Examples and the sample code of SSAX parsing and SXML transformations You need GNU Make to build the targets. To run any of the examples, you need to do make PLATFORM=

where is scmi, petite, biglooi. See the Makefile in this directory for more details. Remove-tags example ------------------- Target name: remove-markup Given an XML document, remove all the markup and print the resulting document. This is a simple in-out application. Outline example --------------- Target name: outline Pretty-print the structure of an XML document (disregarding the character data) This example corresponds to outline.c of the Expat distribution. The example demonstrates how to transform an XML document on the fly, as we parse it. Note that the tags of elements with no XML namespace are printed as they are. Tags of elements within an XML namespace are printed as a pair (Namespace-URI . Local-name) SXML example ------------ Target name: run-sxml Transform an XML document into SXML. See ../docs/SXML.html for the description. Permissive HTML parsing ----------------------- Target name: html-parser This example gets SSAX to _permissively_ parse HTML documents or HTML fragments. Because HTML browsers are so lax, many web pages on the Internet contain invalid HTML. For example, the FrontPage editor is notorious for creates such invalid sequences as text. SSAX is more than an XML parser -- it is a library, which includes lexers and parsers of various kinds. The library comes in handy when we need to permissively parse (or better say, lex) HTML. Because we accept even HTML with unmatched tags, we make no attempt in this example to recover the structure. We faithfully record all the occurring tags and let the user sort out what matches what. Here's the result of parsing of an ill-composed sample document, which includes comments, parsed entities and attributes of various kinds. Source: whatever link

BLah italic bold ened still < bold

But not done yet... Result: a flat list (a token stream) (#(START html ()) " " #(START head ()) " " #(START title ()) " " #(END title) " " #(START title ()) " whatever " #(END title) " " #(END head) "\n " #(START body ()) " " #(START a ((href . "url"))) "link" #(END a) " " #(START p ((align . "center"))) " " #(START ul ((compact . "compact") (style . "aa"))) "\n " #(START p ()) " BLah " #(START i ()) " italic " #(START b ()) " bold " #(START tt ()) " ened " #(END i) " still < bold " #(END b) "\n " #(END body) "\n " #(START P ()) " But not done yet...") Note that the parser handled "<" and other parsed entity references. The parser fully preserved all the whitespace (including newlines). Also note that the three possible styles of HTML attributes are handled properly. In addition, single quotes are also allowed. Comments are silently skipped. This example, after it is built, parses the sample HTML document and prints out the above result. SXML Transformations -------------------- The following three examples demonstrate SXML transformations, from trivial to advanced. The advanced examples exhibit context-sensitive transformations: the result of the conversion of an SXML tag depends on that tag's context, i.e., siblings, parents or descendants. See also ../docs/Makefile for an SXML transformation that turns SXML.scm (the SXML Specification's master file) into HTML and LaTeX documents. The examples in this directory: apply-templates.scm A simple example of a XSLT-like 'apply-templates' form sxml-db-conv.scm Using higher-order transformers for context-sensitive transformations. The example also shows the treatment of XML Namespaces in SXML and the namespace-aware pretty-printing of SXML into XML. pull-punct-sxml.scm A Context-sensitive recursive transformation and breadth-first-like traversals of a SXML document sxml-nesting-depth-label.scm Labeling nested sections of an SXML document by their nesting depth. This is another example of traversing an SXML document with a stylesheet that differs from node to node. sxml-to-sxml.scm Transforming SXML to SXML: Composing SXML transformations. The code introduces and tests a version of pre-post-order that transforms an SXML document into a _strictly conformant_ SXML document. That is, the result of a pre-post-order transformation can be queried with SXPath or transformed again with SXSLT. The names of the targets are the names of the above files without the .scm extension. Parsing and Unparsing of a Namespace-rich XML document ------------------------------------------------------ Target name: daml-parse-unparse The file daml-parse-unparse.scm parses a DAML (DARPA Agent Markup Language -- XML ontology) document and unparses it. A DAML document has a _great_ deal of namespaces. Some of them are represented by namespace shortcuts, some are not. The test prints out the SXML after the parsing, with all the XML namespaces ids and URIs. The file daml-parse-unparse.scm file unparses the SXML tree back into XML. The code then verifies that we can parse that unparsed document again. Furthermore, the code tests that the following invariant holds: PARSE . UNPARSE . PARSE === PARSE An advanced example of SXSLT transformations -------------------------------------------- Target name: sxslt-advanced The example transforms an SHTML-like SXML document with nested sections into a web page with hierarchically numbered and marked up sections and subsections. The example also generates a hierarchical table of contents. This example is due to Jim Bender. The example is aimed to be illustrative of various SXSLT facilities and idioms. In particular, we demonstrate: higher-order tags, pre-order and post-order transformations, re-writing of SXML elements in regular and special ways, context-sensitive applications of re-writing rules. Finally, we illustrate SXSLT reflection: an ability of a rule to query its own stylesheet and to re-apply the stylesheet with "modifications". The file sxslt-advanced.scm should be more properly called a tutorial, with a few occasional lines of code. Parent pointers in SXML trees ----------------------------- Target name: parent-pointers This is the source code for an article 'On parent pointers in SXML trees' posted on the SSAX-SXML mailing list. The code implements and tests four different techniques of determining the parent of an SXML node. The code includes two custom instantiations of the SSAX framework, to add various kinds of parent "pointer" annotations to SXML nodes as we parse an XML document. SSAX parsing with limited XML doctype validation and datatype conversion ------------------------------------------------------------------------ Target name: validate-doctype-simple The present test instantiates the SSAX parser framework to support a limited document type validation and datatype conversion. We demonstrate atomic content validation and structural validation. To be precise, if the content model of an element is declared to be 'bool or 'int, we make sure the corresponding character data in the source XML document indeed represent a boolean or an integer. If the content validates, we convert it to a boolean or an integer value for the resulting SXML. We also show validation of a simple structural constraint, the sequence of child elements matching a user-specified template.