A small, very simple library with bits and pieces of Github's GraphQL and REST API, in one place for my own convenience.
You will need a Github access token with repo
permissions. This is one way to provide that value:
(def token (System/getenv "GITHUB_ACCESS_TOKEN"))
(require '[eamonnsullivan.github-api-lib.core :as core])
;; make your own graphql query
(def get-repo-id
"query getRepoId ($owner: String!, $name: String!) {
repository(owner:$owner, name:$name) {
id
}
}
")
(core/make-graphql-post
token
get-repo-id
{:owner "eamonnsullivan" :name "github-api-lib"})
(require '[eamonnsullivan.github-api-lib.repos :as repos])
(repos/get-repo-id token "https://github.com/eamonnsullivan/github-api-lib")
;; "MDEwOlJlcG9zaXRvcnkzMDYxMjYwNDY="
(repos/get-repo-topics token "eamonnsullivan/github-api-lib")
;; ["clojure" "clojars" "github-graphql"]
(require '[eamonnsullivan.github-api-lib.pull-requests :as pr])
All of these methods return a map of information about the new or updated pull request or comment, such as the :body
(in markdown), :title
, :permalink
or whether the pull request :isDraft
or :mergeable
.
(def options {:title "A title for the pull request"
:body "The body of the pull request"
:base "main or master, usually"
:branch "the name of the branch you want to merge"
:draft true})
(def new-pr-url (:permalink (pr/create-pull-request
token
"https://github.com/eamonnsullivan/github-pr-lib" options)))
The :title
, :base
and :branch
are mandatory. You can omit the :body
, and :draft
defaults to true.
(def updated {:title "A new title"
:body "A new body"})
(pr/update-pull-request token new-pr-url updated)
(pr/mark-ready-for-review token new-pr-url)
Only handles issue comments on pull requests at the moment. The body text can use Github-style markdown.
;; returns the permalink for the comment
(def comment-link (pr/add-pull-request-comment token new-pr-url "Another comment."))
(pr/edit-pull-request-comment token comment-link
"The new body for the comment, with *some markdown* and `stuff`.")
(pr/close-pull-request token new-pr-url)
(pr/reopen-pull-request token new-pr-url)
;; All of these fields are optional. The merge-method will default to "SQUASH".
;; The merge will fail if the pull-request's URL can't be found, if the pull
;; request's head reference is out-of-date or if there are conflicts.
(def merge-options {:title "A title or headline for the commit."
:body "The commit message body."
:mergeMethod "MERGE" or "REBASE" or "SQUASH"
:authorEmail "[email protected]"})
(pr/merge-pull-request token new-pr-url merge-options)
;; Various bits of information, such as whether it is mergeable or a draft.
(pr/get-pull-request-info token new-pr-url)
(require '[eamonnsullivan.github-api-lib.repo-search :as rs])
(def result (rs/get-repos token "my-org" ["topic1" "topic2"]))
(spit "repos.edn" result)
The EDN returned will contain basic information about the repos found. For example:
[{:name "project1",
:description
"A description",
:url "https://github.com/my-org/project1",
:sshUrl "[email protected]:my-org/project1.git",
:updatedAt "2020-04-09T11:01:55Z",
:languages ["Javascript" "Python" "HTML"]}
{:name "project2",
:description "A description for project2",
:url "https://github.com/my-org/project2",
:sshUrl "[email protected]:my-org/project2.git",
:updatedAt "2020-04-09T11:02:28Z",
:languages ["Clojure" "ClojureScript"]}
]
Retrieve information about a file in a repository, on a particular branch. You can use "HEAD" for the branch to retrieve a file from the default branch. The information returns includes :byteSize
and :text
.
(require '[eamonnsullivan.github-api-lib.files :as files])
(files/get-file token "eamonnsullivan" "github-api-lib" "HEAD" "README.md")
{:commitResourcePath
"/eamonnsullivan/github-api-lib/commit/0805f4b95f5e01275e5962e0f8ed23def5129419",
:byteSize 4296,
:filepath "README.md",
:abbreviatedOid "0805f4b",
:isBinary false,
:oid "0805f4b95f5e01275e5962e0f8ed23def5129419",
:commitUrl
"https://github.com/eamonnsullivan/github-api-lib/commit/...",
:isTruncated false,
:text
"# github-api-lib\n\nA small, very simple..."}
You can also try several files and the first one found is returned.
(files/get-first-file token "eamonnsullivan" "github-api-lib" "HEAD"
["build.sbt" ".nvmrc" "deps.edn" "project.edn"])
{:commitResourcePath
"/eamonnsullivan/github-api-lib/commit/74c3092ef552681a7fa5c1a96b3a11479b4f0a28",
:byteSize 1257,
:filepath "deps.edn",
:abbreviatedOid "74c3092",
:isBinary false,
:oid "74c3092ef552681a7fa5c1a96b3a11479b4f0a28",
:commitUrl
"https://github.com/eamonnsullivan/github-api-lib/commit/...",
:isTruncated false,
:text
"{:paths [\"src\" \"resources\"]\n :deps ..."}
The library has a helper function to retrieve all of the pages of a search as a flattened, realised sequence.
See the doc string for more details, but it is normally called with:
- a function to retrieve one of the pages
- a predicate that returns truthy if there are results on that page
- a function to extract the values you want from each page
- A function to retrieve the cursor for the next page.
It obviously would be a poor choice for a very large result set, but it can be convenient for a known limited set. In this example, the function is used to get all of the topics on a given repository.
(require '[eamonnsullivan.github-api-lib.core :as core]
'[eamonnsullivan.github-api-lib.repos :as repos])
(defn get-all-topics [token repo-url page-size]
(let [id (repos/get-repo-id token repo-url)
get-page (partial repos/get-page-of-topics token id page-size)
results? (fn [page] (some? (repos/get-topics page)))
get-next (fn [ret] (if (-> ret :data :node :repsitoryTopics :pageInfo :hasNextPage)
(-> ret :data :node :repositoryTopics :pageInfo :endCursor)
nil))]
(core/get-all-pages get-page results? repos/get-topics get-next)))
(get-all-topics token "https://github.com/eamonnsullivan/github-api-lib" 10)
;; ["clojure" "clojars" "github-graphql"]
To run the project's tests:
$ clojure -M:test:runner
To check test coverage:
$ clojure -M:test:coverage
To build a deployable jar of this library:
$ clojure -Spom # to update any dependencies
$ clojure -M:jar
To install the library locally:
$ clojure -M:install
To deploy it to Clojars -- needs CLOJARS_USERNAME
and CLOJARS_PASSWORD
environment variables or Maven settings:
$ clojure --M:deploy
Copyright © 2020 Eamonn Sullivan
Distributed under the Eclipse Public License either version 1.0 or (at your option) any later version.