debshots/lib/tasks/import_debian.rake

279 lines
11 KiB
Ruby

require 'open-uri' # allows to load URLs using open()
require 'deb_importer'
# The repository format is documented at:
# https://wiki.debian.org/RepositoryFormat
include DebImporter
# List of regular expressions. If the package name matches any
# of these then the package will not be imported.
BLACKLIST_NAME_PATTERN=[
/-data$/,
/-dev$/,
/-doc$/,
/-dbg$/,
/-common$/,
/-l10n-/,
/-locale-/,
]
# List of regular expressions. If the package's section matches
# any of these then the package will not be imported.
# /?...$ is used because Ubuntu adds their own section - e.g. multiverse/debug
BLACKLIST_SECTION_PATTERN = [
/^debian-installer$/,
/\/?translations$/,
/\/?debug$/,
/\/?kernel$/,
/\/?localization$/,
/\/?oldlibs$/,
/\/?libdevel$/,
/\/?cli-mono$/,
]
# Whether to delete a blacklisted package from the database
REMOVE_BLACKLISTED_PACKAGE = true
namespace :debshots do
desc "Import/update package database from configured DEB repositories"
task :update_from_deb_repos => :environment do
# Counters for added, updated or removed packages from the database
stats_added = 0
stats_updated = 0
stats_removed = 0
repositories = Rails.configuration.package_sources
Rails.logger = Logger.new(STDOUT)
Rails.logger.level = Logger::INFO
#Rails.logger.level = Logger::DEBUG
Rails.logger.info "Importing Debian package information"
repositories.each do |repository|
wanted_architectures = repository[:architectures]
Rails.logger.info "Fetching Release file for repository: #{repository[:url]}"
release = DebImporter::Release.new(repository[:url])
Rails.logger.info "> Supported architectures are: #{release.architectures}"
if wanted_architectures
Rails.logger.info "> We only want architectures: #{wanted_architectures}"
end
Rails.logger.info "> Supported components are: #{release.components}"
# TODO: Remember what we imported to show it on the /about page
release.components.split.each do |component|
Rails.logger.info "> Component: #{component}"
release.architectures.split.each do |architecture|
if wanted_architectures and not wanted_architectures.include?(architecture)
Rails.logger.debug "Architecture #{architecture} not wanted. Skipping."
next
end
Rails.logger.info ">> Architecture: #{architecture}"
# Check if this component and architecture is available on this mirror
packages = release.packages(component, architecture)
unless packages
Rails.logger.error "No packages for component #{component} on architecture #{architecture} found on this mirror"
next
end
Rails.logger.info "Packages file for #{component} on #{architecture} found."
Rails.logger.debug "Got: #{packages}"
packages.each do |package|
Rails.logger.info "> Package: #{package[:Package]}"
#Rails.logger.debug "Fetching package informaton from the database"
if package_name_blacklisted? package[:Package] or package_section_blacklisted? package[:Section]
# Should the package get removed from the database?
if REMOVE_BLACKLISTED_PACKAGE
db_package = Package.find_by name: package[:Package]
if db_package
Rails.logger.info "Removing blacklisted package '#{package[:Package]}' from database"
db_package.destroy
stats_removed += 1
end
end
next
end # if blacklisted
db_package = Package.find_by name: package[:Package]
if db_package # package exists in the database
# check if the parsed package data has a newer version
current_version = DebImporter::Version.new(db_package.version)
new_version = DebImporter::Version.new(package[:Version])
Rails.logger.debug "Comparing package version: old=#{db_package.version} new=#{new_version}"
if new_version > current_version
# update the package information from new data
update_data(package, db_package)
Rails.logger.info "Updating package information"
stats_updated += 1
end
else # new package - create it in the database
Rails.logger.debug "No such package in our database. Creating one."
db_package = Package.new
data_has_changed = update_data(package, db_package)
stats_added += 1
end
# unless db_package
# Rails.logger.debug "No such package in our database. Creating one."
# db_package = Package.new
# stats_added += 1
# end
# # Rails.logger.debug "Found in database: #{db_package}"
# data_has_changed = update_data(package, db_package)
# stats_updated += 1 if data_has_changed
end # package.each
end # architectures.each
end # components.each
Rails.logger.info "Done parsing #{repository[:url]} repository."
