com.unboundid.scim2.common.utils.Parser Maven / Gradle / Ivy
/*
* Copyright 2015-2024 Ping Identity Corporation
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License (GPLv2 only)
* or the terms of the GNU Lesser General Public License (LGPLv2.1 only)
* as published by the Free Software Foundation.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, see .
*/
package com.unboundid.scim2.common.utils;
import com.fasterxml.jackson.core.JsonParser;
import com.fasterxml.jackson.databind.node.ValueNode;
import com.unboundid.scim2.common.Path;
import com.unboundid.scim2.common.annotations.NotNull;
import com.unboundid.scim2.common.annotations.Nullable;
import com.unboundid.scim2.common.exceptions.BadRequestException;
import com.unboundid.scim2.common.filters.Filter;
import com.unboundid.scim2.common.filters.FilterType;
import java.io.IOException;
import java.io.Reader;
import java.util.LinkedList;
import java.util.Stack;
/**
* A parser for SCIM filter expressions.
*/
public class Parser
{
/**
* A {@code ThreadLocal} parser options object, to configure nonstandard
* settings such as permitting semicolons in attribute names.
*/
@NotNull
private static final ThreadLocal threadLocalOptions
= ThreadLocal.withInitial(ParserOptions::new);
private static final class StringReader extends Reader
{
@NotNull
private final String string;
private int pos;
private int mark;
/**
* Create a new reader.
*
* @param string The string to read from.
*/
private StringReader(@NotNull final String string)
{
this.string = string;
}
/**
* {@inheritDoc}
*/
@Override
public int read()
{
if(pos >= string.length())
{
return -1;
}
return string.charAt(pos++);
}
/**
* Move the current read position back one character.
*/
public void unread()
{
pos--;
}
/**
* {@inheritDoc}
*/
@Override
public boolean ready()
{
return true;
}
/**
* {@inheritDoc}
*/
@Override
public boolean markSupported()
{
return true;
}
/**
* {@inheritDoc}
*/
@Override
public void mark(final int readAheadLimit)
{
mark = pos;
}
/**
* {@inheritDoc}
*/
@Override
public void reset()
{
pos = mark;
}
/**
* {@inheritDoc}
*
* @throws IllegalArgumentException If {@code n} is negative.
*/
@Override
public long skip(final long n) throws IllegalArgumentException
{
if (n < 0L)
{
throw new IllegalArgumentException(
"Attempted to skip a negative number of characters.");
}
long chars = Math.min(string.length() - pos, n);
// The left side of Math.min() is an integer, and 'chars' will always be
// non-negative, so 'chars' can always be represented as an int.
pos += (int) chars;
return chars;
}
/**
* {@inheritDoc}
*/
@Override
public int read(@NotNull final char[] cbuf, final int off, final int len)
{
if(pos >= string.length())
{
return -1;
}
int chars = Math.min(string.length() - pos, len);
System.arraycopy(string.toCharArray(), pos, cbuf, off, chars);
pos += chars;
return chars;
}
/**
* {@inheritDoc}
*/
@Override
public void close()
{
// do nothing.
}
}
/**
* Parse a filter string.
*
* @param filterString The filter string to parse.
*
* @return A parsed SCIM filter.
* @throws BadRequestException If the filter string could not be parsed.
*/
@NotNull
public static Filter parseFilter(@NotNull final String filterString)
throws BadRequestException
{
return readFilter(new StringReader(filterString.trim()), false);
}
/**
* Get the current {@code ParserOptions} within the current thread.
*
* @return The current parser options.
*/
@NotNull
public static ParserOptions getOptions()
{
return Parser.threadLocalOptions.get();
}
/**
* Set new {@code ParserOptions} within the current thread.
*
* NOTE: SCIM server implementations are not guaranteed to support a given option.
* NOTE: These should be reset as soon as they are no longer needed.
*
*
* Set<ParserOption> priorOptions =
* Parser.setOptions(newOptions);
* try
* {
* performWhateverProcessing();
* }
* finally
* {
* Parser.setOptions(priorOptions);
* }
*
*
* @param newOptions The new parser options.
*
* @return The prior parser options.
