Java Code Examples for org.jsoup.select.Selector#SelectorParseException

The following examples show how to use org.jsoup.select.Selector#SelectorParseException . You can vote up the ones you like or vote down the ones you don't like, and go to the original project or source file by following the links above each example. You may check out the related API usage on the sidebar.
Example 1
Source Project: localization_nifi   File: AbstractHTMLProcessor.java    License: Apache License 2.0 6 votes vote down vote up 
@Override
public ValidationResult validate(final String subject, final String value, final ValidationContext context) {
    if (context.isExpressionLanguageSupported(subject) && context.isExpressionLanguagePresent(value)) {
        return new ValidationResult.Builder().subject(subject).input(value).explanation("Expression Language Present").valid(true).build();
    }

    String reason = null;
    try {
        Document doc = Jsoup.parse("<html></html>");
        doc.select(value);
    } catch (final Selector.SelectorParseException e) {
        reason = "\"" + value + "\" is an invalid CSS selector";
    }

    return new ValidationResult.Builder().subject(subject).input(value).explanation(reason).valid(reason == null).build();
}
Example 2
Source Project: zongtui-webcrawler   File: XPathParser.java    License: GNU General Public License v2.0 6 votes vote down vote up 
private void findElements() {
    if (tq.matches("@")) {
        consumeAttribute();
    } else if (tq.matches("*")) {
        allElements();
    } else if (tq.matchesRegex("\\w+\\(.*\\).*")) {
        consumeOperatorFunction();
    } else if (tq.matchesWord()) {
        byTag();
    } else if (tq.matchesRegex("\\[\\d+\\]")) {
        byNth();
    } else if (tq.matches("[")) {
        evals.add(consumePredicates(tq.chompBalanced('[', ']')));
    } else {
        // unhandled
        throw new Selector.SelectorParseException("Could not parse query '%s': unexpected token at '%s'", query, tq.remainder());
    }

}
Example 3
Source Project: zongtui-webcrawler   File: XPathParser.java    License: GNU General Public License v2.0 6 votes vote down vote up 
private Evaluator byFunction(XTokenQueue predicatesQueue) {
    for (Map.Entry<String, FunctionEvaluator> entry : FUNCTION_MAPPING.entrySet()) {
        if (predicatesQueue.matchChomp(entry.getKey())) {
            String paramString = predicatesQueue.chompBalanced('(', ')');
            List<String> params = XTokenQueue.trimQuotes(XTokenQueue.parseFuncionParams(paramString));

            if (params.get(0).startsWith("@")) {
                params.set(0, params.get(0).substring(1));
                return entry.getValue().call(params.toArray(new String[0]));
            } else {
                return null;
            }
        }
    }

    throw new Selector.SelectorParseException("Could not parse query '%s': unexpected token at '%s'", query, predicatesQueue.remainder());
}
Example 4
Source Project: zongtui-webcrawler   File: XPathParser.java    License: GNU General Public License v2.0 6 votes vote down vote up 
private void functionRegex(String remainder) {
    Validate.isTrue(remainder.endsWith(")"), "Unclosed bracket for function! " + remainder);
    List<String> params = XTokenQueue.trimQuotes(XTokenQueue.parseFuncionParams(remainder.substring("regex(".length(), remainder.length() - 1)));
    if (params.size() == 1) {
        elementOperator = new ElementOperator.Regex(params.get(0));
    } else if (params.size() == 2) {
        if (params.get(0).startsWith("@")) {
            elementOperator = new ElementOperator.Regex(params.get(1), params.get(0).substring(1));
        } else {
            elementOperator = new ElementOperator.Regex(params.get(0), null, Integer.parseInt(params.get(1)));
        }
    } else if (params.size() == 3) {
        elementOperator = new ElementOperator.Regex(params.get(1), params.get(0).substring(1), Integer.parseInt(params.get(2)));
    } else {
        throw new Selector.SelectorParseException("Unknown usage for regex()" + remainder);
    }
}
Example 5
Source Project: nifi   File: AbstractHTMLProcessor.java    License: Apache License 2.0 6 votes vote down vote up 
@Override
public ValidationResult validate(final String subject, final String value, final ValidationContext context) {
    if (context.isExpressionLanguageSupported(subject) && context.isExpressionLanguagePresent(value)) {
        return new ValidationResult.Builder().subject(subject).input(value).explanation("Expression Language Present").valid(true).build();
    }

