Skip to content

Commit

Permalink
Support creating all pages
Browse files Browse the repository at this point in the history
  • Loading branch information
kou committed Jul 17, 2014
1 parent 564e3d4 commit 0508bb6
Show file tree
Hide file tree
Showing 2 changed files with 29 additions and 6 deletions.
8 changes: 8 additions & 0 deletions lib/wikipedia-search/path.rb
Expand Up @@ -93,6 +93,10 @@ def pages
data_dir + "#{@language}-pages.grn"
end

def all_pages
data_dir + "#{@language}-all-pages.grn"
end

def database_dir
data_dir + "db"
end
Expand Down Expand Up @@ -128,6 +132,10 @@ def pages
data_dir + "#{@language}-pages.jsons"
end

def all_pages
data_dir + "#{@language}-all-pages.jsons"
end

def schema
data_dir + "schema.json"
end
Expand Down
27 changes: 21 additions & 6 deletions lib/wikipedia-search/task.rb
Expand Up @@ -69,13 +69,15 @@ def define_data_convert_tasks

def define_data_convert_groonga_tasks
namespace :groonga do
base_command_line = [
"bzcat",
Shellwords.escape(@path.wikipedia.pages.to_s),
"|",
RbConfig.ruby,
"bin/wikipedia-to-groonga.rb",
]
file @path.groonga.pages.to_s => @path.wikipedia.pages.to_s do
command_line = []
command_line << "bzcat"
command_line << Shellwords.escape(@path.wikipedia.pages.to_s)
command_line << "|"
command_line << RbConfig.ruby
command_line << "bin/wikipedia-to-groonga.rb"
command_line = base_command_line.dup
command_line << "--max-n-records"
command_line << "5000"
command_line << "--max-n-characters"
Expand All @@ -85,6 +87,13 @@ def define_data_convert_groonga_tasks
sh(command_line.join(" "))
end

file @path.groonga.all_pages.to_s => @path.wikipedia.pages.to_s do
command_line = base_command_line.dup
command_line << "--output"
command_line << @path.groonga.all_pages.to_s
sh(command_line.join(" "))
end

desc "Convert Japanese Wikipedia page data to Groonga page data."
task :ja => @path.groonga.pages.to_s
end
Expand Down Expand Up @@ -113,6 +122,12 @@ def define_data_convert_droonga_tasks
@path.groonga.pages.to_s)
end

file @path.droonga.all_pages.to_s => @path.groonga.all_pages.to_s do
sh("grn2drn",
"--output", @path.droonga.all_pages.to_s,
@path.groonga.all_pages.to_s)
end

desc "Convert Japanese Wikipedia page data to Droonga page data."
task :ja => @path.droonga.pages.to_s
end
Expand Down

0 comments on commit 0508bb6

Please sign in to comment.