*/
@NotNull
public static ParserOptions setOptions(
@NotNull final ParserOptions newOptions)
{
ParserOptions priorOptions = Parser.threadLocalOptions.get();
Parser.threadLocalOptions.set(newOptions);
return priorOptions;
}
/**
* Parse a path string.
*
* @param pathString The path string to parse.
*
* @return A parsed SCIM path.
* @throws BadRequestException If the path string could not be parsed.
*/
@NotNull
public static Path parsePath(@Nullable final String pathString)
throws BadRequestException
{
if(pathString == null)
{
return Path.root();
}
final String trimmedPathString = pathString.trim();
if(trimmedPathString.isEmpty())
{
return Path.root();
}
Path path = Path.root();
StringReader reader = new StringReader(trimmedPathString);
if(SchemaUtils.isUrn(trimmedPathString))
{
// The attribute name is prefixed with the schema URN.
// Find the last ":" before any open brackets. Everything to the left is
// the schema URN, everything on the right is the attribute name plus a
// potential value filter.
int j = trimmedPathString.indexOf('[');
int i;
if(j >= 0)
{
i = trimmedPathString.substring(0, j).lastIndexOf(':');
}
else
{
i = trimmedPathString.lastIndexOf(':');
}
String schemaUrn = trimmedPathString.substring(0, i++);
String attributeName =
trimmedPathString.substring(i, trimmedPathString.length());
try
{
path = Path.root(schemaUrn);
}
catch(IllegalArgumentException e)
{
throw BadRequestException.invalidPath(e.getMessage());
}
if(attributeName.isEmpty())
{
// The trailing colon signifies that this is an extension root.
return path;
}
reader = new StringReader(attributeName);
}
String token;
while ((token = readPathToken(reader)) != null)
{
if (token.isEmpty())
{
// the only time this is allowed to occur is if the previous attribute
// had a value filter, in which case, consume the token and move on.
if(path.isRoot() ||
path.getElement(path.size()-1).getValueFilter() == null)
{
final String msg = String.format(
"Attribute name expected at position %d", reader.mark);
throw BadRequestException.invalidPath(msg);
}
}
else
{
String attributeName = token;
Filter valueFilter = null;
try
{
if (attributeName.endsWith("["))
{
// There is a value path.
attributeName =
attributeName.substring(0, attributeName.length() - 1);
valueFilter = readFilter(reader, true);
}
path = path.attribute(attributeName, valueFilter);
}
catch(BadRequestException be)
{
Debug.debugException(be);
final String msg = String.format(
"Invalid value filter: %s", be.getMessage());
throw BadRequestException.invalidPath(msg);
}
catch(Exception e)
{
Debug.debugException(e);
final String msg = String.format(
"Invalid attribute name starting at position %d: %s",
reader.mark, e.getMessage());
throw BadRequestException.invalidPath(msg);
}
}
}
return path;
}
/**
* Read a path token. A token is either:
*
* -
* An attribute name terminated by a period.
*
* -
* An attribute name terminated by an opening brace.
*
*
*
* @param reader The reader to read from.
*
* @return The token at the current position, or {@code null} if the end of
* the input has been reached.
* @throws BadRequestException If the path string could not be parsed.
*/
@Nullable
private static String readPathToken(@NotNull final StringReader reader)
throws BadRequestException
{
ParserOptions options = Parser.getOptions();
reader.mark(0);
int c = reader.read();
StringBuilder b = new StringBuilder();
while(c > 0)
{
if (c == '.')
{
if(reader.pos >= reader.string.length())
{
// There is nothing after the period.
throw BadRequestException.invalidPath(
"Unexpected end of path string");
}
// Terminating period. Consume it and return token.
return b.toString();
}
if (c == '[')
{
// Terminating opening brace. Consume it and return token.
b.append((char)c);
return b.toString();
}
if (c == '-' || c == '_' || c == '$' || Character.isLetterOrDigit(c)
|| options.isExtendedAttributeNameCharacter((char) c))
{
b.append((char) c);
}
else
{
final String msg = String.format(
"Unexpected character '%s' at position %d for token starting at %d",
(char)c, reader.pos - 1, reader.mark);
throw BadRequestException.invalidPath(msg);
}
c = reader.read();
}
if(b.length() > 0)
{
return b.toString();
}
return null;
}
/**
* Read a filter token. A token is either:
*
* -
* An attribute path terminated by a space or an opening parenthesis.
