Help Reference and Example

The functionality of xwordlist can be divided into three categories:

  • Input and Output: start with content from your local machine and add more programmatically from the web. Build a list of URLs and crawl multiple pages on the web. Save everything to a file name of your choosing or let the software create one for you.
  • Content Parsing: take content from multiple sources and parse it into a normalized list of words. Grab all the words on a web page or limit to one ID, one or more classes and/or one or more HTML tags. Use a regex pattern to get rid of anything you don’t need.
  • Word Transformation: refine your list into something you can paste directly into your construction software — alphabetized, deduped, properly cased and with all the extraneous words filtered out.

Basic Example

The easiest way to understand xwordlist is to walk through the example of getting a list of song titles, something you might want to do if you are building a themed puzzle about a particular musician. In this case, we’ll be grabbing a list of songs by Tom Petty from this page on SongLyrics:

To see what raw output looks like, we’ll first just get the whole page without any refinement:

xwordlist --webpage

After running this command, you should see a new file called songlyrics_com_xwl.txt in your xwordlist folder. Since no output file name was specified, a default based on the URL was created. Open the file with your favorite text editor and you will see that we have generated a file with all of the text strings from the SongLyrics web page, including random bits of page titles, navigation and advertising verbiage.

That is a good start, but now let’s just grab the text from the table in the middle with the title “Tom Petty Lyrics - by Popularity” which includes the list of song titles we are looking for. By examining the HTML of the page, we see that the text is contained in a table with the class name “tracklist”, which we can specify using the --container option. Next try the command:

xwordlist --webpage --output tompetty.txt --container class=tracklist=1 --webextract html-a

Rather than let a default file be created, this time we have specified that the output should go to tompetty.txt; open that file to see the result. You may have noticed that when we specified the --container option, we included class=tracklist=1. If we had specified an ID, it would have to be unique to the page but there can be multiple uses of the same class name on a page and in this case, we wanted the first. If we had left out the =1 on the end, xwordlist would have given us all of the instances of that class on the page. Try it to see what different outputs look like!

The list of song names is numbered, which is not something we want for our word list. By using --webextract html-a, we ask the software to further refine our query to only give us text within <a> tags, which very nicely contains each song title but not the numbers. If for some reason we wanted multiple tags — say for instance all the content that is either bolded or italicized — we could have specified --webextract html-i_b_em_strong which would include all of the text within <i> <b> <em> <strong> tags. Chain as many HTML tags as you need by separating them with the underscore character.

Now that we have a list of song titles, there are a few bits of clean up since most of these kinds of lists can be a bit noisy. For instance, this list of songs by Tom Petty includes multiple versions of the same song with additional words that we may not want for our final word list. First, we should alphabetize it to make it easier to scan. Rather than go back to the web, we’ll use the file we just created:

xwordlist --input tompetty.txt --output tompetty.txt --alphabetize

Of course, you could save the alphabetized list to a different file name (or let the software create a new file) but for the purposes of this example, we will stick with the same file. Scan through that file and delete duplicates and anything that you wouldn’t want in your word list. Don’t worry about extraneous characters — like for instance the apostrophe on the end of Free Fallin’ — the software will clean that up. Once you have made any edits to that file, be sure to save it before continuing on to the next step.

From here, there are two ways we can go with this list. We may want to create a list where the entire song name is an entry, for instance for long themed entries such as “WONTBACKDOWN”. To do that, we need to strip out all non-alphabetic characters (including spaces) and dedupe the list. So first run this:

xwordlist --input tompetty.txt --output tompetty_songs.txt --strip --dedupe --case --alphabetize

We saved this to a new file name for reasons that will become obvious in a second. Go open that new file tompetty_songs.txt and you will see a list of Tom Petty songs that are ready to be imported into your crossword construction software. But since most of these are long titles that are hard to fit into a standard crossword, we also want the individual words: for instance just the word “BREAK” to clue as “___ Down” for a shorter entry. To do that, we need to add the --convert option on our original list which will separate each title’s words into a new list. Try:

xwordlist --input tompetty.txt --output tompetty_title_words.txt --convert --strip --dedupe --case --alphabetize --minimum

You now have two great lists to use to start building your Tom Petty puzzle! You can paste them into one doc and have xwordlist alphabetize and dedupe or you can just paste each individual list directly into your construction software, perhaps to score one set of words higher than the other. Experiement and figure out what works best for your constructing needs.

