knitr::opts_chunk$set(echo = TRUE)

knitr::opts_knit$set(root.dir = '/work/CogSci_Methods01/portfolio_assignment_03-Ida0601/data') 

library(pacman)
p_load(tidyverse)
p_load(pastecs)
library("car")
library(data.table)
library(vroom)

#### Load and anonymize data ####

if (is_empty(list.files(path =  getwd(), pattern = "*SART_Anonymous*", full.names = T))){
  
files <- list.files(path = getwd(), pattern = "*SART*", full.names = T) 
  
  data_out <- list() 
  
  num_files <- length(files) 
  
  rand_ids <- sample(seq(1,num_files,1)) 
  
  cnt_f <- 0 
  
  for (f in files){
    
    cnt_f <- cnt_f + 1
    
    data_out[[f]] <- read_csv(file = f, col_names = TRUE)
    
    data_out[[f]]$ID <- paste(c("Anonymous", rand_ids[cnt_f]), collapse = "")
    
    out_name <- paste(c(getwd(), "/SART_", unique(data_out[[f]]$ID[1]), ".txt"), collapse = "")
    
    write_csv(data_out[[f]], out_name, na = "NA")
    
    file.remove(f)
    
  }
}
  

filelist = list.files(pattern = ".*.txt")

df <- vroom(filelist)

# Making new dataframes

#Making a dataframe for reaction times:
df_rt <- df %>% dplyr::filter(resp_acc == 1, block_num == 1, number != 3)

df_rt <- df_rt %>%
  group_by(part_ID, Condition) %>%
  summarise(mean = mean(resp_rt))

#Making a dataframe for accuracy
df_acc <- summarise(df, part_ID, block_num = 1, number = 3, resp_acc, Condition)

df_acc <- df_acc %>% group_by(part_ID, Condition) %>% summarise(resp_acc=mean(resp_acc))

# Checking for normality

#Shapiro-Wilks test of normality
df_rt %>% group_by(Condition) %>% 
  summarise(
    normality = shapiro.test(mean)$p.value
  )

df_acc %>% group_by(Condition) %>% 
  summarise(
    normality = shapiro.test(resp_acc)$p.value
  )

# Anova tests

#Making an ANOVA test for accuracy
anova_acc_model <- aov(resp_acc ~ Condition, data = df_acc)
summary(anova_acc_model)

#Making an ANOVA test for RT
anova_model_meanRT <- aov(mean ~ Condition, data = df_rt)
summary(anova_model_meanRT)

# Making boxplots

#Visualization - accuracy
df_acc %>% ggplot(aes(x= Condition, y = resp_acc, fill = Condition))+
  geom_boxplot()+
  theme_light()+
  ggtitle("Accuracy according to condition")

# Making boxplot for mean reaction time 
df_rt %>% ggplot(aes(x = Condition, y = mean, fill = Condition))+
  geom_boxplot()+
  theme_light()+
  ggtitle("Mean reaction time according to condition")

# Doing post-hoc comparisons

# Making multiple t-test for mean reaction time + accuracy between the three condtions (A, B and C)
TukeyHSD(anova_model_meanRT)
TukeyHSD(anova_acc_model)

# Checking assumptions

#homogeneity test of variance 
# Reaction time
leveneTest(mean ~ Condition, data = df_rt)
# Accuracy 
leveneTest(resp_acc ~ Condition, data = df_acc)

## Checking residuals
plot(anova_model_meanRT)
plot(anova_acc_model)