*
* -
* An attribute path terminated by an opening brace.
*
* -
* An operator terminated by a space or an opening parenthesis.
*
* -
* An opening parenthesis.
*
* -
* An closing parenthesis.
*
* -
* An closing brace.
*
*
*
* @param reader The reader to read from.
* @param isValueFilter Whether to read the token for a value filter.
*
* @return The token at the current position, or {@code null} if the end of
* the input has been reached.
* @throws BadRequestException If the filter string could not be parsed.
*/
@Nullable
private static String readFilterToken(@NotNull final StringReader reader,
final boolean isValueFilter)
throws BadRequestException
{
ParserOptions options = Parser.getOptions();
int c;
do
{
// Skip over any leading spaces.
reader.mark(0);
c = reader.read();
}
while(c == ' ');
StringBuilder b = new StringBuilder();
while(c > 0)
{
if (c == ' ')
{
// Terminating space. Consume it and return token.
return b.toString();
}
if (c == '(' || c == ')')
{
if(b.length() > 0)
{
// Do not consume the parenthesis.
reader.unread();
}
else
{
b.append((char)c);
}
return b.toString();
}
if (!isValueFilter && c == '[')
{
// Terminating opening brace. Consume it and return token.
b.append((char)c);
return b.toString();
}
if (isValueFilter && c == ']')
{
if(b.length() > 0)
{
// Do not consume the closing brace.
reader.unread();
}
else
{
b.append((char)c);
}
return b.toString();
}
if (c == '-' || c == '_' || c == '.' || c == ':' || c == '$'
|| Character.isLetterOrDigit(c)
|| options.isExtendedAttributeNameCharacter((char) c))
{
b.append((char)c);
}
else
{
final String msg = String.format(
"Unexpected character '%s' at position %d for token starting at %d",
(char)c, reader.pos - 1, reader.mark);
throw BadRequestException.invalidFilter(msg);
}
c = reader.read();
}
if(b.length() > 0)
{
return b.toString();
}
return null;
}
/**
* Read a filter from the reader.
*
* @param reader The reader to read the filter from.
* @param isValueFilter Whether to read the filter as a value filter.
* @return The parsed filter.
* @throws BadRequestException If the filter string could not be parsed.
*/
@NotNull
private static Filter readFilter(@NotNull final StringReader reader,
final boolean isValueFilter)
throws BadRequestException
{
final Stack outputStack = new Stack();
final Stack precedenceStack = new Stack();
String token;
String previousToken = null;
while((token = readFilterToken(reader, isValueFilter)) != null)
{
if(token.equals("(") && expectsNewFilter(previousToken))
{
precedenceStack.push(token);
}
else if(token.equalsIgnoreCase(FilterType.NOT.getStringValue()) &&
expectsNewFilter(previousToken))
{
// "not" should be followed by an (
String nextToken = readFilterToken(reader, isValueFilter);
if(nextToken == null)
{
throw BadRequestException.invalidFilter(
"Unexpected end of filter string");
}
if(!nextToken.equals("("))
{
final String msg = String.format(
"Expected '(' at position %d", reader.mark);
throw BadRequestException.invalidFilter(msg);
}
precedenceStack.push(token);
}
else if(token.equals(")") && !expectsNewFilter(previousToken))
{
String operator = closeGrouping(precedenceStack, outputStack, false);
if(operator == null)
{
final String msg =
String.format("No opening parenthesis matching closing " +
"parenthesis at position %d", reader.mark);
throw BadRequestException.invalidFilter(msg);
}
if (operator.equalsIgnoreCase(FilterType.NOT.getStringValue()))
{
// Treat "not" the same as "(" except wrap everything in a not filter.
outputStack.push(Filter.not(outputStack.pop()));
}
}
else if(token.equalsIgnoreCase(FilterType.AND.getStringValue()) &&
!expectsNewFilter(previousToken))
{
// and has higher precedence than or.
precedenceStack.push(token);
}
else if(token.equalsIgnoreCase(FilterType.OR.getStringValue()) &&
!expectsNewFilter(previousToken))
{
// pop all the pending ands first before pushing or.
