Encourage Imports not Depends

library(tools)
library(data.table)

# 1. Fetch CRAN metadata and reverse dependencies (Depends: data.table)
cran_db = as.data.table(CRAN_package_db())
cran_deps = package_dependencies("data.table", db = as.matrix(cran_db), reverse = TRUE, which = "Depends")$data.table
cran_sub = cran_db[Package %in% cran_deps]

# 2. Fetch Bioconductor metadata and reverse dependencies (Depends: data.table)
bioc_views_url = "https://bioconductor.org/packages/release/bioc/VIEWS"
tf = tempfile()
download.file(bioc_views_url, tf, quiet = TRUE)
bioc_views = as.data.table(read.dcf(tf))
bioc_deps = package_dependencies("data.table", db = as.matrix(bioc_views), reverse = TRUE, which = "Depends")$data.table
bioc_sub = bioc_views[Package %in% bioc_deps]

# 3. Harvest existing tracked links dynamically from GitHub issue #3076 body
issue_body = tryCatch({
  res = httr::GET("https://api.github.com/repos/Rdatatable/data.table/issues/3076", httr::user_agent("Rdatatable"))
  httr::content(res, "parsed")$body
}, error = function(e) "")

known_links = list()
fixed_links = list()
lines = strsplit(issue_body, "\r?\n")[[1]]

for (l in lines) {
  # Parse table rows: | # | [pkg](repo) | [Host](host_url) | Last Upload | [✅](issue) | [✅](pr) | [date](fix) |
  if (grepl(r"{^\|.*\|.*\|.*\|}", l)) {
    parts = strsplit(l, "\\|")[[1]]
    if (length(parts) > 0 && parts[1] == "") parts = parts[-1]
    parts = trimws(parts)
    if (length(parts) >= 4 && !any(grepl("Last Upload|:---", parts))) {
      pkg_col = if (grepl("^[0-9]+$", parts[1])) 2 else 1
      pkg = sub(r"{^\[([^]]+)\].*}", "\\1", parts[pkg_col])
      
      issue_col = pkg_col + 3
      pr_col    = pkg_col + 4
      fix_col   = pkg_col + 5
      
      if (length(parts) >= issue_col && grepl(r"{https?://}", parts[issue_col])) {
        m_iss = regmatches(parts[issue_col], regexec(r"{\((https?://[^)]+)\)}", parts[issue_col]))[[1]]
        if (length(m_iss) > 1) {
          known_links[[length(known_links) + 1]] = data.table(Package = pkg, Type = "Issue", URL = m_iss[2])
        }
      }
      if (length(parts) >= pr_col && grepl(r"{https?://}", parts[pr_col])) {
        m_pr = regmatches(parts[pr_col], regexec(r"{\((https?://[^)]+)\)}", parts[pr_col]))[[1]]
        if (length(m_pr) > 1) {
          known_links[[length(known_links) + 1]] = data.table(Package = pkg, Type = "PR", URL = m_pr[2])
        }
      }
      if (length(parts) >= fix_col && grepl(r"{https?://}", parts[fix_col])) {
        fixed_links[[length(fixed_links) + 1]] = data.table(Package = pkg, Fixed_In_Dev = parts[fix_col])
      }
    }
  }
}

links_dt = if (length(known_links) > 0) {
  dcast(unique(rbindlist(known_links)), Package ~ Type, value.var = "URL", fun.aggregate = function(x) x[1])
} else {
  data.table(Package = character(), Issue = character(), PR = character())
}

fixed_dt = if (length(fixed_links) > 0) unique(rbindlist(fixed_links)) else data.table(Package = character(), Fixed_In_Dev = character())

# Helper: extract source forge repository (GitHub/GitLab) from package fields or links
extract_repo <- function(pkg, url_fields, issue_url, pr_url) {
  ref_url = na.omit(c(issue_url, pr_url))
  if (length(ref_url) > 0) {
    m = regmatches(ref_url[1], regexec(r"{(https?://[^/]+/[^/]+/[^/#?]+)}", ref_url[1]))[[1]]
    if (length(m) > 1) {
      repo = sub(r"{/(issues|pull|tree|blob|commit)/?.*$}", "", m[2])
      repo = sub(r"{/-/issues/?.*$}", "", repo)
      return(repo)
    }
  }
  urls = unlist(strsplit(url_fields[!is.na(url_fields)], "[,\n\t ]+"))
  urls = urls[urls != ""]
  repo_urls = grep("git(hub|lab)", urls, value = TRUE)
  repo_urls = repo_urls[!grepl("\\.github\\.io", repo_urls)]
  if (length(repo_urls) > 0) {
    m = regmatches(repo_urls[1], regexec(r"{(https?://[^/]+/[^/]+/[^/#?]+)}", repo_urls[1]))[[1]]
    if (length(m) > 1) {
      repo = sub(r"{\.git$}", "", m[2])
      repo = sub(r"{/(issues|pull)/?.*$}", "", repo)
      repo = sub(r"{/-/issues/?.*$}", "", repo)
      return(repo)
    }
  }
  NA_character_
}

# CRAN data
cran_dt = data.table(
  Package_Name = cran_sub$Package,
  Host_Name = "CRAN",
  Host_URL = sprintf("https://cran.r-project.org/package=%s", cran_sub$Package),
  Last_Upload = as.character(as.Date(as.POSIXct(cran_sub[["Date/Publication"]]))),
  URL_Fields = paste(cran_sub$URL, cran_sub$BugReports, sep = ", ")
)

# Bioconductor data
bioc_dt = data.table(
  Package_Name = bioc_sub$Package,
  Host_Name = "Bioconductor",
  Host_URL = sprintf("https://bioconductor.org/packages/%s", bioc_sub$Package),
  Last_Upload = bioc_sub[["Date/Publication"]],
  URL_Fields = paste(bioc_sub$URL, bioc_sub$BugReports, bioc_sub$git_url, sep = ", ")
)

all_pkgs = rbind(cran_dt, bioc_dt)
all_pkgs = merge(all_pkgs, links_dt, by.x = "Package_Name", by.y = "Package", all.x = TRUE)
all_pkgs = merge(all_pkgs, fixed_dt, by.x = "Package_Name", by.y = "Package", all.x = TRUE)

all_pkgs[, Repo := mapply(extract_repo, Package_Name, URL_Fields, Issue, PR)]
all_pkgs[, Package := ifelse(!is.na(Repo) & Repo != "", sprintf("[%s](%s)", Package_Name, Repo), Package_Name)]
all_pkgs[, Host := sprintf("[%s](%s)", Host_Name, Host_URL)]
all_pkgs[, `Issue?` := ifelse(!is.na(Issue) & Issue != "", sprintf("[✅](%s)", Issue), "")]
all_pkgs[, `PR?` := ifelse(!is.na(PR) & PR != "", sprintf("[✅](%s)", PR), "")]
all_pkgs[, `Fixed in dev` := ifelse(!is.na(Fixed_In_Dev), Fixed_In_Dev, "")]

# Sort by Host (CRAN then Bioconductor) and descending Last_Upload
all_pkgs[, Host_Order := factor(Host_Name, levels = c("CRAN", "Bioconductor"))]
setorder(all_pkgs, Host_Order, -Last_Upload)

# Leftmost column numbered 1...n with empty header
all_pkgs[, ` ` := .I]

out_dt = all_pkgs[, .(` `, Package, Host, `Last Upload` = Last_Upload, `Issue?`, `PR?`, `Fixed in dev`)]
writeLines(knitr::kable(out_dt, format = "markdown", align = c("r", "l", "l", "l", "c", "c", "l")))