fix problem with reading parquet files using gh api

R/dpkg_release_gh.R

@@ -48,7 +48,7 @@ dpkg_gh_release <- function(x, draft = TRUE) {
     httr2::req_body_file(written_path) |>
     httr2::req_perform()
 
-  message("created draft release at: ", draft_release_details$html_url)
+  message("created (draft) release at: ", draft_release_details$html_url)
   return(invisible(draft_release_details$html_url))
 }
 
@@ -79,6 +79,7 @@ get_gh_token <- function() {
 
 # stow a github release asset created with dpkg_github_release()
 # @examples
+# dpkg_gh_release(as_dpkg(mtcars))
 # stow_gh_release("cole-brokamp", "dpkg", "mtcars-v0.0.0.9000")
 # stow_gh_release("cole-brokamp", "dpkg", "mtcars-v0.0.0.9000", force = TRUE)
 stow_gh_release <- function(owner, repo, dpkg, overwrite = FALSE) {
@@ -110,13 +111,17 @@ stow_gh_release <- function(owner, repo, dpkg, overwrite = FALSE) {
 
   the_asset <- the_assets[[which(vapply(the_assets, \(.) .$name == paste0(dpkg, ".parquet"), logical(1)))]]
 
-  httr2::request(glue::glue("https://api.github.com/repos/{owner}/{repo}/releases/assets/{the_asset$id}")) |>
-    httr2::req_headers(
-      Accept = "application/vnd.github+json",
-      Authorization = glue::glue("Bearer {get_gh_token()}"),
-      `X-GitHub-Api-Version` = "2022-11-28",
-      .redact = "Authorization"
-    ) |>
-    httr2::req_perform(path = stow_path(dpkg_filename))
+  ## # why does getting file this way break the header of the parquet file??
+  ## httr2::request(glue::glue("https://api.github.com/repos/{owner}/{repo}/releases/assets/{the_asset$id}")) |>
+  ##   httr2::req_headers(
+  ##     Accept = "application/vnd.github+json",
+  ##     Authorization = glue::glue("Bearer {get_gh_token()}"),
+  ##     `X-GitHub-Api-Version` = "2022-11-28",
+  ##     .redact = "Authorization"
+  ##   ) |>
+  ##   httr2::req_perform(path = stow_path(dpkg_filename))
+
+  stow_url(the_asset$browser_download_url)
+
   return(stow_path(dpkg_filename))
 }

R/helpers.R

@@ -1,5 +1,6 @@
 .onLoad <- function(...) {
   S7::methods_register()
+  fs::dir_create(stow_path())
 }
 
 # enable usage of <S7_object>@name in package code

R/stow.R

@@ -22,16 +22,19 @@
 #' with `http://`, `https://`, or `gh://`
 #' @param overwrite logical; re-download the remote file even though
 #' a local file with the same name exists?
+#' @returns path to the stowed file
 #' @export
 #' @examples
 #' # get by using URL
 #' stow("https://github.com/geomarker-io/appc/releases/download/v0.1.0/nei_2020.rds")
 #'
 #' # will be faster (even in later R sessions) next time
-#' stow("https://github.com/geomarker-io/appc/releases/download/v0.1.0/nei_2020.rds")
+#' stow("https://github.com/geomarker-io/appc/releases/download/v0.1.0/nei_2020.rds") |>
+#'   readRDS()
 #'
 #' # get a data package parquet file created with dpkg_gh_release()
-#' stow("gh://cole-brokamp/dpkg/mtcars-v0.0.0.9000")
+#' stow("gh://cole-brokamp/dpkg/cagis_parcels-v0.1.0") |>
+#'   read_dpkg()
 stow <- function(uri, overwrite = FALSE) {
   if (grepl("^https?://", uri)) {
     out <- stow_url(url = uri, overwrite = overwrite)
@@ -63,7 +66,6 @@ stow <- function(uri, overwrite = FALSE) {
 #' stow_url("https://github.com/geomarker-io/appc/releases/download/v0.1.0/nei_2020.rds")
 stow_url <- function(url, overwrite = FALSE) {
   if (!grepl("^https?://", url)) rlang::abort("x must start with `http://` or `https://`")
-  fs::dir_create(stow_path())
   dest_path <- stow_path(fs::path_file(url))
   httr2::req_perform(httr2::request(url), path = dest_path)
   return(dest_path)