LinkedList andComponents = new LinkedList();
while (!precedenceStack.isEmpty())
{
if (precedenceStack.peek().equalsIgnoreCase(
FilterType.AND.getStringValue()))
{
precedenceStack.pop();
andComponents.addFirst(outputStack.pop());
}
else
{
break;
}
if(!andComponents.isEmpty())
{
andComponents.addFirst(outputStack.pop());
outputStack.push(Filter.and(andComponents));
}
}
precedenceStack.push(token);
}
else if(token.endsWith("[") && expectsNewFilter(previousToken))
{
// This is a complex value filter.
final Path filterAttribute;
try
{
filterAttribute = parsePath(
token.substring(0, token.length() - 1));
}
catch (final BadRequestException e)
{
Debug.debugException(e);
final String msg = String.format(
"Invalid attribute path at position %d: %s",
reader.mark, e.getMessage());
throw BadRequestException.invalidFilter(msg);
}
if(filterAttribute.isRoot())
{
final String msg = String.format(
"Attribute path expected at position %d", reader.mark);
throw BadRequestException.invalidFilter(msg);
}
outputStack.push(Filter.hasComplexValue(
filterAttribute, readFilter(reader, true)));
}
else if(isValueFilter && token.equals("]") &&
!expectsNewFilter(previousToken))
{
break;
}
else if(expectsNewFilter(previousToken))
{
// This must be an attribute path followed by operator and maybe value.
final Path filterAttribute;
try
{
filterAttribute = parsePath(token);
}
catch (final BadRequestException e)
{
Debug.debugException(e);
final String msg = String.format(
"Invalid attribute path at position %d: %s",
reader.mark, e.getMessage());
throw BadRequestException.invalidFilter(msg);
}
if(filterAttribute.isRoot())
{
final String msg = String.format(
"Attribute path expected at position %d", reader.mark);
throw BadRequestException.invalidFilter(msg);
}
String op = readFilterToken(reader, isValueFilter);
if(op == null)
{
throw BadRequestException.invalidFilter(
"Unexpected end of filter string");
}
if (op.equalsIgnoreCase(FilterType.PRESENT.getStringValue()))
{
outputStack.push(Filter.pr(filterAttribute));
}
else
{
ValueNode valueNode;
try
{
// Mark the beginning of the JSON value so we can later reset back
// to this position and skip the actual chars that were consumed
// by Jackson. The Jackson parser is buffered and reads everything
// until the end of string.
reader.mark(0);
ScimJsonFactory scimJsonFactory = (ScimJsonFactory)
JsonUtils.getObjectReader().getFactory();
JsonParser parser = scimJsonFactory.createScimFilterParser(reader);
// The object mapper will return a Java null for JSON null.
// Have to distinguish between reading a JSON null and encountering
// the end of string.
if (parser.getCurrentToken() == null && parser.nextToken() == null)
{
// End of string.
valueNode = null;
}
else
{
valueNode = parser.readValueAsTree();
// This is actually a JSON null. Use NullNode.
if(valueNode == null)
{
valueNode = JsonUtils.getJsonNodeFactory().nullNode();
}
}
// Reset back to the beginning of the JSON value.
reader.reset();
// Skip the number of chars consumed by JSON parser.
