Merge pull request #2 from mlverse/updates

Updates

DESCRIPTION

@@ -1,10 +1,13 @@
 Package: lang
-Title: What the Package Does (One Line, Title Case)
-Version: 0.0.0.9000
+Title: Using an Large Language Model of your choice, it translates the R
+    help documentation to a specific language
+Version: 0.0.0.9001
 Authors@R: 
-    person("First", "Last", , "[email protected]", role = c("aut", "cre"),
-           comment = c(ORCID = "YOUR-ORCID-ID"))
-Description: What the package does (one paragraph).
+    person("Edgar", "Ruiz", , "[email protected]", role = c("aut", "cre"))
+Description: Use an 'LLM' to translate a function's help documentation on-the-fly.
+    It  overrides the `?` and `help()` functions in your R session. If you are 
+    using RStudio or Positron, the translated help page will appear in the usual
+    help pane.
 License: MIT + file LICENSE
 Encoding: UTF-8
 Roxygen: list(markdown = TRUE)

R/lang-help.R

@@ -161,7 +161,7 @@ rd_comment_translate <- function(x, lang) {
       rd_char <- llm_vec_translate(rd_char, lang)
       rd_char <- paste0("# ", rd_char, "\n")
     } else {
-      
+
     }
     rd_char <- gsub("%", "\\\\%", rd_char)
     attributes(rd_char) <- attributes(x)
@@ -203,7 +203,7 @@ rd_extract_text <- function(x, collapse = TRUE) {
     rd_txt <- paste0(rd_txt, collapse = "")
   }
   rd_txt <- gsub("\U2018", "'", rd_txt)
-  rd_txt <- gsub("\U2019", "'", rd_txt)  
+  rd_txt <- gsub("\U2019", "'", rd_txt)
   rd_txt
 }
 

R/lang.R

@@ -1,11 +1,10 @@
 #' @importFrom roxygen2 parse_file roxygenize env_package
+#' @importFrom withr local_tempdir with_dir
 #' @importFrom utils capture.output
 #' @import tools
-#' @import tools
 #' @import rlang
+#' @import glue
 #' @import mall
 #' @import cli
-#' @import glue
-#' @importFrom withr local_tempdir with_dir
 #' @import fs
 NULL

R/process-roxygen.R

@@ -4,11 +4,14 @@
 #' Defaults to 'man-lang'.
 #' @param target_folder Base target folder where the different translations will be
 #' located. Defaults to 'inst/man-lang'
+#' @param r_folder Source of the original R scripts. Only used to see if the
+#' Roxygen documentation is different from what is capture in the `source_folder`
 #' @param pkg_path The path to the package
 #' @returns Multiple Rd files based on the source R scripts
 #' @export
 process_roxygen_folder <- function(
     folder,
+    r_folder = "R",
     source_folder = "man-lang",
     target_folder = "inst/man-lang",
     pkg_path = ".") {
@@ -28,8 +31,10 @@ process_roxygen_folder <- function(
   }
   # Copies content of the translated script to the R folder
   # of the temp copy
+  source_path <- path(pkg_path, source_folder, folder)
+  full_source <- dir_ls(path(source_path), recurse = TRUE, glob = "*.R")
   file_copy(
-    path = dir_ls(path(pkg_path, source_folder, folder)),
+    path = full_source,
     new_path = path(copy_path, "R"),
     overwrite = TRUE
   )
@@ -64,8 +69,14 @@ process_roxygen_folder <- function(
 process_roxygen <- function(
     source_folder = "man-lang",
     target_folder = "inst/man-lang",
+    r_folder = "R",
     pkg_path = ".") {
   sub_folders <- dir_ls(path(pkg_path, source_folder), type = "directory")
+  compare_man_lang(
+    r_folder = r_folder,
+    source_folder = source_folder,
+    pkg_path = pkg_path
+  )
   for (folder in sub_folders) {
     process_roxygen_folder(
       folder = path_file(folder),
@@ -75,3 +86,40 @@ process_roxygen <- function(
     )
   }
 }
+
+compare_man_lang <- function(
+    folder = NULL,
+    r_folder = "R",
+    source_folder = "man-lang",
+    pkg_path = ".") {
+  pkg_dir <- path_abs(pkg_path)
+  r_scripts <- dir_ls(path(pkg_dir, r_folder), glob = "*.R")
+  r_comments <- lapply(r_scripts, roxy_comments)
+  source_path <- path(pkg_path, source_folder)
+  if (!is.null(folder)) {
+    source_path <- path(source_path, folder)
+  }
+  full_source <- dir_ls(path(source_path), recurse = TRUE, glob = "*.R")
+  man_comments <- lapply(full_source, roxy_existing)
+  man_diff <- NULL
+  found_diff <- FALSE
+  for (script in r_scripts) {
+    r_curr <- r_comments[names(r_comments) == script]
+    if (!is.null(r_curr[[1]])) {
+      man_curr <- man_comments[path_file(names(man_comments)) == path_file(script)]
+      is_same <- paste0(r_curr, collapse = "") == paste0(man_curr, collapse = "")
+      if (!is_same) {
+        if (!found_diff) {
+          cli_alert_warning(
+            c(
+              "The following R documentation has changed, ",
+              "translation may need to be revised:"
+            )
+          )
+          found_diff <- TRUE
+        }
+        cli_inform("|- {path_rel(script)} -x-> {path_rel(names(man_curr))}")
+      }
+    }
+  }
+}