To make things easier, xwordlist has two catch-all options: --line2word and --word2word. If you want each line to become a single word — as in the case where a song name on each line becomes “WONTBACKDOWN” — you can use --line2word instead of having to add --strip --dedupe --case --alphabetize --minimum. If you want every word in the block of text to become a separate word — as in the case where a song name on each line becomes “WONT”, “BACK” and “DOWN” — use the --word2word option.

Also, check out the recipes page for other possibilities. You can use the software to grab and convert the lyrics from individual songs, spider all of the songs at once or parse Wikipedia for album titles. Let us know what you find!

List of Available Options

When a default option is specified, you can leave it out. For instance, rather than entering --dedupe nocase, you can just add --dedupe. To change a default, see the Changing Global Settings section below.

The order that options are specified in does not matter, the software knows the right order to take these operations so they don’t conflict with each other.

Input and Output

xwordlist only works with text files, if it is given binary data it will let you know and then quit.

--input or -i file1.txt [file2.txt ...]

Use the text file named file1.txt as an input for the parsing and transformation engine. Multiple input files can be specified and all will added, use --input file1.txt file2.txt.

If you specify other input sources (such as --webpage and --urllist), all will be added together before processing by the parsing and transformation engine.

--webpage or -w URL

Use the web page located at the address named by the URL as an input into the parsing and transformation engine. The software will grab all of the available text strings from the web page, unless other options are specified to narrow the request down (see --container and --webextract for more information). The URL should be the fully qualified address, including http:// or https://. The software does not support authenticated sites.

If you specify other input sources (such as --input and --urllist), all will be added together before processing by the parsing and transformation engine.

--urllist filename.txt

Read the text file named filename.txt and loop through each of the URLs to find input for the parsing and transformation engine. This option functions the same as running the program multiple times using the --webpage option for each entry in filename.txt.

To work effectively, you should limit the URLs to the same site (or multiple sites with the same HTML structure) so that the software is able to parse the HTML consistently for each URL. Although you can specify multiple URLs, the text parsing and conversion options will act the same on everything retrieved.

The program pauses 20 seconds between each URL requested to be a good citizen and avoid getting blocked. To change this delay, see the Changing Global Settings section below. If you specify other input sources (such as --input and --webpage), all will be added together before processing by the parsing and transformation engine.

--output or -o filename.txt

Save the list of words output by the parsing and transformation engine to the file named filename.txt. If you do not specify an output file name, a file will be created for you based on either the input file name or the domain name of the input web address plus the string _xwl. To change this string, see the Changing Global Settings section below.

Your input and output files can be the same, the software will always prompt you before writing over an existing file.

--directory /path/to/folder

Set a directory to use for all input and output files. The software will look in the directory specified for files named in the --input and --urllist options and will write any file specified by the --output option (or the default file if no output is specified) to this directory.

To avoid entering the same directory on the command line each time, add your default directory to the configuration file. See the Changing Global Settings section below for more information.

For all inputs and outputs you can also specify a path as part of a file name as in --input /path/to/wordlist.txt.

Content Parsing

--container tag=name

Limit the content read from web page(s) specified by either --webpage or --urllist to a particular HTML id or class tag. You must specify both the tag and the name of the element requested, for example --container id=page_content or --container class=wikitable.

Since the id tag must be unique on a properly structured HTML page, specifying an id element will return a single container. To specify a single instance of a class attribute, add =N which will only grab the Nth time that label appears on the page. For example, class=tracklist=1 will retrieve the first instance of the class named tracklist on the page. If you leave out the =N attribute, the software will grab all instances of that class.

Specify what kind of content to grab from web pages. The default is text and for most web pages you will not need to use this option at all.

