This module pre-processes the HTML files that constitute the manual such that we can access the summary documentation of all predicates for usage in IDE tools.

 manual_object(?Object, ?Summary, ?File, ?Class, ?Offset) is nondet

True if Object is documented. Arguments:

Arguments:

Object	- is the object documented, described by a Prolog term. Defined shapes are: section(Level, Number, Label, File) Name/Arity Name//Arity Module:Name/Arity Module:Name//Arity f(Name/Arity c(Name)
Summary	- is a string object providing a summary of object
File	- is the HTML file in which the object is documented
Class	- is one of manual or packages
Offset	- is the character offset at which the DOM element describing Object appears. This is used by doc_man.pl to quickly extract the node.

 clean_man_index is det

Clean already loaded manual index.

 manual_directory(-Class, -Dir)// is nondet[private]

True if Dir is a directory holding manual files. Class is an identifier used by doc_object_summary/4.

 save_man_index

Create swi('doc/manindex.db'), containing an index into the HTML manuals for use with help/1 and similar predicates. This predicate is called during the build process.

 index_manual is det[private]

Load the manual index if not already done.

 read_index(+File)[private]

Read the manual index from File.

 check_duplicate_ids

Maintenance utility to check that we do not have duplicate section identifiers in the documentation.

 index_man_directory(+Dir, +Options) is det

Index the HTML directory Dir. Options are:

class(Class)

Define category of the found objects.

symbolic(+Term)

Symbolic (file search) specification for Dir

Remaining Options are passed to absolute_file_name/3.

 index_man_file(+Class, +File) is det

 index_man_file(+Class, +File, +Symbolic) is det

Collect the documented objects from the SWI-Prolog manual file File.

 index_on_begin(+Element, +Attributes, +Parser) is semidet[private]

Called from sgml_parse/2 in index_man_file/2. Element is the name of the element, Attributes the list of Name=Value pairs of the open attributes. Parser is the parser objects.

 dom_section(+HeaderDOM, -NR, -Title) is semidet[private]

NR is the section number (e.g. 1.1, 1.23) and Title is the title from a section header. The first clauses processes the style information from latex2html, emitting sections as:

<HN> <A name="sec:nr"><span class='sec-nr'>NR</span>|_|
                      <span class='sec-title'>Title</span>

 summary(+DOM, -Summary:string) is det[private]

Summary is the first sentence of DOM.

 dom_to_text(+DOM, -Text)[private]

Extract the text of a parsed HTML term. White-space in the result is normalised. See normalise_white_space//1.

 current_man_object(?Object) is nondet

 man_object_property(?Object, ?Property) is nondet

True when Property is a property of the given manual object. Defined properties are:

summary(-Text)

Summary text for the object.

id(ID)

Return unique id for the text, so we can remove duplicates

 doc_object_identifier(+DocObject, -Identifier) is semidet

True when Identifier is the name of DocObject.