reader.skip(parser.getCurrentLocation().getCharOffset());
}
catch (IOException e)
{
final String msg = String.format(
"Invalid comparison value at position %d: %s",
reader.mark, e.getMessage());
throw BadRequestException.invalidFilter(msg);
}
if (valueNode == null)
{
throw BadRequestException.invalidFilter(
"Unexpected end of filter string");
}
if (op.equalsIgnoreCase(FilterType.EQUAL.getStringValue()))
{
outputStack.push(Filter.eq(filterAttribute, valueNode));
} else if (op.equalsIgnoreCase(
FilterType.NOT_EQUAL.getStringValue()))
{
outputStack.push(Filter.ne(filterAttribute, valueNode));
} else if (op.equalsIgnoreCase(
FilterType.CONTAINS.getStringValue()))
{
outputStack.push(Filter.co(filterAttribute, valueNode));
} else if (op.equalsIgnoreCase(
FilterType.STARTS_WITH.getStringValue()))
{
outputStack.push(Filter.sw(filterAttribute, valueNode));
} else if (op.equalsIgnoreCase(
FilterType.ENDS_WITH.getStringValue()))
{
outputStack.push(Filter.ew(filterAttribute, valueNode));
} else if (op.equalsIgnoreCase(
FilterType.GREATER_THAN.getStringValue()))
{
outputStack.push(Filter.gt(filterAttribute, valueNode));
} else if (op.equalsIgnoreCase(
FilterType.GREATER_OR_EQUAL.getStringValue()))
{
outputStack.push(Filter.ge(filterAttribute, valueNode));
} else if (op.equalsIgnoreCase(
FilterType.LESS_THAN.getStringValue()))
{
outputStack.push(Filter.lt(filterAttribute, valueNode));
} else if (op.equalsIgnoreCase(
FilterType.LESS_OR_EQUAL.getStringValue()))
{
outputStack.push(Filter.le(filterAttribute, valueNode));
} else
{
final String msg = String.format(
"Unrecognized attribute operator '%s' at position %d. " +
"Expected: eq,ne,co,sw,ew,pr,gt,ge,lt,le", op, reader.mark);
throw BadRequestException.invalidFilter(msg);
}
}
}
else
{
final String msg = String.format(
"Unexpected character '%s' at position %d", token,
reader.mark);
throw BadRequestException.invalidFilter(msg);
}
previousToken = token;
}
closeGrouping(precedenceStack, outputStack, true);
if(outputStack.isEmpty())
{
throw BadRequestException.invalidFilter(
"Unexpected end of filter string");
}
return outputStack.pop();
}
/**
* Close a grouping of filters enclosed by parenthesis.
*
* @param operators The stack of operators tokens.
* @param output The stack of output tokens.
* @param isAtTheEnd Whether the end of the filter string was reached.
* @return The last operator encountered that signaled the end of the group.
* @throws BadRequestException If the filter string could not be parsed.
*/
@NotNull
private static String closeGrouping(@NotNull final Stack operators,
@NotNull final Stack output,
final boolean isAtTheEnd)
throws BadRequestException
{
String operator = null;
String repeatingOperator = null;
LinkedList components = new LinkedList();
// Iterate over the logical operators on the stack until either there are
// no more operators or an opening parenthesis or not is found.
while (!operators.isEmpty())
{
operator = operators.pop();
if(operator.equals("(") ||
operator.equalsIgnoreCase(FilterType.NOT.getStringValue()))
{
if(isAtTheEnd)
{
throw BadRequestException.invalidFilter(
"Unexpected end of filter string");
}
break;
}
if(repeatingOperator == null)
{
repeatingOperator = operator;
}
if(!operator.equals(repeatingOperator))
{
if(output.isEmpty())
{
throw BadRequestException.invalidFilter(
"Unexpected end of filter string");
}
components.addFirst(output.pop());
if(repeatingOperator.equalsIgnoreCase(FilterType.AND.getStringValue()))
{
output.push(Filter.and(components));
}
else
{
output.push(Filter.or(components));
}
components.clear();
repeatingOperator = operator;
}
if(output.isEmpty())
{
throw BadRequestException.invalidFilter(
"Unexpected end of filter string");
}
components.addFirst(output.pop());
}
if(repeatingOperator != null && !components.isEmpty())
{
if(output.isEmpty())
{
throw BadRequestException.invalidFilter(
"Unexpected end of filter string");
}
components.addFirst(output.pop());
if(repeatingOperator.equalsIgnoreCase(FilterType.AND.getStringValue()))
{
output.push(Filter.and(components));
}
else
{
output.push(Filter.or(components));
}
}
return operator;
}
/**
* Whether a new filter token is expected given the previous token.
*
* @param previousToken The previous filter token.
* @return Whether a new filter token is expected.
*/
private static boolean expectsNewFilter(@Nullable final String previousToken)
{
return previousToken == null ||
previousToken.equals("(") ||
previousToken.equalsIgnoreCase(FilterType.NOT.getStringValue()) ||
previousToken.equalsIgnoreCase(FilterType.AND.getStringValue()) ||
previousToken.equalsIgnoreCase(FilterType.OR.getStringValue());
}
}