Linux Package Search is a simple search tool to retrieve information for Linux packages. Supported distributions are:
- Ubuntu
- Debian
- Fedora
Run ./build.sh to get started.
The application is bundled with an Express server to make proxy search requests. See package.json for a full list of dependencies.
Access localhost/v1/api-docs to access Swagger documentation for the available API endpoints. Access localhost/api/v1/api-docs to view the swagger configuration in JSON format.
Endpoints are protected using an API secret. This value is set using the NODE_API_SECRET environment variable. Include the secret in the Authorization header of your HTTP requests.
A command line version of the application is also included. Run npm install . to install the linsearch command globally. Execute linsearch --help to view all available commands.
search|s Search for a list of packages by name.
view|v View details for the specified package.
arsearch|as Search for a list of archived packages by name.
arview|av View archived details for the specified package.
arsave|asv Save the specified package to the archive.
ardel|ad Delete the specified package from the archive.
isearch|is Perform a full-text search of indexed content.
index|i Index the specified package for the search engine.
idel|id Delete the specified package from the index.
Ensure that a .env file is included in the current directory when executing linsearch.
Package information is archived into a Postgres database.
Execute make dbdump to view the contents of the database as raw SQL (the output should ideally be redirected to a file). Execute make dbexp to export the database into an SQL file inside the data directory. Execute make dbimp to import the SQL file into the database.
By default, content will be stored inside of the linux_packages database. This consists of a distribution table, which stores a list of supported Linux distros, and a packages table, which contains archived package information. The packages table contains the name (string, 255), displayName (string, 255), version (string, 255), search_query (string, 255), and additionalProperties (JSON binary) columns.
Access Adminer from localhost:$ADMINER_PORT. This defaults to port 8080.
This app uses the Knexjs module to handle seeds, migrations, and SQL queries. Knex-related scripts can be accessed from the server directory.
The seed script will fill the distribution table with supported distributions.
The migration table will define the structure of the distribution and packages tables.
The knexfile.js file exclusively uses environment variables. Ensure that you export the contents of the .env file into your current shell environment.
This application leverages Redis to cache your JSON-formatted search results. Set the cache lifetime using the NODE_CACHE_LIFETIME environment variable.
Cache keys use the pattern {command}-{distribution}-{package}.
A fallback in-memory cache is also included should Redis be unavailable. Set NODE_CACHE_BACKEND to memory to enable this cache.
Content can be indexed into Elasticsearch to enable full-text search functionality.
ES_JAVA_OPTS can be used to allocate the amount of memory available to the JVM. The default value is "ES_JAVA_OPTS=-Xmx1024m -Xms1024m". You may increase or decrease these values to fit your own requirements.
The Elasticsearch container is initially configured to use a single node cluster for local development. Comment out DISCOVERY_TYPE from the docker-compose file (or override the Elasticsearch configuration using another docker-compose file) to enable multi-node clustering. Change ELASTICSEARCH_CLUSTER_INITIAL_MASTER_NODES to specify which nodes should be included in the cluster.
Search results are cached by Redis using the query string as a key.