quiet_library <- function(...) { suppressPackageStartupMessages(library(...)) }
quiet_library(hise)
quiet_library(dplyr)
quiet_library(purrr)
quiet_library(ggplot2)

BR1_rna_desc <- getFileDescriptors(
    fileType = "scRNA-seq-labeled", 
    filter = list(cohort.cohortGuid = "BR1"))
BR2_rna_desc <- getFileDescriptors(
    fileType = "scRNA-seq-labeled", 
    filter = list(cohort.cohortGuid = "BR2"))

BR1_rna_desc <- fileDescToDataframe(BR1_rna_desc)
BR2_rna_desc <- fileDescToDataframe(BR2_rna_desc)

meta_data <- plyr::rbind.fill(BR1_rna_desc , BR2_rna_desc )

meta_data <- meta_data %>%
  filter(!grepl("EXP",file.batchID)) %>%
  filter(!file.batchID == "B004")

non_healthy <- meta_data %>%
  filter(sample.diseaseStatesRecordedAtVisit != "") %>%
  select(subject.subjectGuid, sample.diseaseStatesRecordedAtVisit) %>%
  unique()

non_healthy

meta_data <- meta_data %>%
  filter(!subject.subjectGuid %in% non_healthy$subject.subjectGuid)

reference_sample_uuid <- '2da66a1a-17cc-498b-9129-6858cf639caf'

hise_res <- cacheFiles(list(reference_sample_uuid))

reference_meta <- read.csv("cache/2da66a1a-17cc-498b-9129-6858cf639caf/ref_h5_meta_data_2024-02-18.csv")

meta_data$pbmc_sample_id <- gsub("_","",paste0("PB0",substr(sub(".*PB0", "", meta_data$file.name),1,8)))

reference_meta_data <- meta_data %>%
  filter(sample.sampleKitGuid %in% reference_meta$sample.sampleKitGuid) %>%
  filter(file.id %in% reference_meta$file.id) %>%
  mutate(sample.reference = "in_reference")

non_reference_meta_data <- meta_data %>%
  filter(!sample.sampleKitGuid %in% reference_meta$sample.sampleKitGuid) %>%
  arrange(pbmc_sample_id) %>%
  filter(!duplicated(sample.sampleKitGuid, fromLast = TRUE)) %>%
  mutate(sample.reference = "not_in_reference")

meta_data <- rbind(reference_meta_data, non_reference_meta_data)
meta_data <- meta_data %>%
  arrange(sample.sampleKitGuid)

table(meta_data$sample.reference)

nrow(meta_data)

table(meta_data$cohort.cohortGuid)

meta_data <- meta_data %>%
  mutate(sample.visitName = ifelse(
      sample.visitName == "Other - Non-Flu",
      sample.visitDetails, sample.visitName
  ))

options(repr.plot.width = 6, repr.plot.height = 16)
ggplot() +
  geom_tile(
      data = meta_data,
      aes(x = sample.visitName,
          y = subject.subjectGuid,
          fill = sample.visitName),
      color = "black"
  ) +
  scale_fill_brewer(
      type = "qual",
      palette = 3
  ) +
  scale_x_discrete(expand = c(0,0)) +
  scale_y_discrete(expand = c(0,0)) +
  theme_bw() +
  theme(axis.text.x = element_text(angle = 90, hjust = 1, vjust = 0.3))

options(repr.plot.width = 16, repr.plot.height = 16)
ggplot() +
  geom_tile(
      data = meta_data,
      aes(x = file.batchID,
          y = subject.subjectGuid,
          fill = sample.visitName),
      color = "black"
  ) +
  scale_fill_brewer(
      type = "qual",
      palette = 3
  ) +
  scale_x_discrete(expand = c(0,0)) +
  scale_y_discrete(expand = c(0,0)) +
  theme_bw() +
  theme(axis.text.x = element_text(angle = 90, hjust = 1, vjust = 0.3))

if(!dir.exists("output")) {
    dir.create("output")
}

out_file <- file.path(
    "output",
    paste0("diha_sample_meta_data_",Sys.Date(),".csv")
)

write.csv(
    meta_data,
    out_file,
    row.names = FALSE,
    quote = FALSE
)

res <- map(
    meta_data$file.id,
    function(file.id) {
        capture.output(cacheFiles(list(file.id)))
    }
)

study_space_uuid <- "de025812-5e73-4b3c-9c3b-6d0eac412f2a"
title <- paste("DIHA Sample Metadata", Sys.Date())

in_list <- as.list(meta_data$file.id)

out_list <- list(out_file)

uploadFiles(
    files = out_list,
    studySpaceId = study_space_uuid,
    title = title,
    inputFileIds = in_list,
    store = "project",
    doPrompt = FALSE
)

sessionInfo()