    String reason = null;
    try {
        Document doc = Jsoup.parse("<html></html>");
        doc.select(value);
    } catch (final Selector.SelectorParseException e) {
        reason = "\"" + value + "\" is an invalid CSS selector";
    }

    return new ValidationResult.Builder().subject(subject).input(value).explanation(reason).valid(reason == null).build();
}
Example 6
Source Project: xsoup   File: XPathParser.java    License: MIT License 6 votes vote down vote up 
private void findElements() {
    if (tq.matches("@")) {
        consumeAttribute();
    } else if (tq.matches("*")) {
        allElements();
    } else if (tq.matchesRegex("\\w+\\(.*\\).*")) {
        consumeOperatorFunction();
    } else if (tq.matchesWord()) {
        byTag();
    } else if (tq.matchesRegex("\\[\\d+\\]")) {
        byNth();
    } else if (tq.matches("[")) {
        evals.add(consumePredicates(tq.chompBalanced('[', ']')));
    } else {
        // unhandled
        throw new Selector.SelectorParseException("Could not parse query '%s': unexpected token at '%s'", query, tq.remainder());
    }

}
Example 7
Source Project: xsoup   File: XPathParser.java    License: MIT License 6 votes vote down vote up 
private Evaluator byFunction(XTokenQueue predicatesQueue) {
    for (Map.Entry<String, FunctionEvaluator> entry : FUNCTION_MAPPING.entrySet()) {
        if (predicatesQueue.matchChomp(entry.getKey())) {
            String paramString = predicatesQueue.chompBalanced('(', ')');
            List<String> params = XTokenQueue.trimQuotes(XTokenQueue.parseFuncionParams(paramString));

            if (params.get(0).startsWith("@")) {
                params.set(0, params.get(0).substring(1));
                return entry.getValue().call(params.toArray(new String[0]));
            } else {
                return null;
            }
        }
    }

