots
+using Distributions
+
+
+# Get the path for the HGF superfolder
+hgf_path = dirname(dirname(pathof(HierarchicalGaussianFiltering)))
+# Add the path to the data files
+data_path = hgf_path * "/docs/src/tutorials/data/"
+
+#Load data
+data = CSV.read(data_path * "classic_cannonball_data.csv", DataFrame)
+inputs = data[(data.ID.==20).&(data.session.==1), :].outcome
+
+#Create HGF
+hgf = premade_hgf("JGET", verbose = false)
+#Create agent
+agent = premade_agent("hgf_gaussian", hgf)
+#Set parameters
+parameters = Dict(
+    "action_noise" => 1,
+    ("u", "input_noise") => 0,
+    ("x", "initial_mean") => first(inputs) + 2,
+    ("x", "initial_precision") => 0.001,
+    ("x", "volatility") => -8,
+    ("xvol", "volatility") => -8,
+    ("xnoise", "volatility") => -7,
+    ("xnoise_vol", "volatility") => -2,
+    ("x", "xvol", "coupling_strength") => 1,
+    ("xnoise", "xnoise_vol", "coupling_strength") => 1,
+)
+set_parameters!(agent, parameters)
+reset!(agent)
+
+#Simulate updates and actions
+actions = give_inputs!(agent, inputs);
+#Plot belief trajectories
+plot_trajectory(agent, "u")
+plot_trajectory!(agent, "x")
+plot_trajectory(agent, "xvol")
+plot_trajectory(agent, "xnoise")
+plot_trajectory(agent, "xnoise_vol")

docs/src/tutorials/classic_binary.jl → docs/julia_files/tutorials/classic_binary.jl

@@ -14,7 +14,7 @@ using Distributions
 # Get the path for the HGF superfolder
 hgf_path = dirname(dirname(pathof(HierarchicalGaussianFiltering)))
 # Add the path to the data files
-data_path = hgf_path * "/docs/src/tutorials/data/"
+data_path = hgf_path * "/docs/julia_files/tutorials/data/"
 
 # Load the data 
 inputs = CSV.read(data_path * "classic_binary_inputs.csv", DataFrame)[!, 1];
@@ -23,59 +23,57 @@ inputs = CSV.read(data_path * "classic_binary_inputs.csv", DataFrame)[!, 1];
 hgf_parameters = Dict(
     ("u", "category_means") => Real[0.0, 1.0],
     ("u", "input_precision") => Inf,
-    ("x2", "volatility") => -2.5,
-    ("x2", "initial_mean") => 0,
-    ("x2", "initial_precision") => 1,
-    ("x3", "volatility") => -6.0,
-    ("x3", "initial_mean") => 1,
-    ("x3", "initial_precision") => 1,
-    ("x1", "x2", "value_coupling") => 1.0,
-    ("x2", "x3", "volatility_coupling") => 1.0,
+    ("xprob", "volatility") => -2.5,
+    ("xprob", "initial_mean") => 0,
+    ("xprob", "initial_precision") => 1,
+    ("xvol", "volatility") => -6.0,
+    ("xvol", "initial_mean") => 1,
+    ("xvol", "initial_precision") => 1,
+    ("xbin", "xprob", "coupling_strength") => 1.0,
+    ("xprob", "xvol", "coupling_strength") => 1.0,
 );
 
 hgf = premade_hgf("binary_3level", hgf_parameters, verbose = false);
 
 # Create an agent
-agent_parameters = Dict("sigmoid_action_precision" => 5);
-agent =
-    premade_agent("hgf_unit_square_sigmoid_action", hgf, agent_parameters, verbose = false);
+agent_parameters = Dict("action_noise" => 0.2);
+agent = premade_agent("hgf_unit_square_sigmoid", hgf, agent_parameters, verbose = false);
 
 # Evolve agent and save actions
 actions = give_inputs!(agent, inputs);
 
 # Plot the trajectory of the agent
 plot_trajectory(agent, ("u", "input_value"))
-plot_trajectory!(agent, ("x1", "prediction"))
-
+plot_trajectory!(agent, ("xbin", "prediction"))
 
 # -
 
-plot_trajectory(agent, ("x2", "posterior"))
-plot_trajectory(agent, ("x3", "posterior"))
+plot_trajectory(agent, ("xprob", "posterior"))
+plot_trajectory(agent, ("xvol", "posterior"))
 
 # Set fixed parameters
 fixed_parameters = Dict(
-    "sigmoid_action_precision" => 5,
+    "action_noise" => 0.2,
     ("u", "category_means") => Real[0.0, 1.0],
     ("u", "input_precision") => Inf,
-    ("x2", "initial_mean") => 0,
-    ("x2", "initial_precision") => 1,
-    ("x3", "initial_mean") => 1,
-    ("x3", "initial_precision") => 1,
-    ("x1", "x2", "value_coupling") => 1.0,
-    ("x2", "x3", "volatility_coupling") => 1.0,
-    ("x3", "volatility") => -6.0,
+    ("xprob", "initial_mean") => 0,
+    ("xprob", "initial_precision") => 1,
+    ("xvol", "initial_mean") => 1,
+    ("xvol", "initial_precision") => 1,
+    ("xbin", "xprob", "coupling_strength") => 1.0,
+    ("xprob", "xvol", "coupling_strength") => 1.0,
+    ("xvol", "volatility") => -6.0,
 );
 
