using CurricularAnalytics, CurricularVisualization
#using CurricularOptimization # this package requires special access rights, if you commented out, you won't be able to run the optimzation-related functions
using Glob
using CSV
using DataFrames
using Statistics

UA_ChemE_plan = read_csv("./programs/plans/ChemE-DP-UA.csv")
visualize(UA_ChemE_plan, notebook=true)

metrics = basic_metrics(UA_ChemE_plan)
println(String(take!(metrics)))

metrics = basic_metrics(UA_ChemE_plan.curriculum)
println(String(take!(metrics)))

df = DataFrame(CSV.File("./Complexity_x_Grad_UA.csv"));
histogram(df.complexity_score, nbins=50, xlabel="Program Complexity", ylabel="# of programs", label="")

 = Array{Integer,1}() # curricular complexity score of program
Y = Array{Bool,1}() # did student graduate within 4 years?
Z = Array{Bool,1}() # did student graduate within 6 years?
for row in eachrow(df)
  for t in 1:row["four"] 
    push!(X, row["complexity_score"])
    push!(Y, 1); push!(Z, 1)
  end
  s = row["total_degrees_awarded"] - row["six"]
  for t in 1:(row["total_degrees_awarded"] - row["four"])
    push!(X, row["complexity_score"])
    push!(Y, 0)
    s > 0 ? (push!(Z, 1); s=s-1) : push!(Z, 0)
  end
end
df_binary = DataFrame(Program_Complexity = X, grad4 = Y, grad6 = Z);
histogram(X, nbins=50, xlabel="Program Complexity", ylabel="# of students", label="")

probit4 = glm(@formula(grad4 ~ Program_Complexity), df_binary, Binomial(), ProbitLink())

probit6 = glm(@formula(grad6 ~ Program_Complexity), df_binary, Binomial(), ProbitLink())

CC_scores = DataFrame(Program_Complexity = 0:10:650)
pred4 = predict(probit4, CC_scores, interval = :confidence)
pred6 = predict(probit6, CC_scores, interval = :confidence)
plot(title="Graduates")
plot!(CC_scores.Program_Complexity, pred4.prediction , seriestype = :scatter, labels="4 years",
       ribbon = (pred4.prediction .- pred4.lower, pred4.upper .- pred4.prediction))
plot!(CC_scores.Program_Complexity, pred6.prediction , seriestype = :scatter, labels="6 years",
       ribbon = (pred6.prediction .- pred6.lower, pred6.upper .- pred6.prediction))

currics = Array{Curriculum,1}()
program_files = glob("*", "./programs/curricula")
for program in program_files
    curric = read_csv(program)
    complexity(curric)  # compute the curricular complexity of the curriculum
    push!(currics, curric)    # store the curriculum in the darray 
end

for c in currics
    println(c.institution)
end

metric = "complexity"
stats = basic_statistics(currics, "complexity")
println(String(take!(stats))); flush(stdout)  
display(metric_histogram(currics, metric, title="Chemical Engineering Programs", xlabel="$(metric)", 
          ylabel="# of curricula", xlim=(0,900)))

metrics = basic_metrics(AE_plan.curriculum)
println(String(take!(metrics)))

metrics = basic_metrics(EE_plan.curriculum)
println(String(take!(metrics)))

curricula = Array{Curriculum,1}()
metric = "complexity"
for (key, val) in plans
    push!(curricula, val.curriculum)
end
display(metric_histogram(curricula, metric, title=" Programs", xlabel="$(metric)", ylabel="# of curricula", xlim=(0,500)))

for plan in plans
    extraneous_requisites(plan[2].curriculum, print=true)
end

prefixes = Dict{String, Array{String,1}}()
prefixes["Electrical Engineering"] = ["ECE"];
prefixes["Electrical Engineering - Electronic Devices and Circuit Technology"] = ["ECE"];
prefixes["Aerospace Engineering"] = ["AE"];

for plan in plans
    de = dead_ends(plan[2].curriculum, prefixes[plan[2].curriculum.name])
    println("\nDead end courses in the $(plan[2].curriculum.name) curriculum:")
    for course in de[2]
        println("$(course.prefix) $(course.num): $(course.name)")
    end
end

# Uncomment the following two lines if the Gurobi package has not yet been included in your Julia environment.
#using Pkg
#Pkg.add("Gurobi")
using Gurobi

metrics = basic_metrics(EE_plan)
println(String(take!(metrics)))

EE_new_1 = optimize_plan(EE_plan.curriculum, 8, 12, 18, balance_obj);
visualize(EE_new_1, notebook=true)

metrics = basic_metrics(EE_new_1)
println(String(take!(metrics)))

#curric = deepcopy(EE_plan.curriculum)
#convert_ids(curric)
#toxic = Array{Pair{Course,Course},1}()
# Differential Calc is toxic to Linear Algebra
#push!(toxic, course(curric, "MATH", "1551", "Differential Calculus", "") => course(curric, "MATH", "1554", "Linear Algebra", ""))

EE_new_2 = optimize_plan(EE_plan.curriculum, 8, 12, 18, req_distance_obj);
visualize(EE_new_2, notebook=true)

metrics = basic_metrics(EE_new_2)
println(String(take!(metrics)))

EE_new_3 = optimize_plan(EE_plan.curriculum, 8, 12, 18, [balance_obj, req_distance_obj]);
visualize(EE_new_3, notebook=true)

metrics = basic_metrics(EE_new_3)
println(String(take!(metrics)))

#EE_new_4 = optimize_plan(EE_plan.curriculum, 8, 12, 18, [balance_obj, req_distance_obj], toxic_courses=toxic);
#visualize(EE_new_4, notebook=true)