Merge pull request #499 from r-world-devs/492-fix-get_repos

492 get_repos() for users
r-world-devs · Oct 11, 2024 · 18a2574 · 18a2574
2 parents 8b751aa + 3f6e9ef
commit 18a2574
Show file tree

Hide file tree

Showing 24 changed files with 978 additions and 197 deletions.
diff --git a/.github/workflows/test-coverage.yaml b/.github/workflows/test-coverage.yaml
@@ -1,5 +1,3 @@
-# Workflow derived from https://github.com/r-lib/actions/tree/v2/examples
-# Need help debugging build failures? Start at https://github.com/r-lib/actions#where-to-find-help
 on:
   push:
     branches: [master]
@@ -14,6 +12,7 @@ jobs:
     env:
       GITHUB_PAT: ${{ secrets.GITHUB_TOKEN }}
       GITLAB_PAT_PUBLIC: ${{ secrets.GITLAB_PAT}}
+      CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}
       USE_RENV: "FALSE"
 
     steps:
@@ -28,7 +27,7 @@ jobs:
           extra-packages: any::covr
           needs: coverage
 
-      - name: Test coverage
+      - name: Code coverage
         run: |
           covr::codecov(
             quiet = FALSE,

diff --git a/DESCRIPTION b/DESCRIPTION
@@ -1,6 +1,6 @@
 Package: GitStats
 Title: Get Statistics from GitHub and GitLab
-Version: 2.1.0.9000
+Version: 2.1.0.9001
 Authors@R: c(
     person(given = "Maciej", family = "Banas", email = "[email protected]", role = c("aut", "cre")),
     person(given = "Kamil", family = "Koziej", email = "[email protected]", role = "aut"),

diff --git a/NEWS.md b/NEWS.md
@@ -1,5 +1,6 @@
 # GitStats (development version)
 
+- Added possibility to get repositories for individual users with `get_repos()` ([#492](https://github.com/r-world-devs/GitStats/issues/492)). Earlier this was only possible for GitHub organizations and GitLab groups.
 - Fixed getting large search responses for GitHub ([#491](https://github.com/r-world-devs/GitStats/issues/491)).
 
 # GitStats 2.1.0

diff --git a/R/EngineGraphQLGitHub.R b/R/EngineGraphQLGitHub.R
@@ -41,16 +41,22 @@ EngineGraphQLGitHub <- R6::R6Class(
     },
 
     # Pull all repositories from organization
-    get_repos_from_org = function(org = NULL) {
+    get_repos_from_org = function(org = NULL,
+                                  type = c("organization", "user")) {
       full_repos_list <- list()
       next_page <- TRUE
       repo_cursor <- ""
       while (next_page) {
         repos_response <- private$get_repos_page(
-          org = org,
+          login = org,
+          type  = type,
           repo_cursor = repo_cursor
         )
-        repositories <- repos_response$data$repositoryOwner$repositories
+        repositories <- if (type == "organization") {
+          repos_response$data$repositoryOwner$repositories
+        } else {
+          repos_response$data$user$repositories
+        }
         repos_list <- repositories$nodes
         next_page <- repositories$pageInfo$hasNextPage
         if (is.null(next_page)) next_page <- FALSE
@@ -87,6 +93,7 @@ EngineGraphQLGitHub <- R6::R6Class(
 
     # Pull all given files from all repositories of an organization.
     get_files_from_org = function(org,
+                                  type,
                                   repos,
                                   file_paths,
                                   host_files_structure,
@@ -95,6 +102,7 @@ EngineGraphQLGitHub <- R6::R6Class(
                                   progress = TRUE) {
       repo_data <- private$get_repos_data(
         org = org,
+        type = type,
         repos = repos
       )
       repositories <- repo_data[["repositories"]]
@@ -117,13 +125,15 @@ EngineGraphQLGitHub <- R6::R6Class(
 
     # Pull all files from all repositories of an organization.
     get_files_structure_from_org = function(org,
+                                            type,
                                             repos,
                                             pattern  = NULL,
                                             depth    = Inf,
                                             verbose  = FALSE,
                                             progress = TRUE) {
       repo_data <- private$get_repos_data(
         org = org,
+        type = type,
         repos = repos
       )
       repositories <- repo_data[["repositories"]]
@@ -162,14 +172,20 @@ EngineGraphQLGitHub <- R6::R6Class(
   private = list(
 
