Add ability to set up without the UI about xowa

[gnosygnu]

Hi! This is actually available already. You can refer to home/wiki/Help:Import/Command-line within XOWA. There's a section for command-line import at: home/wiki/Help:Contents

Let me know if you run into any issues, or if the instructions aren't clear. Thanks!

[pioterj]

Good to hear. I can't find the instructions you are referring to. Are they somewhere at https://gnosygnu.github.io/xowa/?

[gnosygnu]

Nope. XOWA currently has most of it documentation within the app. In this case, you would do the following:

• Start XOWA:
• Copy-paste "home/wiki/Help:Import/Command-line" to the url bar

I list the wikitext below, but you're better off reading it within XOWA.

I am planning to upload these to https://gnosygnu.github.io/xowa/. However, there are a lot of pages and I'd like to automate generation and synchronization of them. If I can't get around to coding a system in the next few months, I'll just upload them all by hand.

XOWA can import a wiki using a plain text file and a command-line.
{{Help/Css}}

== Import simple.wikipedia.org through the command-line ==
* Open up a command-line. For example, on Windows, run <span class='bold'>cmd</span>
* Run the following: <span class='console'>java -jar {{#invoke:Xowa_url|plat_jar}} --cmd_file {{#invoke:Xowa_url|plat_url|xowa_build.gfs}} --app_mode cmd</span>
* Wait about 10 minutes for the script to complete
* Launch XOWA and enter <span class='url'>simple.wikipedia.org</span> in the URL bar

== Import a different wiki by editing the build script ==
* Open the following file in a [[Help:Text_editor|text editor]]: <span class='path'>{{#invoke:Xowa_url|plat_url|xowa_build.gfs}}</span>. See Script below for the full text.
* Replace all instances of <span class='bold'>simple.wikipedia.org</span> with the domain name. For example, for English Wikipedia, use <span class='bold'>en.wikipedia.org</span>
* Run the command-line import again.
* Launch XOWA and enter in the domain name in the the URL bar.

== Import a wiki with a manual download ==
=== Download the wiki dump ===
* Navigate to https://dumps.wikimedia.org/enwiki
* Click on the '''latest''' directory
* Download the file just under "'''Articles, templates, media/file descriptions, and primary meta-pages.'''". It should read '''enwiki-latest-pages-articles.xml.bz2'''
: The download is 11+ GB and may take anywhere between 2 and 5 hours to complete.
: If you also want talk pages, you should download the "'''Recombine all pages, current versions only.'''" version. It should read '''enwiki-latest-pages-meta-current.xml.bz2'''. Note that this dump is twice the size of the regular dump.

=== Specify location of the wiki dump ===
* In the build script, replace the following line:
: <span class='code'>add ('simple.wikipedia.org', 'text.init') {src_bz2_fil = '/your_directory/simplewiki-20130103-pages-articles.xml.bz2';}</span>

== Script ==
<pre class='code'>
// do not show a "Press enter to continue" at the end of the script
app.bldr.pause_at_end = 'n';

// run xowa.gfs
app.scripts.run_file_by_type('xowa_cfg_app');

// import wiki; for more info see [[Help:Import/Command-line]]
app.bldr.cmds {
  // delete all files in directory; note that subdirectories and file databases ("-file.xowa") will not be deleted
  add     ('simple.wikipedia.org' , 'util.cleanup')          {delete_all = 'y';}

  // download main dump file; contains all articles
  add     ('simple.wikipedia.org' , 'util.download')         {dump_type = 'pages-articles';}

  // download categorylinks file; contains links from category to pages
  add     ('simple.wikipedia.org' , 'util.download')         {dump_type = 'categorylinks';}

  // download page_props file; contains information on hidden categories
  add     ('simple.wikipedia.org' , 'util.download')         {dump_type = 'page_props';}

  // start wiki import
  add     ('simple.wikipedia.org' , 'text.init');

  // import articles
  add     ('simple.wikipedia.org' , 'text.page');

  // generate search data
  add     ('simple.wikipedia.org' , 'text.search');

  // generate main category data
  add     ('simple.wikipedia.org' , 'text.cat.core');

  // import category links
  add     ('simple.wikipedia.org' , 'text.cat.link');

  // apply hidden categories
  add     ('simple.wikipedia.org' , 'text.cat.hidden');

  // end import
  add     ('simple.wikipedia.org' , 'text.term');

  // import css into wiki
  add     ('simple.wikipedia.org' , 'text.css');

  // cleanup temp files; delete xml and bz2
  add     ('simple.wikipedia.org' , 'util.cleanup')          {delete_tmp = 'y'; delete_by_match('*.xml|*.sql|*.bz2|*.gz');}
}

// run cmds
app.bldr.run;
</pre>

[pioterj]

Thanks. One of the steps is "Launch XOWA" that involves starting a UI. So actually there's no way to use it in completely headless mode, right?

[gnosygnu]

So actually there's no way to use it in completely headless mode, right?

Well, there are two other ways, but I'm not sure how they'll work for your environment:

Run XOWA as an HTTP-server

• Open up a shell
• Run the following: java -jar xowa_linux_64.jar --app_mode http_server
• Open up a web-browser
• Navigate to any of the following links:
  • http://localhost:8080/
  • http://localhost:8080/home/wiki/Help:Contents
  • http://localhost:8080/home/wiki/Help:Import/Command-line

Run XOWA in command-line mode

• Open up a shell
• Run either of the following:
  • (wikitext) java -jar xowa_linux_64.jar --app_mode cmd --show_license n --show_args n --cmd_text "app.shell.fetch_page('Help:Import/Command-line' 'wiki');"
  • (html) java -jar xowa_linux_64.jar --app_mode cmd --show_license n --show_args n --cmd_text "app.shell.fetch_page('Help:Import/Command-line' 'html');"
• Read the text in the shell, or pipe to a text file and read in a text-editor

Let me know if neither of the above works. Thanks.

[gnosygnu]

I'm going to mark this ticket closed. The command-line route should handle setup without a UI. Keep in mind this is what I use to generate all the wikis for archive.org. If you have questions, please feel free to reopen the issue and I'll respond. Thanks.