graphql query

graphql_reviews.R

library(tidyverse)
library(httr)
library(jsonlite)
library(glue)
library(lubridate)

# JSON friendly version of glue
jlue <- function(..., .envir = parent.frame()) {
  glue(..., .envir = .envir, .open = "<", .close = ">")
}

org_data <- function(org) {
  # TODO: paginate repos and maybe pullRequests?
  query <- jlue('{
      organization(login: "<org>") {
        repositories(first: 50, orderBy: {field: PUSHED_AT, direction: DESC}) {
          nodes {
            name
            pullRequests(first: 25, orderBy: {field: UPDATED_AT, direction: DESC}) {
              nodes {
                number
                author {
                  login
                }
                createdAt
                reviews(first: 10) {
                  nodes {
                    author {
                      login
                    }
                    submittedAt
                  }
                }
              }
            }
          }
        }
      }
  }')

  res <- gh::gh("/graphql", query = query)

  d <- content(res)$data$organization$repositories$nodes

  has_pr <- map_int(d, ~ length(.x$pullRequests$nodes)) > 0
  d1 <- d[has_pr]

  d2 <- map_dfr(d1, function(.x) {
    res <- flatten_df(.x$pullRequests)
    res$repo <- .x$name
    res$author <- map_chr(res$author, pluck)
    res
  })

  has_review <- lengths(d2$reviews) > 0
  d3 <- bind_rows(
    mutate(d2[has_review, ], 
     reviews = map(reviews, 
       ~ tibble(
           reviewer = map_chr(.x, c("author", "login")),
           reviewedAt = map_chr(.x, "submittedAt")))) %>% unnest(),
    select(d2[!has_review, ], -reviews))

  d3
}

d <- bind_rows(
  org_data("r-lib"),
  org_data("tidyverse"),
  org_data("r-dbi"))

prs <- d %>%
  mutate(createdAt = lubridate::as_datetime(d$createdAt)) %>%
  filter(createdAt > today() - ddays(30)) %>%
  select(repo, number, author, reviewer) %>%
  filter(author != reviewer) %>%
  unique()

pr_authors <- prs %>% select(-reviewer) %>% unique() %>% group_by(author)
pr_reviewers <- prs %>% select(-author) %>% unique() %>% group_by(reviewer)

tally(pr_authors, sort = TRUE)
tally(pr_reviewers, sort = TRUE)

pull_requests.Rmd

---
title: " Repository Dashboard"
output: 
  flexdashboard::flex_dashboard:
    orientation: columns
    vertical_layout: fill
---

```{r setup, include=FALSE, cache = TRUE}
library(flexdashboard)
library(tidyverse)
library(httr)
library(jsonlite)
library(glue)
library(lubridate)
library(dplyr)
library(codetools)

# JSON friendly version of glue
jlue <- function(..., .envir = parent.frame()) {
  glue(..., .envir = .envir, .open = "<", .close = ">")
}

org_data <- function(org) {
  # TODO: paginate repos and maybe pullRequests?
  query <- jlue('{
      organization(login: "<org>") {
        repositories(first: 50, orderBy: {field: PUSHED_AT, direction: DESC}) {
          nodes {
            name
            pullRequests(first: 25, orderBy: {field: UPDATED_AT, direction: DESC}) {
              nodes {
                number
                author {
                  login
                }
                createdAt
                reviews(first: 10) {
                  nodes {
                    author {
                      login
                    }
                    submittedAt
                  }
                }
              }
            }
          }
        }
      }
  }')

  res <- gh::gh("POST /graphql", query = query)

  d <- res$data$organization$repositories$nodes

  has_pr <- map_int(d, ~ length(.x$pullRequests$nodes)) > 0
  d1 <- d[has_pr]

  d2 <- map_dfr(d1, function(.x) {
    res <- flatten_df(.x$pullRequests)
    res$repo <- .x$name
    res$author <- map_chr(res$author, pluck)
    res
  })

  has_review <- lengths(d2$reviews) > 0
  get_review_info <- function(x) {
       tibble(
           reviewer = map_chr(x, c("author", "login")),
           reviewedAt = map_chr(x, ~ .$submittedAt %||% NA_character_))
   }
  d3 <- bind_rows(
    mutate(d2[has_review, ], reviews = map(reviews, get_review_info)) %>% unnest(),
    select(d2[!has_review, ], -reviews))

  d3
}

d <- bind_rows(
  org_data("r-lib"),
  org_data("tidyverse"),
  org_data("r-dbi"))

prs <- d %>%
  mutate(createdAt = as_datetime(d$createdAt)) %>%
  filter(createdAt > today() - ddays(30)) %>%
  select(repo, number, author, reviewer) %>%
  filter(author != reviewer) %>%
  unique()

pr_authors <- prs %>% select(-reviewer) %>% unique() %>% group_by(author)
pr_reviewers <- prs %>% select(-author) %>% unique() %>% group_by(reviewer)
```

```{r repo_stats, include = FALSE, cache = TRUE}
parse_issue <- function(x) {
  labels <- map_chr(x$labels$nodes, "name")

  list(p1 = x$p1$totalCount %||% 0, bugs = "bug" %in% labels, features = "feature" %in% labels, unlabeled = !length(labels))
}
parse_repository <- function(x) {
  if (x$open_issues$totalCount == 0) {
    issues <- list(p1 = 0, bugs = FALSE, features = FALSE, unlabeled = FALSE)
  } else {
    issues <- map_dfr(x$open_issues$nodes, parse_issue)
  }
  if (x$open_issues$pageInfo$hasNextPage) {
    message(x$repo, " has more than 100 open issues!")
  }
  #message(x$repo)

  tibble::tibble(
    owner = x$owner$login,
    repo = x$repo,
    prs = x$prs$totalCount,
    watchers = x$watchers$totalCount,
    open_issues = x$open_issues$totalCount,
    p1 = sum(issues$p1),
    bugs = sum(issues$bugs),
    features = sum(issues$features),
    unlabeled = sum(issues$unlabeled))
}

org_summary <- function(org) {
  query <- readChar("summary.json", fs::file_info("summary.json")$size)
  res <- gh::gh("POST /graphql", query = query, variables = list(org = org))

  map_dfr(res$data$organization$repositories$nodes, parse_repository)
}

summary_data <- bind_rows(
  org_summary("tidyverse"),
  org_summary("r-lib"),
  org_summary("r-dbi"))
```
Repository Statistics
=====================================
```{r}
DT::datatable(summary_data)
``` 

Pull Request Review counts
=====================================
Column
-------------------------------------
### PR authors
```{r}
DT::datatable(dplyr::tally(pr_authors, sort = TRUE))
```

Column
-------------------------------------
### PR Reviewers
```{r}
DT::datatable(dplyr::tally(pr_reviewers, sort = TRUE))
```

summary.json

query repoSummary($org:String!) {
  organization(login: $org) {
    repositories(first:100) {
      nodes {
        owner {login},
        repo:name,
        open_issues : issues(first:100, states:OPEN) {
          totalCount
          nodes {
            p1 : reactions(content:THUMBS_UP) {
              totalCount
            }
            labels(first:10) {
              nodes {
                name
              }
            }
          }
          pageInfo {
            endCursor
            hasNextPage
          }
        }
        prs : pullRequests(states:OPEN) {
          totalCount
        }
        watchers : stargazers() {
          totalCount
        }
      }
    }
  }
}