library(tidyverse)
library(arrow)
library(here)
library(yaml)

# from fetchngs, the multiqc metadata has an easy way to map between
# accessions and samples
multiqc_chec_config = yaml::read_yaml("data/mahendrawada_multiqc_config.yml")
chec_meta <- map_dfr(multiqc_chec_config$sample_names_rename,
                     ~tibble(!!!setNames(., multiqc_chec_config$sample_names_rename_buttons)))

genetable = arrow::read_parquet("~/code/hf/yeast_genome_resources/brentlab_features.parquet")

parsed_chec_meta = chec_meta %>%
    mutate(mahendrawada_symbol = str_remove(`sample_description"`, "_\\w_ChEC-seq"),
           replicate = str_remove_all(str_extract(`sample_description"`, "_(A|B|C)"), "_")) %>%
    select(sample, mahendrawada_symbol, replicate) %>%
    mutate(mahendrawada_symbol = str_remove(mahendrawada_symbol, "_(A|B|C)$")) %>%
    mutate(
        condition = str_remove(str_extract(mahendrawada_symbol, "_.*"), "^_"),
        mahendrawada_symbol = toupper(str_remove(mahendrawada_symbol, "_.*"))) %>%
    replace_na(list(condition = "standard")) %>%
    mutate(condition = str_remove(condition, "_(A|B|C)_S\\d+_L001")) %>%
    mutate(tmp_m_symbol = case_when(
        mahendrawada_symbol == "MED15" ~ "GAL11",
        mahendrawada_symbol == "YNR063W" ~ "PUL4",
        .default = mahendrawada_symbol
    )) %>%
    left_join(
        select(genetable, locus_tag, symbol) %>%
            dplyr::rename(tmp_m_symbol = symbol,
                          regulator_locus_tag = locus_tag)) %>%
    dplyr::rename(regulator_symbol = tmp_m_symbol) %>%
    mutate(regulator_locus_tag = case_when(
        regulator_symbol == "FREEMNASE" ~ "FREEMNASE",
        .default = regulator_locus_tag)) %>%
    dplyr::rename(sra_accession = sample)

parsed_chec_meta_with_ids = parsed_chec_meta %>%
    filter(regulator_symbol != "FREEMNASE") %>%
    mutate(replicate = factor(replicate, levels = c("A", "B", "C"))) %>%
    arrange(regulator_locus_tag, condition, replicate) %>%
    group_by(regulator_locus_tag, condition) %>%
    mutate(sample_id = cur_group_id()) %>%
    ungroup() %>%
    mutate(replicate = as.character(replicate))

# arrow::write_parquet(parsed_chec_meta_with_ids, "~/code/hf/mahendrawada_2025/chec_genome_map_meta.parquet")

annotated_feature_meta = parsed_chec_meta_with_ids %>%
    select(regulator_locus_tag, regulator_symbol, condition, sample_id) %>%
    distinct()

freemnase_meta = parsed_chec_meta %>%
    filter(regulator_symbol == "FREEMNASE") %>%
    mutate(replicate = factor(replicate, levels = c("A", "B", "C"))) %>%
    arrange(regulator_locus_tag, condition, replicate) %>%
    group_by(regulator_locus_tag, condition) %>%
    mutate(sample_id = cur_group_id() + max(parsed_chec_meta_with_ids$sample_id)) %>%
    ungroup() %>%
    mutate(replicate = as.character(replicate)) %>%
    select(sra_accession, replicate, condition)

# arrow::write_parquet(freemnase_meta, "~/code/hf/mahendrawada_2025/chec_genome_map_control_meta.parquet")

mahendrawada_genome_map_bed = list.files("~/code/hf/mahendrawada_2025/chec_genome_map_bed", ".bed$")

mahendrawada_genome_map = map(
    mahendrawada_genome_map_bed,
    ~read_tsv(file.path("~/code/hf/mahendrawada_2025/chec_genome_map_bed", .),
              col_names = c("chr", "start", "end", "name", "score", "strand"))
)

names(mahendrawada_genome_map) = str_remove(mahendrawada_genome_map_bed, "_REP1.mLb.mkD.sorted_5p.bed")

mahendrawada_genome_map_df = bind_rows(
    mahendrawada_genome_map,
    .id = 'sra_accession')

rm(mahendrawada_genome_map)
gc()

# arrow::write_dataset(
#     mahendrawada_genome_map_df,
#     path = "/home/chase/code/hf/mahendrawada_2025/chec_genome_map",
#     format = "parquet",
#     partitioning = c("sra_accession"),
#     existing_data_behavior = "overwrite",
#     compression = "zstd",
#     write_statistics = TRUE,
#     use_dictionary = c(
#         chr = TRUE))