inst/README.md

@@ -0,0 +1,17 @@
+# CAGIS Open Data Parcels
+
+## About
+
+A tabular data resource derived from the Hamilton County, OH Auditor data distributed through [CAGIS Open Data](https://cagismaps.hamilton-co.org/cagisportal/mapdata/download) of parcel-level characteristics collected by Hamilton County.
+
+## Sources
+
+- `parcel_example.R`
+- <https://cagismaps.hamilton-co.org/cagisportal/mapdata/download>
+
+## Details
+
+Parcels were excluded if they were missing an identifier, missing an address number, missing an address street name, or were not considered to be residential.
+
+Because 'second line' address components (e.g., 'Unit 2B') are not captured, a single address can refer to multiple parcels in the case of condos or otherwise shared building ownership.
+Large apartment complexes often use multiple mailing addresses that are not the same as the parcel address(es).

inst/parcel_example.R

@@ -1,21 +1,7 @@
-library(lbl)
+library(dpkg)
 library(dplyr, warn.conflicts = FALSE)
 library(sf)
 
-md <-
-  list(
-    name = "cagis_parcels",
-    title = "CAGIS Open Data Parcels",
-    version = "0.1.0",
-    description = paste(
-        "A tabular data resource derived from the Hamilton County, OH Auditor data distributed through [CAGIS Open Data](https://cagismaps.hamilton-co.org/cagisportal/mapdata/download) of parcel-level characteristics collected by Hamilton County.",
-        "Briefly, parcels were excluded if they were missing an identifier, missing an address number, missing an address street name, or were not considered to be residential.",
-        "Because 'second line' address components (e.g., 'Unit 2B') are not captured, a single address can refer to multiple parcels in the case of condos or otherwise shared building ownership.",
-        "Large apartment complexes often use multiple mailing addresses that are not the same as the parcel address(es).",
-        .sep = "\n"
-      )
-  )
-
 # download CAGIS OpenData Auditor geodatabase
 the_gdb <- fs::path(tools::R_user_dir("lbl_example", "cache"), "Parcels2024.gdb")
 if (!fs::file_exists(the_gdb)) {
@@ -85,27 +71,13 @@ d <-
     sep = " ", na.rm = TRUE, remove = TRUE
   )
 
-# add metadata and updated labels
-out <-
-  d |>
-  as_lbl_tbl(!!!md) |>
-  update_labels(
-    parcel_address = "Derived by pasting `parcel_addr_{number, street, suffix}` together",
-    parcel_id = "Uniquely identifies parcels of land (i.e., auditor parcel number)",
-    land_use = "Parcel land usage code; distinct from city land use codes"
-  )
-
-## login using profile sso account
-system2("aws", c("sso", "login", "--profile", "geomarker-io"))
-Sys.setenv("AWS_PROFILE" = "geomarker-io")
-
-codec_board <-
-  pins::board_s3(
-    bucket = "io.geomarker.codec",
-    versioned = FALSE,
-    prefix = "data/",
-    profile = "geomarker-io",
-    cache = tools::R_user_dir("io.geomarker.codec.data", "cache")
+d_dpkg <- d |>
+  as_dpkg(
+    name = "cagis_parcels",
+    title = "CAGIS Open Data Parcels",
+    version = "0.1.0",
+    homepage = "https://github.com/geomarker-io/parcel",
+    description = paste(readLines(fs::path_package("dpkg", "README.md")), collapse = "\n")
   )
 
-pins::pin_write(codec_board, out, name = glue::glue("{attr(out, 'name')}_v{attr(out, 'version')}"), type = "rds")
+## dpkg_gh_release(d_dpkg, draft = FALSE)