GitHub Search ·

This project is made of two components:

1. A Spring Boot powered back-end, responsible for:
   1. Continuously crawling GitHub API endpoints for repository information, and storing it in a central database;
   2. Acting as an API for providing access to the stored data.
2. A Bootstrap-styled and jQuery-powered web user interface, serving as an accessible front for the API.

Running Locally

Prerequisites

Dependency	Version Requirement
Java	17
Maven	3.9
MySQL	8.3
Flyway	10.13
cloc1	2.00
Git1	2.43

Database

Before choosing whether to start with a clean slate or pre-populated database, make sure the following requirements are met:

1. The database timezone is set to +00:00. You can verify this via:

   SELECT @@global.time_zone, @@session.time_zone;

2. The event scheduler is turned ON. You can verify this via:

   SELECT @@global.event_scheduler;

3. The binary logging during the creation of stored functions is set to 1. You can verify this via:

   SELECT @@global.log_bin_trust_function_creators;

4. The gse database exists. To create it:

   CREATE DATABASE gse CHARACTER SET utf8 COLLATE utf8_bin;

5. The gseadmin user exists. To create one, run:

   CREATE USER IF NOT EXISTS 'gseadmin'@'%' IDENTIFIED BY 'Lugano2020';
   GRANT ALL ON gse.* TO 'gseadmin'@'%';

If you prefer to begin with an empty database, there is nothing more for you to do. The required tables will be generated through Flyway migrations during the initial startup of the server. However, if you would like your local database to be pre-populated with the data we've collected, you can use the compressed SQL dump we offer. We host this dump, along with the four previous iterations, on Dropbox. After choosing and downloading a database dump, you can import the data by executing:

gzcat < gse.sql.gz | mysql -u gseadmin -pLugano2020 gse

Server

Before attempting to run the server, you should generate your own GitHub personal access token (PAT). The crawler relies on the GraphQL API, which is inaccessible without authentication. To access the information provided by the GitHub API, the token must include the repo scope.

Once that is done, you can run the server locally using Maven:

mvn spring-boot:run

If you want to make use of the token when crawling, specify it in the run arguments:

mvn spring-boot:run -Dspring-boot.run.arguments=--ghs.github.tokens=<your_access_token>

Alternatively, you can compile and run the JAR directly:

mvn clean package
ln target/ghs-application-*.jar target/ghs-application.jar
java -Dghs.github.tokens=<your_access_token> -jar target/ghs-application.jar

Here is a list of project-specific arguments supported by the application that you can find in the application.properties:

Variable Name	Type	Default Value	Description
ghs.github.tokens	List<String>		List of GitHub personal access tokens (PATs) that will be used for mining the GitHub API. Must not contain blank strings.
ghs.github.api-version	String	2022-11-28	GitHub API version used across various operations.
ghs.git.username	String		Git account login used to interact with the version control system.
ghs.git.password	String		Password used to authenticate the specified Git account.
ghs.git.config	Map<String,String>	See application.properties	Git configurations specific to the application2.
ghs.git.folder-prefix	String	ghs-clone-	Prefix used for the temporary directories into which analyzed repositories are cloned. Must not be blank.
ghs.git.ls-remote-timeout-duration	Duration	1m	Maximum time allowed for listing remotes of Git repositories.
ghs.git.clone-timeout-duration	Duration	5m	Maximum time allowed for cloning Git repositories.
ghs.cloc.max-file-size	DataSize	25MB	Maximum file size threshold for analysis with cloc.
ghs.cloc.timeout-duration	Duration	5m	Maximum time allowed for a cloc command to execute.
ghs.crawler.enabled	Boolean	true	Specifies if the repository crawling job is enabled.
ghs.crawler.minimum-stars	int	10	Inclusive lower bound for the number of stars a project needs to have in order to be picked up by the crawler. Must not be negative.
ghs.crawler.languages	List<String>	See application.properties	List of language names that will be targeted during crawling. Must not contain blank strings. To ensure proper operations, the names must match those specified in linguist.
ghs.crawler.start-date	Date	2008-01-01T00:00:00Z	Default crawler start date: the earliest date for repository crawling in the absence of prior crawl jobs. Value format: yyyy-MM-ddTHH:MM:SSZ.
ghs.crawler.delay-between-runs	Duration	PT6H	Delay between successive crawler runs, expressed as a duration string.
ghs.analysis.enabled	Boolean	true	Specifies if the analysis job is enabled.
ghs.analysis.delay-between-runs	Duration	PT6H	Delay between successive analysis runs, expressed as a duration string.
ghs.analysis.max-pool-threads	int	3	Maximum amount of live threads dedicated to concurrently analyzing repositories. Must be positive.
ghs.clean-up.enabled	Boolean	true	Specifies if the job responsible for removing unavailable repositories (clean-up) is enabled.
ghs.clean-up.cron	CronTrigger	0 0 0 * * 1	Delay between successive repository clean-up runs, expressed as a Spring CRON expression.