 # Set priors for parameter recovery
-param_priors = Dict(("x2", "volatility") => Normal(-3.0, 0.5));
+param_priors = Dict(("xprob", "volatility") => Normal(-3.0, 0.5));
 #-
 # Prior predictive plot
 plot_predictive_simulation(
     param_priors,
     agent,
     inputs,
-    ("x1", "prediction_mean"),
+    ("xbin", "prediction_mean"),
     n_simulations = 100,
 )
 #-
@@ -104,6 +102,6 @@ plot_predictive_simulation(
     fitted_model,
     agent,
     inputs,
-    ("x1", "prediction_mean"),
+    ("xbin", "prediction_mean"),
     n_simulations = 3,
 )

docs/src/tutorials/classic_usdchf.jl → docs/julia_files/tutorials/classic_usdchf.jl

@@ -12,7 +12,7 @@ using Distributions
 # Get the path for the HGF superfolder
 hgf_path = dirname(dirname(pathof(HierarchicalGaussianFiltering)))
 # Add the path to the data files
-data_path = hgf_path * "/docs/src/tutorials/data/"
+data_path = hgf_path * "/docs/julia_files/tutorials/data/"
 
 # Load the data
 inputs = Float64[]
@@ -24,20 +24,19 @@ end
 
 #Create HGF
 hgf = premade_hgf("continuous_2level", verbose = false);
-agent = premade_agent("hgf_gaussian_action", hgf, verbose = false);
+agent = premade_agent("hgf_gaussian", hgf, verbose = false);
 
 # Set parameters for parameter recover
 parameters = Dict(
-    ("u", "x1", "value_coupling") => 1.0,
-    ("x1", "x2", "volatility_coupling") => 1.0,
+    ("x", "xvol", "coupling_strength") => 1.0,
     ("u", "input_noise") => -log(1e4),
-    ("x1", "volatility") => -13,
-    ("x2", "volatility") => -2,
-    ("x1", "initial_mean") => 1.04,
-    ("x1", "initial_precision") => 1 / (0.0001),
-    ("x2", "initial_mean") => 1.0,
-    ("x2", "initial_precision") => 1 / 0.1,
-    "gaussian_action_precision" => 100,
+    ("x", "volatility") => -13,
+    ("xvol", "volatility") => -2,
+    ("x", "initial_mean") => 1.04,
+    ("x", "initial_precision") => 1 / (0.0001),
+    ("xvol", "initial_mean") => 1.0,
+    ("xvol", "initial_precision") => 1 / 0.1,
+    "action_noise" => 0.01,
 );
 
 set_parameters!(agent, parameters)
@@ -59,7 +58,7 @@ plot_trajectory(
     xlabel = "Trading days since 1 January 2010",
 )
 #-
-plot_trajectory!(agent, ("x1", "posterior"), color = "red")
+plot_trajectory!(agent, ("x", "posterior"), color = "red")
 plot_trajectory!(
     agent,
     "action",
@@ -71,7 +70,7 @@ plot_trajectory!(
 #-
 plot_trajectory(
     agent,
-    "x2",
+    "xvol",
     color = "blue",
     size = (1300, 500),
     xlims = (0, 615),
@@ -81,27 +80,26 @@ plot_trajectory(
 #-
 # Set priors for fitting
 fixed_parameters = Dict(
-    ("u", "x1", "value_coupling") => 1.0,
-    ("x1", "x2", "volatility_coupling") => 1.0,
-    ("x1", "initial_mean") => 0,
-    ("x1", "initial_precision") => 2000,
-    ("x2", "initial_mean") => 1.0,
-    ("x2", "initial_precision") => 600.0,
-    "gaussian_action_precision" => 100,
+    ("x", "xvol", "coupling_strength") => 1.0,
+    ("x", "initial_mean") => 0,
+    ("x", "initial_precision") => 2000,
+    ("xvol", "initial_mean") => 1.0,
+    ("xvol", "initial_precision") => 600.0,
 );
 
 param_priors = Dict(
     ("u", "input_noise") => Normal(-6, 1),
-    ("x1", "volatility") => Normal(-4, 1),
-    ("x2", "volatility") => Normal(-4, 1),
+    ("x", "volatility") => Normal(-4, 1),
+    ("xvol", "volatility") => Normal(-4, 1),
+    "action_noise" => LogNormal(log(0.01), 1),
 );
 #-
 # Prior predictive simulation plot
 plot_predictive_simulation(
     param_priors,
     agent,
     inputs,
-    ("x1", "posterior_mean");
+    ("x", "posterior_mean");
     n_simulations = 100,
 )
 #-