Use the links option when you want to retrieve URLs from a web page, for instance to build a text file of URLs to then feed to the --urllist option. links only returns URLs specified by <a> tags, it parses out the values located in the href element. If it finds relative URLs, it builds fully-qualified links based on the page’s URL.

Use the html-XX to further refine which text to pull from a web page by limiting to only text within specified tags. For instance, use html-b to only return text inside all <b> and </b> blocks. You can chain multiple tags by separating them with the underscore character, for example html-i_em will get all content in either <i> or <em> blocks.

--regex pattern

Refine which text to keep by using regex patterns. Whereas other options such as --container or --webextract will grab all of the text within whatever containers and/or elements specified, --regex allows you to narrow down to particular parts of text strings to keep. For examples of patterns to use, see regex examples. (TK: regex examples)

Word Transformation


Convert each line of text into a word by stripping out spaces and other non-alphabetic characters, dedupe the list, alphabetize and capitalize every word. Remove all words below the minimum word size. This is the equivalent of doing the default behavior for all the following options: alphabetize, case, dedupe, minimum, strip.

To change the default behaviors for any of the items called by line2word, see the Changing Global Settings section below.


Convert blocks of text into individual words delimited by spaces, strip out non-alphabetic characters, dedupe the list, alphabetize and capitalize every word. Remove all words below the minimum word size. This is the equivalent of doing the default behavior for all the following options: alphabetize, case, convert, dedupe, minimum, strip.

To change the default behaviors for any of the items called by word2word, see the Changing Global Settings section below.

--alphabetize or -a normal (default) | reverse

Alphabetize the list of words.

--case upper (default) | lower | none

Change the case of all words in the list.

--convert "chars" (default is " " space)

Take any block of text and turn it into a list of words. If "chars" is not specified, the software uses spaces to separate text into words and in most cases, that will suffice. If you do need to specify "chars", put all characters to be used as delimiters inside the quotes. For instance, if you want to separate text blocks connected by either dashes or spaces, you would use --convert "- ". Be careful with back slashes (\) which can act as an escape character.

--dedupe or -d nocase (default) | bycase

Remove all duplicates in the list. The dedupe function by default is case insensitive: “apple” and “APPLE” are treated as the same word and the first instance found is kept. Use --dedupe bycase if you want the dedupe routine to be case sensitive.

--minimum or -m N (default is 3)

Remove all words with less than N characters.

--strip or -s diacritic (default) | keepdiacritic

Remove all non-alphabetic characters, to make your list crossword puzzle ready. By default, the software first converts diacritics to English equivalents, for instance “Renée” is turned into “Renee”. To leave diacritics as is, use --strip keepdiacritic.

Meta Options

--help or -h

View the help message and exit.

--version or -v

View the version number and exit. For a record of changes and to see the most current version number, see the changelog.

Changing Global Settings

Global defaults and settings for line2word and word2word are stored in a configuration file that is created on first run. The configuration file is named xwordlist.conf and is stored within a subdirectory of the user’s home directory located at ~/.config/xwordlist. The default configuration file should look like this:

urllist_delay = 20
file_add = xwl
impact_color = ansired

convert = 
strip = diacritic
minimum = 3
case = upper
dedupe = nocase
webextract = text
alphabetize = normal

There is currently no user interface built into xwordlist to modify these defaults; use your text editor to change an entry in the configuration file to change a default. For instance, by changing the line minimum = 3 to minimum = 4, you will cause every instance of line2word or word2word to remove all words with less than 4 letters instead of the default of 3.

To set a default directory to avoid having to enter --directory ~/my_files each time on the command line, add a line in the [globals] section with your default directory — for instance, directory = ~/my_files. You can always override the default setting by using --directory /something/else in the command line.

If you enter an incorrect entry or option, xwordlist will delete the entry and revert back to the default.

impact_color is the color used in the command line to emphasize important parts of a response. Possible colors are ansiblack, ansired, ansigreen, ansiyellow, ansiblue, ansimagenta, ansicyan, ansigray, ansibrightblack, ansibrightred, ansibrightgreen, ansibrightyellow, ansibrightblue, ansibrightmagenta, ansibrightcyan and ansiwhite.