Rails.logger.info "#{stats_added} packages added"
Rails.logger.info "#{stats_updated} packages updated"
Rails.logger.info "#{stats_removed} packages removed"
Rails.logger.info "--------------------------------------------"
end # repositories.each
end # task
desc 'Update long description from i18n file'
task :update_longdescription_from_deb_repos => :environment do
repositories = Rails.configuration.package_sources
Rails.logger = Logger.new(STDOUT)
Rails.logger.level = Logger::INFO
# Rails.logger.level = Logger::DEBUG
Rails.logger.info "Importing long description from Debian repository (i18n)"
repositories.each do |repository|
Rails.logger.info "Fetching Release file for repository: #{repository[:url]} with components: #{repository[:components]}"
release = DebImporter::Release.new(repository[:url])
Rails.logger.info "> Supported components are: #{release.components}"
release.components.split.each do |component|
Rails.logger.info "> Component: #{component}"
release.i18n(component, 'en').each do |pkg|
Rails.logger.debug "i18n information: #{pkg}"
unless pkg[:'Description-en']
Rails.logger.debug "No Description-en found in section."
next
end
if package_name_blacklisted? pkg[:Package]
next
end
# See if we have that package in the database
if db_pkg = Package.find_by(name: pkg[:Package])
Rails.logger.info "Updating long description for package #{db_pkg.name}"
text = pkg[:'Description-en']
short_description, long_description = text.split("\n", 2)
db_pkg.long_description = long_description
db_pkg.save
else
Rails.logger.debug "Package not in database. Nothing to update."
end
end # pkg.each
end # components.each
Rails.logger.info "Done updating long descriptions from translation file."
Rails.logger.info "--------------------------------------------"
end # repositories.each
end # task
# TODO: Some code duplication. Should be refactord.
desc "List configured DEB repositories"
task :list_deb_repos => :environment do
repositories = Rails.configuration.package_sources
Rails.logger = Logger.new(STDOUT)
Rails.logger.level = Logger::INFO
# Rails.logger.level = Logger::DEBUG
Rails.logger.info "Listing configured DEB repositories"
repositories.each do |repository|
wanted_architectures = repository[:architectures]
Rails.logger.info "Fetching Release file for repository: #{repository[:url]}"
release = DebImporter::Release.new(repository[:url])
Rails.logger.info "> Supported architectures are: #{release.architectures}"
if wanted_architectures
Rails.logger.info "> We only want architectures: #{wanted_architectures}"
end
Rails.logger.info "> Supported components are: #{release.components}"
release.components.split.each do |component|
Rails.logger.info "> Component: #{component}"
release.architectures.split.each do |architecture|
if wanted_architectures and not wanted_architectures.include?(architecture)
Rails.logger.debug "Architecture #{architecture} not wanted. Skipping."
next
end
Rails.logger.info ">> Architecture: #{architecture}"
# Check if this component and architecture is available on this mirror
packages = release.packages(component, architecture)
if packages
Rails.logger.info "#{packages.count} packages found."
else
Rails.logger.info "No packages."
end
end # architectures.each
end # components.each
end # repositories.each
end # task
end # namespace
# Update the information about a package in the database
# package: Information about new package (parsed from remote source)
# db_package: Information that currently exists in the database
def update_data(package, db_package)
Rails.logger.debug "New information: #{package.inspect}"
# Rails.logger.info "New package version found. Updating details in database."
db_package.version = package[:Version]
db_package.name = package[:Package] unless db_package.name # set the name for new packages
db_package.description = package[:Description][0..79]
db_package.homepage = package[:Homepage]
db_package.section = package[:Section]
db_package.origin = package[:Origin]
package[:Maintainer]=~/^(.+) <(.+)?>/
db_package.maintainer = $1
db_package.maintainer_email = $2
db_package.save
Rails.logger.debug "Saved package data: #{db_package.inspect}"
end
# Check if a package name is blacklisted and should not be imported.
# This prevents importing boring software that likely does not get a useful screenshot.
def package_name_blacklisted?(name)
BLACKLIST_NAME_PATTERN.each do |pattern|
if name=~pattern
Rails.logger.debug " > Blacklisted by name ('#{name}' matches '#{pattern}'')"
return true
end
end
return false
end
# Check if a package's section is blacklisted and should not be imported.
# This prevents importing boring software that likely does not get a useful screenshot.
def package_section_blacklisted?(section)
BLACKLIST_SECTION_PATTERN.each do |pattern|
if section=~pattern
Rails.logger.debug " > Blacklisted by section ('#{section}' matches '#{pattern.to_s}')"
return true
end
end
return false
end