This repository has been archived on 2023-08-20. You can view files and clone it, but cannot push or open issues or pull requests.
yap-6.3/packages/xml/xml_example/xml_example.pl

388 lines
12 KiB
Perl
Raw Normal View History

2012-02-14 12:41:12 +00:00
/* Using xml.pl to solve XML Query Cases - An Example
*
* The following is a complete example to illustrate how the module can be used;
* it exercises both the input and output parsing modes of xml_parse/[2,3], and
* illustrates the use of xml_subterm/2 to access the nodes of a "document value
* model". It's written for Quintus Prolog, but should port to other Prologs
* easily.
*
* The entry-point of the program is the test/1 predicate.
*
* test( +QueryId ) executes a Prolog implementation of a Query from Use Case
* "XMP": Experiences and Exemplars, in the W3C's XML Query Use Cases, which
* "contains several example queries that illustrate requirements gathered from
* the database and document communities".
* <http://www.w3.org/TR/2002/WD-xmlquery-use-cases-20021115/#xmp>
*
* QueryId is one of q1<EFBFBD>q12 selecting which of the 12 use cases is executed.
* The XML output is written to the file [QueryId].xml in the current directory.
*
* xml_pp/1 is used to display the resulting "document value model"
% data-structures on the user output (stdout) stream.
*/
:- use_module(library(lists),[append/3]).
test( Query ) :-
xml_query( Query, ResultElement ),
% Parse output XML into the Output chars
xml_parse( Output, xml([], [ResultElement]) ),
absolute_file_name( Query, [extensions([xml])], OutputFile ),
% Write OutputFile from the Output list of chars
tell( OutputFile ),
put_chars( Output ),
told,
% Pretty print OutputXML
write( 'Output XML' ), nl,
xml_pp( xml([], [ResultElement]) ).
/* xml_query( +QueryNo, ?OutputXML ) when OutputXML is an XML Document Value Model
* produced by running an example taken, identified by QueryNo from the XML Query
* "XMP" use case.
*/
% Q1: List books published by Addison-Wesley after 1991, including their year and
% title.
xml_query( q1, element(bib, [], Books) ) :-
element_name( Title, title ),
element_name( Publisher, publisher ),
input_document( 'bib.xml', Bibliography ),
findall(
element(book, [year=Year], [Title]),
(
xml_subterm( Bibliography, element(book, Attributes, Content) ),
xml_subterm( Content, Publisher ),
xml_subterm( Publisher, Text ),
text_value( Text, "Addison-Wesley" ),
member( year=Year, Attributes ),
number_codes( YearNo, Year ),
YearNo > 1991,
xml_subterm( Content, Title )
),
Books
).
% Q2: Create a flat list of all the title-author pairs, with each pair enclosed
% in a "result" element.
xml_query( q2, element(results, [], Results) ) :-
element_name( Title, title ),
element_name( Author, author ),
element_name( Book, book ),
input_document( 'bib.xml', Bibliography ),
findall(
element(result, [], [Title,Author]),
(
xml_subterm( Bibliography, Book ),
xml_subterm( Book, Title ),
xml_subterm( Book, Author )
),
Results
).
% Q3: For each book in the bibliography, list the title and authors, grouped
% inside a "result" element.
xml_query( q3, element(results, [], Results) ) :-
element_name( Title, title ),
element_name( Author, author ),
element_name( Book, book ),
input_document( 'bib.xml', Bibliography ),
findall(
element(result, [], [Title|Authors]),
(
xml_subterm( Bibliography, Book ),
xml_subterm( Book, Title ),
findall( Author, xml_subterm(Book, Author), Authors )
),
Results
).
% Q4: For each author in the bibliography, list the author's name and the titles
% of all books by that author, grouped inside a "result" element.
xml_query( q4, element(results, [], Results) ) :-
element_name( Title, title ),
element_name( Author, author ),
element_name( Book, book ),
input_document( 'bib.xml', Bibliography ),
findall( Author, xml_subterm(Bibliography, Author), AuthorBag ),
sort( AuthorBag, Authors ),
findall(
element(result, [], [Author|Titles]),
(
member( Author, Authors ),
findall( Title, (
xml_subterm( Bibliography, Book ),
xml_subterm( Book, Author ),
xml_subterm( Book, Title )
),
Titles
)
),
Results
).
% Q5: For each book found at both bn.com and amazon.com, list the title of the
% book and its price from each source.
xml_query( q5, element('books-with-prices', [], BooksWithPrices) ) :-
element_name( Title, title ),
element_name( Book, book ),
element_name( Review, entry ),
input_document( 'bib.xml', Bibliography ),
input_document( 'reviews.xml', Reviews ),
findall(
element('book-with-prices', [], [
Title,
element('price-bn',[], BNPrice ),
element('price-amazon',[], AmazonPrice )
] ),
(
xml_subterm( Bibliography, Book ),
xml_subterm( Book, Title ),
xml_subterm( Reviews, Review ),
xml_subterm( Review, Title ),
xml_subterm( Book, element(price,_, BNPrice) ),
xml_subterm( Review, element(price,_, AmazonPrice) )
),
BooksWithPrices
).
% Q6: For each book that has at least one author, list the title and first two
% authors, and an empty "et-al" element if the book has additional authors.
xml_query( q6, element(bib, [], Results) ) :-
element_name( Title, title ),
element_name( Author, author ),
element_name( Book, book ),
input_document( 'bib.xml', Bibliography ),
findall(
element(book, [], [Title,FirstAuthor|Authors]),
(
xml_subterm( Bibliography, Book ),
xml_subterm( Book, Title ),
findall( Author, xml_subterm(Book, Author), [FirstAuthor|Others] ),
other_authors( Others, Authors )
),
Results
).
% Q7: List the titles and years of all books published by Addison-Wesley after
% 1991, in alphabetic order.
xml_query( q7, element(bib, [], Books) ) :-
element_name( Title, title ),
element_name( Publisher, publisher ),
input_document( 'bib.xml', Bibliography ),
findall(
Title-element(book, [year=Year], [Title]),
(
xml_subterm( Bibliography, element(book, Attributes, Book) ),
xml_subterm( Book, Publisher ),
xml_subterm( Publisher, Text ),
text_value( Text, "Addison-Wesley" ),
member( year=Year, Attributes ),
number_codes( YearNo, Year ),
YearNo > 1991,
xml_subterm( Book, Title )
),
TitleBooks
),
keysort( TitleBooks, TitleBookSet ),
range( TitleBookSet, Books ).
% Q8: Find books in which the name of some element ends with the string "or" and
% the same element contains the string "Suciu" somewhere in its content. For each
% such book, return the title and the qualifying element.
xml_query( q8, element(bib, [], Books) ) :-
element_name( Title, title ),
element_name( Book, book ),
element_name( QualifyingElement, QualifyingName ),
append( "Suciu", _Back, Suffix ),
input_document( 'bib.xml', Bibliography ),
findall(
element(book, [], [Title,QualifyingElement]),
(
xml_subterm( Bibliography, Book ),
xml_subterm( Book, QualifyingElement ),
atom_codes( QualifyingName, QNChars ),
append( _QNPrefix, "or", QNChars ),
xml_subterm( QualifyingElement, TextItem ),
text_value( TextItem, TextValue ),
append( _Prefix, Suffix, TextValue ),
xml_subterm( Book, Title )
),
Books
).
% Q9: In the document "books.xml", find all section or chapter titles that
% contain the word "XML", regardless of the level of nesting.
xml_query( q9, element(results, [], Titles) ) :-
element_name( Title, title ),
append( "XML", _Back, Suffix ),
input_document( 'books.xml', Books ),
findall(
Title,
(
xml_subterm( Books, Title ),
xml_subterm( Title, TextItem ),
text_value( TextItem, TextValue ),
append( _Prefix, Suffix, TextValue )
),
Titles
).