     # Wrapper over building GraphQL query and response.
-    get_repos_page = function(org = NULL,
+    get_repos_page = function(login = NULL,
+                              type = c("organization", "user"),
                               repo_cursor = "") {
-      repos_query <- self$gql_query$repos_by_org(
-        repo_cursor = repo_cursor
-      )
+      repos_query <- if (type == "organization") {
+        self$gql_query$repos_by_org()
+      } else {
+        self$gql_query$repos_by_user()
+      }
       response <- self$gql_response(
         gql_query = repos_query,
-        vars = list("org" = org)
+        vars = list(
+          "login" = login,
+          "repoCursor" = repo_cursor
+        )
       )
       return(response)
     },
@@ -225,9 +241,10 @@ EngineGraphQLGitHub <- R6::R6Class(
       return(response)
     },
 
-    get_repos_data = function(org, repos = NULL) {
+    get_repos_data = function(org, type, repos = NULL) {
       repos_list <- self$get_repos_from_org(
-        org = org
+        org = org,
+        type = type
       )
       if (!is.null(repos)) {
         repos_list <- purrr::keep(repos_list, ~ .$repo_name %in% repos)

diff --git a/R/EngineGraphQLGitLab.R b/R/EngineGraphQLGitLab.R
@@ -46,19 +46,22 @@ EngineGraphQLGitLab <- R6::R6Class(
     },
 
     # Iterator over pulling pages of repositories.
-    get_repos_from_org = function(org = NULL) {
+    get_repos_from_org = function(org  = NULL,
+                                  type = c("organization", "user")) {
       full_repos_list <- list()
       next_page <- TRUE
       repo_cursor <- ""
       while (next_page) {
         repos_response <- private$get_repos_page(
           org = org,
+          type = type,
           repo_cursor = repo_cursor
         )
-        if (length(repos_response$data$group) == 0) {
-          cli::cli_abort("Empty")
+        core_response <- if (type == "organization") {
+          repos_response$data$group$projects
+        } else {
+          repos_response$data$projects
         }
-        core_response <- repos_response$data$group$projects
         repos_list <- core_response$edges
         next_page <- core_response$pageInfo$hasNextPage
         if (is.null(next_page)) next_page <- FALSE
@@ -81,6 +84,7 @@ EngineGraphQLGitLab <- R6::R6Class(
     # pulled files_structure. In such a case GitStats will switch from this function
     # to iterator over repositories (multiple queries), as it is done for GitHub.
     get_files_from_org = function(org,
+                                  type,
                                   repos,
                                   file_paths,
                                   host_files_structure,
@@ -100,73 +104,87 @@ EngineGraphQLGitLab <- R6::R6Class(
       } else if (is.null(host_files_structure) && only_text_files) {
         file_paths <- file_paths[!grepl(non_text_files_pattern, file_paths)]
       }
-      while (next_page) {
-        files_query <- self$gql_query$files_by_org(
-          end_cursor = end_cursor
-        )
-        files_response <- tryCatch(
-          {
-            self$gql_response(
-              gql_query = files_query,
-              vars = list(
-                "org" = org,
-                "file_paths" = file_paths
+      if (type == "organization") {
+        while (next_page) {
+          files_query <- self$gql_query$files_by_org(
+            end_cursor = end_cursor
+          )
+          files_response <- tryCatch(
+            {
+              self$gql_response(
+                gql_query = files_query,
+                vars = list(
+                  "org" = org,
+                  "file_paths" = file_paths
+                )
               )
-            )
-          },
-          error = function(e) {
-            list()
-          }
-        )
-        if (private$is_query_error(files_response)) {
-          if (verbose) {
-            purrr::walk(files_response$errors, ~ cli::cli_alert_warning(.))
-          }
-          if (private$is_complexity_error(files_response)) {
+            },
+            error = function(e) {
+              list()
+            }
+          )
+          if (private$is_query_error(files_response)) {
             if (verbose) {
-              cli::cli_alert_info(
-                cli::col_br_cyan("I will switch to pulling files per repository.")
+              purrr::walk(files_response$errors, ~ cli::cli_alert_warning(.))
+            }
+            if (private$is_complexity_error(files_response)) {
+              if (verbose) {
+                cli::cli_alert_info(
+                  cli::col_br_cyan("I will switch to pulling files per repository.")
+                )
+              }
+              full_files_list <- self$get_files_from_org_per_repo(
+                org = org,
+                type = type,
+                repos = repos,
+                file_paths = file_paths,
+                host_files_structure = host_files_structure,
+                only_text_files = only_text_files,
+                verbose = verbose,
+                progress = progress
               )
+              return(full_files_list)
             }
-            full_files_list <- self$get_files_from_org_per_repo(
-              org = org,
-              repos = repos,
-              file_paths = file_paths,
-              host_files_structure = host_files_structure,
-              only_text_files = only_text_files,
-              verbose = verbose,
-              progress = progress
-            )
-            return(full_files_list)
           }
+          if (length(files_response$data$group) == 0 && verbose) {
+            cli::cli_alert_danger("Empty response.")
+          }
+          projects <- files_response$data$group$projects
+          files_list <- purrr::map(projects$edges, function(edge) {
+            edge$node
+          }) %>%
+            purrr::discard(~ length(.$repository$blobs$nodes) == 0)
+          if (is.null(files_list)) files_list <- list()
+          if (length(files_list) > 0) {
+            next_page <- files_response$pageInfo$hasNextPage
+          } else {
+            next_page <- FALSE
+          }
+          if (is.null(next_page)) next_page <- FALSE
+          if (next_page) {
+            end_cursor <- files_response$pageInfo$endCursor
+          } else {
+            end_cursor <- ""
+          }
+          full_files_list <- append(full_files_list, files_list)
         }
-        if (length(files_response$data$group) == 0 && verbose) {
-          cli::cli_alert_danger("Empty response.")
-        }
-        projects <- files_response$data$group$projects
-        files_list <- purrr::map(projects$edges, function(edge) {
-          edge$node
-        }) %>%
-          purrr::discard(~ length(.$repository$blobs$nodes) == 0)
-        if (is.null(files_list)) files_list <- list()
-        if (length(files_list) > 0) {
-          next_page <- files_response$pageInfo$hasNextPage
-        } else {
-          next_page <- FALSE
-        }
-        if (is.null(next_page)) next_page <- FALSE
-        if (next_page) {
-          end_cursor <- files_response$pageInfo$endCursor
-        } else {
-          end_cursor <- ""
+        if (!is.null(repos)) {
+          full_files_list <- purrr::keep(full_files_list, function(project) {
+            repo_name <- private$get_repo_name_from_url(project$webUrl)
+            repo_name %in% repos
+          })
         }
-        full_files_list <- append(full_files_list, files_list)
-      }
-      if (!is.null(repos)) {
-        full_files_list <- purrr::keep(full_files_list, function(project) {
-          repo_name <- private$get_repo_name_from_url(project$webUrl)
-          repo_name %in% repos
-        })
+      } else {
+        full_files_list <- self$get_files_from_org_per_repo(
+          org = org,
+          type = type,
+          repos = repos,
+          file_paths = file_paths,
+          host_files_structure = host_files_structure,
+          only_text_files = only_text_files,
+          verbose = verbose,
+          progress = progress
+        )
       }
       return(full_files_list)
     },
@@ -175,6 +193,7 @@ EngineGraphQLGitLab <- R6::R6Class(
     # one query way applied with get_files_from_org() fails due to its complexity.
     # For more info see docs above.
     get_files_from_org_per_repo = function(org,
+                                           type,
                                            repos,
                                            file_paths = NULL,
                                            host_files_structure = NULL,
@@ -184,6 +203,7 @@ EngineGraphQLGitLab <- R6::R6Class(
       if (is.null(repos)) {
         repo_data <- private$get_repos_data(
           org = org,
+          type = type,
           repos = repos
         )
         repos <- repo_data[["repositories"]]
@@ -213,13 +233,15 @@ EngineGraphQLGitLab <- R6::R6Class(
       return(org_files_list)
     },
     get_files_structure_from_org = function(org,
+                                            type,
                                             repos,
                                             pattern = NULL,
                                             depth = Inf,
                                             verbose = TRUE,
                                             progress = TRUE) {
       repo_data <- private$get_repos_data(
         org = org,
+        type = type,
         repos = repos
       )
       repositories <- repo_data[["repositories"]]
@@ -259,14 +281,25 @@ EngineGraphQLGitLab <- R6::R6Class(
 