Web UI

The easiest way to launch the front-end is through the provided NPM script:

npm run dev

You can also use the built-in web server of your IDE, or any other web server of your choice. Regardless of which method you choose for hosting, the back-end CORS restricts you to using ports 3030 and 7030.

Dockerisation 🐳

The deployment stack consists of the following containers:

Service/Container name	Image	Description	Enabled by Default
gse-database	mysql	Platform database	✅
gse-migration	flyway	Database schema migration executions	✅
gse-backup	tiredofit/db-backup	Automated database backups	❎
gse-server	seart/ghs-server	Spring Boot server application	✅
gse-website	seart/ghs-website	NGINX web server acting as HTML supplier	✅
gse-watchtower	containrrr/watchtower	Automatic Docker image updates	❎

The service dependency chain can be represented as follows:

graph RL
    gse-migration --> |service_healthy| gse-database
    gse-backup --> |service_completed_successfully| gse-migration
    gse-server --> |service_completed_successfully| gse-migration
    gse-website --> |service_healthy| gse-server
    gse-watchtower --> |service_healthy| gse-website

Loading

Deploying is as simple as, in the docker-compose directory, run:

docker-compose -f docker-compose.yml up -d

It is important to note that the database setup steps explained in the preceding section aren't necessary when running with Docker. This is because the environment properties passed to the service will automatically create the MySQL user and database during the initial startup. However, this convenience doesn't extend to the database data, as the default deployment generates an empty database. If you wish to use existing data from the dumps, you will need to override the docker-compose deployment to employ a custom database image that includes the dump. To achieve this, create your docker-compose.override.yml file with the following contents:

version: "3.9"
name: "gse"

services:
    gse-database:
        image: seart/ghs-database:latest

The above image will include the freshest database dump, at most 15 days behind the actual platform data. For a more specific database version, refer to the Docker Hub page. Remember to specify the override file during deployment:

docker-compose -f docker-compose.yml -f docker-compose.override.yml up -d

The database data itself is kept in the gse-data volume, while detailed back-end logs are kept in a local mount called logs. You can also use this override file to change the configurations of other services. For example, specifying your own PAT for the crawler:

version: "3.9"
name: "gse"

services:
    # other services omitted...

    gse-server:
        environment:
            GHS_GITHUB_TOKENS: "A single or comma-separated list of token(s)"
            GHS_CRAWLER_ENABLED: "true"

Any of the Spring Boot properties or aforementioned application-specific properties can be overridden. Keep in mind that a property such as ghs.x.y corresponds to the GHS_X_Y service environment setting.

Another example is the automated database backup service, which is disabled by default. If you would like to re-enable it, you would have to add the following to the override file:

version: "3.9"
name: "gse"

services:
    # other services omitted...

    gse-backup:
        restart: always
        entrypoint: "/init"

FAQ

How can I request a feature or ask a question?

If you have ideas for a feature you would like to see implemented or if you have any questions, we encourage you to create a new discussion. By initiating a discussion, you can engage with the community and our team, and we will respond promptly to address your queries or consider your feature requests.

How can I report a bug?

To report any issues or bugs you encounter, create a new issue. Providing detailed information about the problem you're facing will help us understand and address it more effectively. Rest assured, we're committed to promptly reviewing and responding to the issues you raise, working collaboratively to resolve any bugs and improve the overall user experience.

How do I contribute to the project?

Refer to CONTRIBUTING.md for more information.

How do I extend/modify the existing database schema?

To do that, you should be familiar with database migration tools and practices. This project uses Flyway by Redgate. The general rule for schema manipulation is: create new migrations, and refrain from editing existing ones.

Footnotes

1. Only required in versions prior to 1.7.0 ↩ ↩²

2. We separate the application-level Git configurations from the ones used by the user to avoid any potential conflicts or confusion. As such, an application-specific configuration file is created in the temporary directory on startup. Settings added to the file depend on the ghs.git.config entries in the application.properties. Note that configuration subsections are currently not supported. ↩