R/translate-roxygen.R

@@ -73,9 +73,16 @@ translate_roxygen_imp <- function(path,
     }
     pkg_env <- env_package(pkg_path)
   }
+  current_roxy <- roxy_comments(path)
   dir_create(dir)
   rd_path <- path(dir, path_file(path))
-  cli_inform("[{no}/{of}] {path} --> {rd_path}")
+  if (file_exists(rd_path)) {
+    tr_roxy <- roxy_existing(rd_path)
+    if (paste0(tr_roxy, collapse = " ") == paste0(current_roxy, collapse = " ")) {
+      cli_inform("[{no}/{of}] {path} --> [Skipping, no changes]")
+      return(invisible())
+    }
+  }
   parsed <- parse_file(path, env = pkg_env)
   contents <- NULL
   tg_label <- NULL
@@ -108,7 +115,6 @@ translate_roxygen_imp <- function(path,
         if (tg == "section") {
           raw <- glue("{raw[1]}:\n{raw[2]}")
         }
-        # raw <- gsub("\n", "\n#'", raw)
         if (tg != "title") {
           if (length(raw) != 0 && raw != "") {
             pre_raw <- paste0(tg, name, collapse = " ")
@@ -132,6 +138,39 @@ translate_roxygen_imp <- function(path,
     contents <- c(contents, fn_str)
   }
   if (!is.null(contents)) {
+    cli_inform("[{no}/{of}] {path} --> {rd_path}")
+    contents <- c(
+      contents,
+      "# --- Created by `lang` do not edit by hand ---",
+      current_roxy
+    )
     writeLines(contents, rd_path)
+  } else {
+    cli_inform("[{no}/{of}] {path} --> [Skipping, no content]")
+  }
+}
+
+roxy_comments <- function(x) {
+  script_contents <- readLines(x)
+  roxy_comment <- substr(script_contents, 1, 2) == "#'"
+  just_roxy <- script_contents[roxy_comment]
+  just_roxy <- just_roxy[just_roxy != "#'"]
+
+  if (length(just_roxy) == 0) {
+    return(NULL)
+  } else {
+    just_roxy <- paste0("#-", just_roxy)
+    no_exports <- !any(grepl("#' @export", just_roxy))
+    no_name <- !any(grepl("#' @name", just_roxy))
+    if (no_exports && no_name) {
+      return(NULL)
+    }
   }
+  just_roxy
+}
+
+roxy_existing <- function(x) {
+  script_contents <- readLines(x)
+  roxy_comment <- substr(script_contents, 1, 4) == "#-#'"
+  script_contents[roxy_comment]
 }