     # Wrapper over building GraphQL query and response.
     get_repos_page = function(org = NULL,
+                              type = "organization",
                               repo_cursor = "") {
-      repos_query <- self$gql_query$repos_by_org(
-        repo_cursor = repo_cursor
-      )
-      response <- self$gql_response(
-        gql_query = repos_query,
-        vars = list("org" = org)
-      )
+      if (type == "organization") {
+        response <- self$gql_response(
+          gql_query = self$gql_query$repos_by_org(),
+          vars = list(
+            "org" = org,
+            "repo_cursor" = repo_cursor
+          )
+        )
+      } else {
+        response <- self$gql_response(
+          gql_query = self$gql_query$repos_by_user(),
+          vars = list(
+            "username" = org,
+            "repo_cursor" = repo_cursor
+          )
+        )
+      }
       return(response)
     },
 
@@ -277,9 +310,10 @@ EngineGraphQLGitLab <- R6::R6Class(
       return(repo_name)
     },
 
-    get_repos_data = function(org, repos = NULL) {
+    get_repos_data = function(org, type, repos = NULL) {
       repos_list <- self$get_repos_from_org(
-        org = org
+        org = org,
+        type = type
       )
       if (!is.null(repos)) {
         repos_list <- purrr::keep(repos_list, ~ .$node$repo_path %in% repos)