• Parses a context string as a regular URL (RFC 2396, "Uniform Resource Identifier") and transforms it according to the given URL pattern. A new DocumentChunk is created with the substitution. Pattern used to transform the URL (in the form <scheme>://<authority><path>?<query>#<fragment>):

◦ Characters other than '$' or '\' are kept as-is

◦ The '$' character and the '\' character must be escaped with a leading \

◦ The ${expression} form allows to compute a string expression based on URL components (see "Expression" below)

Expression used inside the enclosing ${}:

◦ url: Original URL

◦ scheme: Scheme name ("http", "https", "file", ...)

◦ authority: Authority (host:port or host) (may be empty)

◦ host: Hostname part of the authority (may be empty)

◦ port: Port number part of the authority (may be empty)

◦ userInfo: username:password field of the authority (may be empty)

◦ file: File starting with / and query string, if any

◦ pathurl: Normalized absolute path starting with /

◦ path: Normalized absolute path (may start with C:\ on Windows)

◦ query: Normalized query part starting with ? (may be empty)

◦ args: Query part without the leading ? (may be empty)

◦ fragment: Fragment part starting with #(may be empty)

◦ reference: Reference part ; i.e., fragment without the leading # (may be empty)

◦ arg:name: Query part argument identified by its name, unescaped (you must re-escape it using "urlencode:" when necessary)

◦ str:string: The final argument is not a variable name, but a string (only useful for clarity purpose)

◦ tolower:<i>expression</i>: Transform into lowercase (ONLY A-Z)

◦ toupper:<i>expression</i>: Transform into uppercase (ONLY a-z)

◦ urlencode:<i>expression</i> :URL encoding (%NN or +)

◦ urlpathencode:expression</i>: URL encoding outside / fragments

◦ urldecode:<i>expression</i>: URL decoding

◦ pathslash:<i>expression</i>: Convert \ into /

◦ pathantislash:<i>expression</i>: Convert / into \

Notes:

◦ Unreserved characters are unescaped during URL processing (i.e., never '%' or '\')

◦ The lower other similar prefix accept recursion (i.e., the expression "${urlpathencode:pathantislash:toupper:path}" is valid)

◦ Both "file://C:\path" and "file:///C:\path" will produce path="/C:\path"

Examples:

◦ With the input context value "http://www.example.com/bar/foo?bar=42"

▪ "hello, world" => "hello, world"

▪ "the scheme is ${scheme}" => "the scheme is http"

▪ "the scheme is \${scheme}" => "the scheme is \${scheme}

▪ "http://myserver${path}${query}" => "http://myserver/bar/foo?bar=42"

▪ "http://myserver/applet?f=${urlpathencode:path}&t=${arg:bar}" => "http://myserver/applet?f=/bar/foo&t=42"

▪ "http://myserver/applet?f=${urlencode:path}&t=${arg:bar}" => "http://myserver/applet?f=%2Fbar%2Ffoo&t=42"

▪ "http://myserver/applet?f=${urlpathencode:pathantislash:toupper:path}" => "http://myserver/applet?f=%5CBAR%5CFOO"

◦ With the input context value "file:///C:/My%20Documents/Document.doc"

▪ "${pathantislash:urldecode:path}" => "C:\My Documents\Document.doc"

• Parent elements:

◦ com.exalead.indexing.analysis.v10.AnalysisPipeline (as AnalysisPipeline)

◦ com.exalead.indexing.analysis.v10.DocumentProcessorGroup (as DocumentProcessorGroup)

• Attributes:

• Nested elements: