net.sf.saxon.expr.Atomizer Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of Saxon-HE Show documentation
Show all versions of Saxon-HE Show documentation
The XSLT and XQuery Processor
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
// Copyright (c) 2018-2022 Saxonica Limited
// This Source Code Form is subject to the terms of the Mozilla Public License, v. 2.0.
// If a copy of the MPL was not distributed with this file, You can obtain one at http://mozilla.org/MPL/2.0/.
// This Source Code Form is "Incompatible With Secondary Licenses", as defined by the Mozilla Public License, v. 2.0.
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
package net.sf.saxon.expr;
import net.sf.saxon.event.ReceiverOption;
import net.sf.saxon.expr.instruct.Block;
import net.sf.saxon.expr.instruct.Choose;
import net.sf.saxon.expr.instruct.TerminationException;
import net.sf.saxon.expr.instruct.ValueOf;
import net.sf.saxon.expr.parser.*;
import net.sf.saxon.functions.Error;
import net.sf.saxon.ma.arrays.ArrayItemType;
import net.sf.saxon.ma.map.MapType;
import net.sf.saxon.om.*;
import net.sf.saxon.pattern.NameTest;
import net.sf.saxon.pattern.NodeKindTest;
import net.sf.saxon.pattern.NodeTest;
import net.sf.saxon.trace.ExpressionPresenter;
import net.sf.saxon.trans.XPathException;
import net.sf.saxon.tree.iter.AtomizingIterator;
import net.sf.saxon.tree.iter.EmptyIterator;
import net.sf.saxon.tree.iter.UntypedAtomizingIterator;
import net.sf.saxon.type.*;
import net.sf.saxon.value.AtomicValue;
import net.sf.saxon.value.Cardinality;
import net.sf.saxon.value.EmptySequence;
/**
* An Atomizer is an expression corresponding essentially to the fn:data() function: it
* maps a sequence by replacing nodes with their typed values
*/
public final class Atomizer extends UnaryExpression {
private boolean untyped = false; //set to true if it is known that the nodes being atomized will be untyped
private boolean singleValued = false; // set to true if all atomized nodes will atomize to a single atomic value
private ItemType operandItemType = null;
private RoleDiagnostic roleDiagnostic = null;
/**
* Constructor
*
* @param sequence the sequence to be atomized
* @param role (may be null) additional information for use in diagnostics
*/
public Atomizer(Expression sequence, RoleDiagnostic role) {
super(sequence);
this.roleDiagnostic = role;
sequence.setFlattened(true);
}
/**
* Make an atomizer with a given operand
*
* @param sequence the operand
* @param role (may be null) additional information for diagnostics
* @return an Atomizer that atomizes the given operand, or another expression that returns the same result
*/
public static Expression makeAtomizer(Expression sequence, RoleDiagnostic role) {
if (sequence instanceof Literal && ((Literal) sequence).getGroundedValue() instanceof AtomicSequence) {
return sequence;
} else {
return new Atomizer(sequence, role);
}
}
@Override
protected OperandRole getOperandRole() {
return OperandRole.ATOMIC_SEQUENCE;
}
/**
* An implementation of Expression must provide at least one of the methods evaluateItem(), iterate(), or process().
* This method indicates which of these methods is provided directly. The other methods will always be available
* indirectly, using an implementation that relies on one of the other methods.
*
* @return the implementation method, for example {@link #ITERATE_METHOD} or {@link #EVALUATE_METHOD} or
* {@link #PROCESS_METHOD}
*/
@Override
public int getImplementationMethod() {
return ITERATE_METHOD | WATCH_METHOD;
}
public ItemType getOperandItemType() {
if (operandItemType == null) {
operandItemType = getBaseExpression().getItemType();
}
return operandItemType;
}
public void setRoleDiagnostic(RoleDiagnostic role) {
this.roleDiagnostic = role;
}
/**
* Simplify an expression
*
*/
/*@NotNull*/
@Override
public Expression simplify() throws XPathException {
//untyped = !getContainer().getPackageData().isSchemaAware();
untyped = !getPackageData().isSchemaAware();
computeSingleValued(getConfiguration().getTypeHierarchy());
Expression operand = getBaseExpression().simplify();
if (operand instanceof Literal) {
GroundedValue val = ((Literal) operand).getGroundedValue();
if (val instanceof AtomicValue) {
return operand;
}
SequenceIterator iter = val.iterate();
Item i;
while ((i = iter.next()) != null) {
if (i instanceof NodeInfo) {
return this;
}
if (i instanceof Function) {
if (((Function)i).isArray()) {
return this;
} else if (((Function)i).isMap()) {
XPathException err = new XPathException(
expandMessage("Cannot atomize a map (" + i.toShortString() + ")"), "FOTY0013");
err.setIsTypeError(true);
err.setLocation(getLocation());
throw err;
} else {
XPathException err = new XPathException(
expandMessage("Cannot atomize a function item"), "FOTY0013");
err.setIsTypeError(true);
err.setLocation(getLocation());
throw err;
}
}
}
// if all items in the sequence are atomic (they generally will be, since this is
// done at compile time), then return the sequence
return operand;
} else if (operand instanceof ValueOf &&
!ReceiverOption.contains(((ValueOf) operand).getOptions(), ReceiverOption.DISABLE_ESCAPING)) {
// XSLT users tend to use ValueOf unnecessarily
return ((ValueOf) operand).convertToCastAsString();
}
setBaseExpression(operand);
return this;
}
/**
* Type-check the expression
*/
/*@NotNull*/
@Override
public Expression typeCheck(ExpressionVisitor visitor, ContextItemStaticInfo contextInfo) throws XPathException {
getOperand().typeCheck(visitor, contextInfo);
untyped = untyped | !visitor.getStaticContext().getPackageData().isSchemaAware();
// If the configuration allows typed data, check whether the content type of these particular nodes is untyped
final TypeHierarchy th = visitor.getConfiguration().getTypeHierarchy();
computeSingleValued(th);
resetLocalStaticProperties();
ItemType operandType = getOperandItemType();
if (th.isSubType(operandType, BuiltInAtomicType.ANY_ATOMIC)) {
return getBaseExpression();
}
if (!operandType.isAtomizable(th)) {
XPathException err;
if (operandType instanceof FunctionItemType) {
String thing = operandType instanceof MapType ? "map" : "function item";
err = new XPathException(
expandMessage("Cannot atomize a " + thing), "FOTY0013");
} else {
err = new XPathException(
expandMessage("Cannot atomize an element that is defined in the schema to have element-only content"), "FOTY0012");
}
err.setIsTypeError(true);
err.setLocation(getLocation());
throw err;
}
getBaseExpression().setFlattened(true);
return this;
}
private void computeSingleValued(TypeHierarchy th) {
ItemType operandType = getOperandItemType();
if (th.relationship(operandType, ArrayItemType.ANY_ARRAY_TYPE) != Affinity.DISJOINT) {
singleValued = false;
} else {
singleValued = untyped;
if (!singleValued) {
ItemType nodeType = getBaseExpression().getItemType();
if (nodeType instanceof NodeTest) {
SchemaType st = ((NodeTest) nodeType).getContentType();
if (st == Untyped.getInstance() || st.isAtomicType() || (st.isComplexType() && st != AnyType.getInstance())) {
singleValued = true;
}
if (!nodeType.getUType().overlaps(UType.ELEMENT.union(UType.ATTRIBUTE))) {
singleValued = true;
}
}
}
}
}
/**
* Expand an error message with information about the context in which atomization is taking place
* @param message the message to be expanded
*/
private String expandMessage(String message) {
if (roleDiagnostic == null) {
return message;
} else {
return message + ". Found while atomizing the " + roleDiagnostic.getMessage();
}
}
/**
* Perform optimisation of an expression and its subexpressions.
* This method is called after all references to functions and variables have been resolved
* to the declaration of the function or variable, and after all type checking has been done.
*
* @param visitor an expression visitor
* @param contextInfo the static type of "." at the point where this expression is invoked.
* The parameter is set to null if it is known statically that the context item will be undefined.
* If the type of the context item is not known statically, the argument is set to
* {@link net.sf.saxon.type.Type#ITEM_TYPE}
* @return the original expression, rewritten if appropriate to optimize execution
* @throws net.sf.saxon.trans.XPathException
* if an error is discovered during this phase
* (typically a type error)
*/
/*@NotNull*/
@Override
public Expression optimize(ExpressionVisitor visitor, ContextItemStaticInfo contextInfo) throws XPathException {
Expression exp = super.optimize(visitor, contextInfo);
if (exp == this) {
final TypeHierarchy th = visitor.getConfiguration().getTypeHierarchy();
Expression operand = getBaseExpression();
if (th.isSubType(operand.getItemType(), BuiltInAtomicType.ANY_ATOMIC)) {
return operand;
}
if (operand instanceof ValueOf &&
!ReceiverOption.contains(((ValueOf) operand).getOptions(), ReceiverOption.DISABLE_ESCAPING)) {
// XSLT users tend to use ValueOf unnecessarily
Expression cast = ((ValueOf) operand).convertToCastAsString();
return cast.optimize(visitor, contextInfo);
}
if (operand instanceof LetExpression || operand instanceof ForExpression) {
// replace data(let $x := y return z) by (let $x := y return data(z))
Expression action = ((Assignation) operand).getAction();
((Assignation) operand).setAction(new Atomizer(action, roleDiagnostic));
return operand.optimize(visitor, contextInfo);
}
if (operand instanceof Choose) {
// replace data(if x then y else z) by (if x then data(y) else data(z)
((Choose)operand).atomizeActions();
return operand.optimize(visitor, contextInfo);
}
if (operand instanceof Block) {
// replace data((x,y,z)) by (data(x), data(y), data(z)) as some of the atomizers
// may prove to be redundant. (Also, it helps streaming)
Operand[] children = ((Block) operand).getOperanda();
Expression[] atomizedChildren = new Expression[children.length];
for (int i = 0; i < children.length; i++) {
atomizedChildren[i] = new Atomizer(children[i].getChildExpression(), roleDiagnostic);
}
Block newBlock = new Block(atomizedChildren);
return newBlock.typeCheck(visitor, contextInfo).optimize(visitor, contextInfo);
}
if (untyped && operand instanceof AxisExpression &&
((AxisExpression)operand).getAxis() == AxisInfo.ATTRIBUTE &&
((AxisExpression) operand).getNodeTest() instanceof NameTest &&
!((AxisExpression) operand).isContextPossiblyUndefined()) {
StructuredQName name = ((AxisExpression) operand).getNodeTest().getMatchingNodeName();
FingerprintedQName qName = new FingerprintedQName(name, visitor.getConfiguration().getNamePool());
AttributeGetter ag = new AttributeGetter(qName);
int checks = 0;
if (!(((AxisExpression) operand).getContextItemType() instanceof NodeTest)) {
checks = AttributeGetter.CHECK_CONTEXT_ITEM_IS_NODE;
}
ag.setRequiredChecks(checks);
ExpressionTool.copyLocationInfo(this, ag);
return ag;
}
if (untyped && operand instanceof SimpleStepExpression &&
((SimpleStepExpression) operand).getAxisExpression().getAxis() == AxisInfo.ATTRIBUTE &&
((SimpleStepExpression) operand).getAxisExpression().getNodeTest() instanceof NameTest) {
StructuredQName name = ((SimpleStepExpression) operand).getAxisExpression().getNodeTest().getMatchingNodeName();
FingerprintedQName qName = new FingerprintedQName(name, visitor.getConfiguration().getNamePool());
AttributeGetter ag = new AttributeGetter(qName);
int checks = 0;
if (!(((SimpleStepExpression) operand).getStart().getItemType() instanceof NodeTest)) {
checks = AttributeGetter.CHECK_CONTEXT_ITEM_IS_NODE;
}
ag.setRequiredChecks(checks);
ExpressionTool.copyLocationInfo(this, ag);
return new SlashExpression(((SimpleStepExpression) operand).getStart(), ag);
}
}
return exp;
}
/**
* Ask whether it is known that any nodes in the input will always be untyped
*
* @return true if it is known that all nodes in the input will be untyped
*/
public boolean isUntyped() {
return untyped;
}
/**
* Determine the special properties of this expression
*
* @return {@link StaticProperty#NO_NODES_NEWLY_CREATED}.
*/
@Override
protected int computeSpecialProperties() {
int p = super.computeSpecialProperties();
p &= ~StaticProperty.NODESET_PROPERTIES;
// if (!untyped) {
// p |= StaticProperty.NOT_UNTYPED_ATOMIC;
// }
return p | StaticProperty.NO_NODES_NEWLY_CREATED;
}
/**
* Reset the static properties of the expression to -1, so that they have to be recomputed
* next time they are used.
*/
@Override
public void resetLocalStaticProperties() {
super.resetLocalStaticProperties();
operandItemType = null;
}
/**
* Copy an expression. This makes a deep copy.
*
* @return the copy of the original expression
* @param rebindings variables that need to be rebound
*/
/*@NotNull*/
@Override
public Expression copy(RebindingMap rebindings) {
Atomizer copy = new Atomizer(getBaseExpression().copy(rebindings), roleDiagnostic);
copy.untyped = untyped;
copy.singleValued = singleValued;
ExpressionTool.copyLocationInfo(this, copy);
return copy;
}
@Override
public String getStreamerName() {
return "Atomizer";
}
/**
* Iterate over the sequence of values
*/
/*@NotNull*/
@Override
public SequenceIterator iterate(XPathContext context) throws XPathException {
try {
SequenceIterator base = getBaseExpression().iterate(context);
return getAtomizingIterator(base, untyped && operandItemType instanceof NodeTest);
} catch (TerminationException | Error.UserDefinedXPathException e) {
throw e;
} catch (XPathException e) {
if (roleDiagnostic == null) {
throw e;
} else {
String message = expandMessage(e.getMessage());
XPathException e2 = new XPathException(message, e.getErrorCodeLocalPart(), e.getLocator());
e2.setXPathContext(context);
e2.maybeSetLocation(getLocation());
throw e2;
}
}
}
/**
* Evaluate as an Item. This should only be called if the Atomizer has cardinality zero-or-one,
* which will only be the case if the underlying expression has cardinality zero-or-one.
*/
@Override
public AtomicValue evaluateItem(XPathContext context) throws XPathException {
Item i = getBaseExpression().evaluateItem(context);
if (i == null) {
return null;
} else {
return i.atomize().head();
}
}
/**
* Determine the data type of the items returned by the expression, if possible
*
* @return a value such as Type.STRING, Type.BOOLEAN, Type.NUMBER. For this class, the
* result is always an atomic type, but it might be more specific.
*/
/*@NotNull*/
@Override
public ItemType getItemType() {
operandItemType = getBaseExpression().getItemType();
TypeHierarchy th = getConfiguration().getTypeHierarchy();
return getAtomizedItemType(getBaseExpression(), untyped, th);
}
/**
* Compute the type that will result from atomizing the result of a given expression
*
* @param operand the given expression
* @param alwaysUntyped true if it is known that nodes will always be untyped
* @param th the type hierarchy cache
* @return the item type of the result of evaluating the operand expression, after atomization, or
* xs:error if it is known that atomization will return an error
*/
public static ItemType getAtomizedItemType(Expression operand, boolean alwaysUntyped, TypeHierarchy th) {
ItemType in = operand.getItemType();
if (in.isPlainType()) {
return in;
} else if (in instanceof NodeTest) {
UType kinds = in.getUType();
if (alwaysUntyped) {
// Some node-kinds always have a typed value that's a string
if (STRING_KINDS.subsumes(kinds)) {
return BuiltInAtomicType.STRING;
}
// Some node-kinds are always untyped atomic; some are untypedAtomic provided that the configuration
// is untyped
if (UNTYPED_IF_UNTYPED_KINDS.subsumes(kinds)) {
return BuiltInAtomicType.UNTYPED_ATOMIC;
}
} else {
if (UNTYPED_KINDS.subsumes(kinds)) {
return BuiltInAtomicType.UNTYPED_ATOMIC;
}
}
return in.getAtomizedItemType();
} else if (in instanceof JavaExternalObjectType) {
return in.getAtomizedItemType();
} else if (in instanceof ArrayItemType) {
PlainType ait = ((ArrayItemType)in).getMemberType().getPrimaryType().getAtomizedItemType();
return ait == null ? ErrorType.getInstance() : ait;
} else if (in instanceof FunctionItemType) {
return ErrorType.getInstance();
}
return BuiltInAtomicType.ANY_ATOMIC;
}
/**
* Node kinds whose typed value is always a string
*/
public static final UType STRING_KINDS =
UType.NAMESPACE.union(UType.COMMENT).union(UType.PI);
/**
* Node kinds whose typed value is always untypedAtomic
*/
public static final UType UNTYPED_KINDS =
UType.TEXT.union(UType.DOCUMENT);
/**
* Node kinds whose typed value is untypedAtomic if the configuration is untyped
*/
public static final UType UNTYPED_IF_UNTYPED_KINDS =
UType.TEXT.union(UType.ELEMENT).union(UType.DOCUMENT).union(UType.ATTRIBUTE);
/**
* Determine the static cardinality of the expression
*/
@Override
protected int computeCardinality() {
ItemType in = getOperandItemType();
Expression operand = getBaseExpression();
if (singleValued) {
return operand.getCardinality();
} else if (untyped && in instanceof NodeTest) {
return operand.getCardinality();
} else if (Cardinality.allowsMany(operand.getCardinality())) {
return StaticProperty.ALLOWS_ZERO_OR_MORE;
} else if (in.isPlainType()) {
return operand.getCardinality();
} else if (in instanceof NodeTest) {
SchemaType schemaType = ((NodeTest) in).getContentType();
if (schemaType.isAtomicType()) {
// can return at most one atomic value per node
return operand.getCardinality();
}
}
return StaticProperty.ALLOWS_ZERO_OR_MORE;
}
/**
* Add a representation of this expression to a PathMap. The PathMap captures a map of the nodes visited
* by an expression in a source tree.
* The default implementation of this method assumes that an expression does no navigation other than
* the navigation done by evaluating its subexpressions, and that the subexpressions are evaluated in the
* same context as the containing expression. The method must be overridden for any expression
* where these assumptions do not hold. For example, implementations exist for AxisExpression, ParentExpression,
* and RootExpression (because they perform navigation), and for the doc(), document(), and collection()
* functions because they create a new navigation root. Implementations also exist for PathExpression and
* FilterExpression because they have subexpressions that are evaluated in a different context from the
* calling expression.
*
* @param pathMap the PathMap to which the expression should be added
* @param pathMapNodeSet the PathMapNodeSet to which the paths embodied in this expression should be added
* @return the pathMapNodeSet representing the points in the source document that are both reachable by this
* expression, and that represent possible results of this expression. For an expression that does
* navigation, it represents the end of the arc in the path map that describes the navigation route. For other
* expressions, it is the same as the input pathMapNode.
*/
@Override
public PathMap.PathMapNodeSet addToPathMap(PathMap pathMap, PathMap.PathMapNodeSet pathMapNodeSet) {
PathMap.PathMapNodeSet result = getBaseExpression().addToPathMap(pathMap, pathMapNodeSet);
if (result != null) {
TypeHierarchy th = getConfiguration().getTypeHierarchy();
ItemType operandItemType = getBaseExpression().getItemType();
if (th.relationship(NodeKindTest.ELEMENT, operandItemType) != Affinity.DISJOINT ||
th.relationship(NodeKindTest.DOCUMENT, operandItemType) != Affinity.DISJOINT) {
result.setAtomized();
}
}
return null;
}
/**
* Get an iterator that returns the result of atomizing the sequence delivered by the supplied
* iterator
*
* @param base the supplied iterator, the input to atomization
* @param oneToOne this can safely be set to true if it is known that all nodes in the base sequence will
* be untyped and that there will be no arrays in the base sequence; but it is always OK
* to set it to false.
* @return an iterator that returns atomic values, the result of the atomization
* @throws net.sf.saxon.trans.XPathException
* if a dynamic evaluation error occurs
*/
public static SequenceIterator getAtomizingIterator(SequenceIterator base, boolean oneToOne) throws XPathException {
if (SequenceTool.supportsGetLength(base)) {
int count = SequenceTool.getLength(base);
if (count == 0) {
return EmptyIterator.emptyIterator();
} else if (count == 1) {
Item first = base.next();
return first.atomize().iterate();
}
} else if (base instanceof AtomizedValueIterator) {
return new AxisAtomizingIterator((AtomizedValueIterator)base);
}
if (oneToOne) {
return new UntypedAtomizingIterator(base);
} else {
return new AtomizingIterator(base);
}
}
public static AtomicSequence atomize(Sequence sequence) throws XPathException {
if (sequence instanceof AtomicSequence) {
return (AtomicSequence)sequence;
} else if (sequence instanceof EmptySequence) {
return EmptyAtomicSequence.getInstance();
} else {
SequenceIterator iter = getAtomizingIterator(sequence.iterate(), false);
return new AtomicArray(iter);
}
}
/**
* Diagnostic print of expression structure. The abstract expression tree
* is written to the supplied output destination.
*/
@Override
public String getExpressionName() {
return "data";
}
/**
* The toString() method for an expression attempts to give a representation of the expression
* in an XPath-like form, but there is no guarantee that the syntax will actually be true XPath.
* In the case of XSLT instructions, the toString() method gives an abstracted view of the syntax
*/
@Override
public String toString() {
return "data(" + getBaseExpression().toString() + ")";
}
@Override
public String toShortString() {
return getBaseExpression().toShortString();
}
@Override
protected void emitExtraAttributes(ExpressionPresenter out) {
if (roleDiagnostic != null) {
out.emitAttribute("diag", roleDiagnostic.save());
}
}
}