    throw new Selector.SelectorParseException("Could not parse query '%s': unexpected token at '%s'", query, predicatesQueue.remainder());
}
Example 8
Source Project: xsoup   File: XPathParser.java    License: MIT License 6 votes vote down vote up 
private void functionRegex(String remainder) {
    Validate.isTrue(remainder.endsWith(")"), "Unclosed bracket for function! " + remainder);
    List<String> params = XTokenQueue.trimQuotes(XTokenQueue.parseFuncionParams(remainder.substring("regex(".length(), remainder.length() - 1)));
    if (params.size() == 1) {
        elementOperator = new ElementOperator.Regex(params.get(0));
    } else if (params.size() == 2) {
        if (params.get(0).startsWith("@")) {
            elementOperator = new ElementOperator.Regex(params.get(1), params.get(0).substring(1));
        } else {
            elementOperator = new ElementOperator.Regex(params.get(0), null, Integer.parseInt(params.get(1)));
        }
    } else if (params.size() == 3) {
        elementOperator = new ElementOperator.Regex(params.get(1), params.get(0).substring(1), Integer.parseInt(params.get(2)));
    } else {
        throw new Selector.SelectorParseException("Unknown usage for regex()" + remainder);
    }
}
Example 9
Source Project: zongtui-webcrawler   File: XPathParser.java    License: GNU General Public License v2.0 5 votes vote down vote up 
private Evaluator consumePredicates(String queue) {
    XTokenQueue predicatesQueue = new XTokenQueue(queue);
    EvaluatorStack evaluatorStack = new EvaluatorStack();
    Operation currentOperation = null;
    predicatesQueue.consumeWhitespace();
    while (!predicatesQueue.isEmpty()) {
        if (predicatesQueue.matchChomp("and")) {
            currentOperation = Operation.AND;
        } else if (predicatesQueue.matchChomp("or")) {
            currentOperation = Operation.OR;
        } else {
            if (currentOperation == null && evaluatorStack.size() > 0) {
                throw new IllegalArgumentException(String.format("Need AND/OR between two predicate! %s", predicatesQueue.remainder()));
            }
            Evaluator evaluator;
            if (predicatesQueue.matches("(")) {
                evaluator = consumePredicates(predicatesQueue.chompBalanced('(', ')'));
            } else if (predicatesQueue.matches("@")) {
                evaluator = byAttribute(predicatesQueue);
            } else if (predicatesQueue.matchesRegex("\\w+.*")) {
                evaluator = byFunction(predicatesQueue);
            } else {
                throw new Selector.SelectorParseException("Could not parse query '%s': unexpected token at '%s'", query, predicatesQueue.remainder());
            }
            evaluatorStack.calc(evaluator, currentOperation);
            //consume operator
            currentOperation = null;
        }
        predicatesQueue.consumeWhitespace();
    }
    evaluatorStack.mergeOr();
    return evaluatorStack.peek();
}
Example 10
Source Project: xsoup   File: XPathParser.java    License: MIT License 5 votes vote down vote up 
private Evaluator consumePredicates(String queue) {
    XTokenQueue predicatesQueue = new XTokenQueue(queue);
    EvaluatorStack evaluatorStack = new EvaluatorStack();
    Operation currentOperation = null;
    predicatesQueue.consumeWhitespace();
    while (!predicatesQueue.isEmpty()) {
        if (predicatesQueue.matchChomp("and")) {
            currentOperation = Operation.AND;
        } else if (predicatesQueue.matchChomp("or")) {
            currentOperation = Operation.OR;
        } else {
            if (currentOperation == null && evaluatorStack.size() > 0) {
                throw new IllegalArgumentException(String.format("Need AND/OR between two predicate! %s", predicatesQueue.remainder()));
            }
            Evaluator evaluator;
            if (predicatesQueue.matches("(")) {
                evaluator = consumePredicates(predicatesQueue.chompBalanced('(', ')'));
            } else if (predicatesQueue.matches("@")) {
                evaluator = byAttribute(predicatesQueue);
            } else if (predicatesQueue.matchesRegex("\\w+.*")) {
                evaluator = byFunction(predicatesQueue);
            } else {
                throw new Selector.SelectorParseException("Could not parse query '%s': unexpected token at '%s'", query, predicatesQueue.remainder());
            }
            evaluatorStack.calc(evaluator, currentOperation);
            //consume operator
            currentOperation = null;
        }
        predicatesQueue.consumeWhitespace();
    }
    evaluatorStack.mergeOr();
    return evaluatorStack.peek();
}
Example 11
Source Project: localization_nifi   File: TestGetHTMLElement.java    License: Apache License 2.0 4 votes vote down vote up 
@Test(expected = Selector.SelectorParseException.class)
public void testCSSSelectorSyntaxValidator() throws IOException {
    Document doc = Jsoup.parse(new File("src/test/resources/Weather.html"), StandardCharsets.UTF_8.name());
    doc.select("---invalidCssSelector");
}
Example 12
Source Project: zongtui-webcrawler   File: XPathParser.java    License: GNU General Public License v2.0 4 votes vote down vote up 
private Evaluator byAttribute(XTokenQueue cq) {
    cq.matchChomp("@");
    String key = cq.consumeToAny("=", "!=", "^=", "$=", "*=", "~="); // eq, not, start, end, contain, match, (no val)
    Validate.notEmpty(key);
    cq.consumeWhitespace();
    Evaluator evaluator;
    if (cq.isEmpty()) {
        if ("*".equals(key)) {
            evaluator = new XEvaluators.HasAnyAttribute();
        } else {
            evaluator = new Evaluator.Attribute(key);
        }
    } else {
        if (cq.matchChomp("=")) {
            String value = chompEqualValue(cq);
            //to support select one class out of all
            if (key.equals("class")) {
                String className = XTokenQueue.trimQuotes(value);
                if (!className.contains(" ")) {
                    evaluator = new Evaluator.Class(className);
                } else {
                    evaluator = new Evaluator.AttributeWithValue(key, className);
                }
            } else {
                evaluator = new Evaluator.AttributeWithValue(key, XTokenQueue.trimQuotes(value));
            }
        } else if (cq.matchChomp("!="))
            evaluator = new Evaluator.AttributeWithValueNot(key, XTokenQueue.trimQuotes(chompEqualValue(cq)));

        else if (cq.matchChomp("^="))
            evaluator = new Evaluator.AttributeWithValueStarting(key, XTokenQueue.trimQuotes(chompEqualValue(cq)));

        else if (cq.matchChomp("$="))
            evaluator = new Evaluator.AttributeWithValueEnding(key, XTokenQueue.trimQuotes(chompEqualValue(cq)));

        else if (cq.matchChomp("*="))
            evaluator = new Evaluator.AttributeWithValueContaining(key, XTokenQueue.trimQuotes(chompEqualValue(cq)));

        else if (cq.matchChomp("~="))
            evaluator = new Evaluator.AttributeWithValueMatching(key, Pattern.compile(XTokenQueue.trimQuotes(chompEqualValue(cq))));
        else
            throw new Selector.SelectorParseException("Could not parse attribute query '%s': unexpected token at '%s'", query, chompEqualValue(cq));
    }
    return evaluator;
}
Example 13
Source Project: nifi   File: TestGetHTMLElement.java    License: Apache License 2.0 4 votes vote down vote up 
@Test(expected = Selector.SelectorParseException.class)
public void testCSSSelectorSyntaxValidator() throws IOException {
    Document doc = Jsoup.parse(new File("src/test/resources/Weather.html"), StandardCharsets.UTF_8.name());
    doc.select("---invalidCssSelector");
}
Example 14
Source Project: xsoup   File: XPathParser.java    License: MIT License 4 votes vote down vote up 
private Evaluator byAttribute(XTokenQueue cq) {
    cq.matchChomp("@");
    String key = cq.consumeToAny("=", "!=", "^=", "$=", "*=", "~="); // eq, not, start, end, contain, match, (no val)
    Validate.notEmpty(key);
    cq.consumeWhitespace();
    Evaluator evaluator;
    if (cq.isEmpty()) {
        if ("*".equals(key)) {
            evaluator = new XEvaluators.HasAnyAttribute();
        } else {
            evaluator = new Evaluator.Attribute(key);
        }
    } else {
        if (cq.matchChomp("=")) {
            String value = chompEqualValue(cq);
            //to support select one class out of all
            if (key.equals("class")) {
                String className = XTokenQueue.trimQuotes(value);
                if (!className.contains(" ")) {
                    evaluator = new Evaluator.Class(className);
                } else {
                    evaluator = new Evaluator.AttributeWithValue(key, className);
                }
            } else {
                evaluator = new Evaluator.AttributeWithValue(key, XTokenQueue.trimQuotes(value));
            }
        } else if (cq.matchChomp("!="))
            evaluator = new Evaluator.AttributeWithValueNot(key, XTokenQueue.trimQuotes(chompEqualValue(cq)));

        else if (cq.matchChomp("^="))
            evaluator = new Evaluator.AttributeWithValueStarting(key, XTokenQueue.trimQuotes(chompEqualValue(cq)));

        else if (cq.matchChomp("$="))
            evaluator = new Evaluator.AttributeWithValueEnding(key, XTokenQueue.trimQuotes(chompEqualValue(cq)));

        else if (cq.matchChomp("*="))
            evaluator = new Evaluator.AttributeWithValueContaining(key, XTokenQueue.trimQuotes(chompEqualValue(cq)));

        else if (cq.matchChomp("~="))
            evaluator = new Evaluator.AttributeWithValueMatching(key, Pattern.compile(XTokenQueue.trimQuotes(chompEqualValue(cq))));
        else
            throw new Selector.SelectorParseException("Could not parse attribute query '%s': unexpected token at '%s'", query, chompEqualValue(cq));
    }
    return evaluator;
}