% Q10: In the document "prices.xml", find the minimum price for each book, in the
% form of a "minprice" element with the book title as its title attribute.
xml_query( q10, element(results, [], MinPrices) ) :-
element_name( Title, title ),
element_name( Price, price ),
input_document( 'prices.xml', Prices ),
findall( Title, xml_subterm(Prices, Title), TitleBag ),
sort( TitleBag, TitleSet ),
element_name( Book, book ),
findall(
element(minprice, [title=TitleString], [MinPrice]),
(
member( Title, TitleSet ),
xml_subterm( Title, TitleText ),
text_value( TitleText, TitleString ),
findall( PriceValue-Price, (
xml_subterm( Prices, Book ),
xml_subterm( Book, Title ),
xml_subterm( Book, Price ),
xml_subterm( Price, Text ),
text_value( Text, PriceChars ),
number_codes( PriceValue, PriceChars )
),
PriceValues
),
minimum( PriceValues, PriceValue-MinPrice )
),
MinPrices
).
% Q11: For each book with an author, return the book with its title and authors.
% For each book with an editor, return a reference with the book title and the
% editor's affiliation.
xml_query( q11, element(bib, [], Results) ) :-
element_name( Title, title ),
element_name( Author, author ),
element_name( Book, book ),
element_name( Editor, editor ),
element_name( Affiliation, affiliation ),
input_document( 'bib.xml', Bibliography ),
findall(
element(book, [], [Title,FirstAuthor|Authors]),
(
xml_subterm( Bibliography, Book ),
xml_subterm( Book, Title ),
findall( Author, xml_subterm(Book, Author), [FirstAuthor|Authors] )
),
Books
),
findall(
element(reference, [], [Title,Affiliation]),
(
xml_subterm( Bibliography, Book ),
xml_subterm( Book, Title ),
xml_subterm( Book, Editor ),
xml_subterm( Editor, Affiliation )
),
References
),
append( Books, References, Results ).
% Q12: Find pairs of books that have different titles but the same set of authors
% (possibly in a different order).
xml_query( q12, element(bib, [], Pairs) ) :-
element_name( Author, author ),
element_name( Book1, book ),
element_name( Book2, book ),
element_name( Title1, title ),
element_name( Title2, title ),
input_document( 'bib.xml', Bibliography ),
findall(
element('book-pair', [], [Title1,Title2]),
(
xml_subterm( Bibliography, Book1 ),
findall( Author, xml_subterm(Book1, Author), AuthorBag1 ),
sort( AuthorBag1, AuthorSet ),
xml_subterm( Bibliography, Book2 ),
Book2 @< Book1,
findall( Author, xml_subterm(Book2, Author), AuthorBag2 ),
sort( AuthorBag2, AuthorSet ),
xml_subterm( Book1, Title1 ),
xml_subterm( Book2, Title2 )
),
Pairs
).
% Auxilliary Predicates
other_authors( [], [] ).
other_authors( [Author|Authors], [Author|EtAl] ) :-
et_al( Authors, EtAl ).
et_al( [], [] ).
et_al( [_|_], [element('et-al',[],[])] ).
text_value( [pcdata(Text)], Text ).
text_value( [cdata(Text)], Text ).
element_name( element(Name, _Attributes, _Content), Name ).
/* range( +Pairs, ?Range ) when Pairs is a list of key-datum pairs and Range
* is the list of data.
*/
range( [], [] ).
range( [_Key-Datum|Pairs], [Datum|Data] ) :-
range( Pairs, Data ).
/* minimum( +List, ?Min ) is true if Min is the least member of List in the
* standard order.
*/
minimum( [H|T], Min ):-
minimum1( T, H, Min ).
minimum1( [], Min, Min ).
minimum1( [H|T], Min0, Min ) :-
compare( Relation, H, Min0 ),
minimum2( Relation, H, Min0, T, Min ).
minimum2( '=', Min0, Min0, T, Min ) :-
minimum1( T, Min0, Min ).
minimum2( '<', Min0, _Min1, T, Min ) :-
minimum1( T, Min0, Min ).
minimum2( '>', _Min0, Min1, T, Min ) :-
minimum1( T, Min1, Min ).
/* input_document( +File, ?XML ) reads File and parses the input into the
* "Document Value Model" XML.
*/
input_document( File, XML ) :-
% Read InputFile as a list of chars
see( File ),
get_chars( Input ),
seen,
% Parse the Input chars into the term XML
xml_parse( Input, XML ).
% Load the XML module.
:- use_module( library(xml) ).
% Load a small library of utilities.
:- ensure_loaded( misc ).