Examples with FuzzyHashMatcher org.apache.nifi.processors.cybersecurity.matchers.FuzzyHashMatcher used on opensource projects

Example 1 with FuzzyHashMatcher

use of org.apache.nifi.processors.cybersecurity.matchers.FuzzyHashMatcher in project nifi by apache.

the class CompareFuzzyHash method onTrigger.

@Override
public void onTrigger(ProcessContext context, ProcessSession session) throws ProcessException {
    FlowFile flowFile = session.get();
    if (flowFile == null) {
        return;
    }
    final ComponentLog logger = getLogger();
    String algorithm = context.getProperty(HASH_ALGORITHM).getValue();
    final String attributeName = context.getProperty(ATTRIBUTE_NAME).getValue();
    String inputHash = flowFile.getAttribute(attributeName);
    if (inputHash == null) {
        getLogger().info("FlowFile {} lacks the required '{}' attribute, routing to failure.", new Object[] { flowFile, attributeName });
        session.transfer(flowFile, REL_FAILURE);
        return;
    }
    FuzzyHashMatcher fuzzyHashMatcher = null;
    switch(algorithm) {
        case tlsh:
            fuzzyHashMatcher = new TLSHHashMatcher(getLogger());
            break;
        case ssdeep:
            fuzzyHashMatcher = new SSDeepHashMatcher(getLogger());
            break;
        default:
            getLogger().error("Seems like the processor is configured to use unsupported algorithm '{}' ? Yielding.", new Object[] { algorithm });
            context.yield();
            return;
    }
    if (fuzzyHashMatcher.isValidHash(inputHash) == false) {
        // and if that is the case we log
        logger.error("Invalid hash provided. Sending to failure");
        // and send to failure
        session.transfer(flowFile, REL_FAILURE);
        session.commit();
        return;
    }
    double similarity = 0;
    double matchThreshold = context.getProperty(MATCH_THRESHOLD).asDouble();
    try {
        Map<String, Double> matched = new ConcurrentHashMap<String, Double>();
        BufferedReader reader = fuzzyHashMatcher.getReader(context.getProperty(HASH_LIST_FILE).getValue());
        String line = null;
        iterateFile: while ((line = reader.readLine()) != null) {
            if (line != null) {
                similarity = fuzzyHashMatcher.getSimilarity(inputHash, line);
                if (fuzzyHashMatcher.matchExceedsThreshold(similarity, matchThreshold)) {
                    String match = fuzzyHashMatcher.getMatch(line);
                    // Because this would simply look odd, we ignore such entry and log
                    if (!StringUtils.isEmpty(match)) {
                        matched.put(match, similarity);
                    } else {
                        logger.error("Found a match against a malformed entry '{}'. Please inspect the contents of" + "the {} file and ensure they are properly formatted", new Object[] { line, HASH_LIST_FILE.getDisplayName() });
                    }
                }
            }
            // Check if single match is desired and if a match has been made
            if (context.getProperty(MATCHING_MODE).getValue() == singleMatch.getValue() && (matched.size() > 0)) {
                // and save time by breaking the outer loop
                break iterateFile;
            }
        }
        // no matter if the break was called or not, Continue processing
        // First by creating a new map to hold attributes
        Map<String, String> attributes = new ConcurrentHashMap<String, String>();
        // Then by iterating over the hashmap of matches
        if (matched.size() > 0) {
            int x = 0;
            for (Map.Entry<String, Double> entry : matched.entrySet()) {
                // defining attributes accordingly
                attributes.put(attributeName + "." + x + ".match", entry.getKey());
                attributes.put(attributeName + "." + x + ".similarity", String.valueOf(entry.getValue()));
                x++;
            }
            // Finally, append the attributes to the flowfile and sent to match
            flowFile = session.putAllAttributes(flowFile, attributes);
            session.transfer(flowFile, REL_FOUND);
            session.commit();
            return;
        } else {
            // Otherwise send it to non-match
            session.transfer(flowFile, REL_NOT_FOUND);
            session.commit();
            return;
        }
    } catch (IOException e) {
        logger.error("Error while reading the hash input source");
        context.yield();
    }
}

Example 2 with FuzzyHashMatcher

use of org.apache.nifi.processors.cybersecurity.matchers.FuzzyHashMatcher in project nifi by apache.

the class TestCompareFuzzyHash method testlooksLikeSpamSum.

@Test
public void testlooksLikeSpamSum() {
    FuzzyHashMatcher matcher = new SSDeepHashMatcher();
    List<String> invalidPayloads = Arrays.asList(// invalidFirstField
    "4AD:c1xs8Z/m6H0eRH31S8p8bHENANkPrNy4tkPytwPyh2jTytxPythPytNdPytDgYyF:OuO/mg3HFSRHEb44RNMi6uHU2hcq3", // emptyFirstField
    ":c1xs8Z/m6H0eRH31S8p8bHENANkPrNy4tkPytwPyh2jTytxPythPytNdPytDgYyF:OuO/mg3HFSRHEb44RNMi6uHU2hcq3", // emptySecondField
    "48::OuO/mg3HFSRHEb44RNMi6uHU2hcq3", // emptyThirdField
    "48:c1xs8Z/m6H0eRH31S8p8bHENANkPrNy4tkPytwPyh2jTytxPythPytNdPytDgYyF:", // withoutThirdField
    "48:c1xs8Z/m6H0eRH31S8p8bHENANkPrNy4tkPytwPyh2jTytxPythPytNdPytDgYyF", // Just a simple string
    "c1xs8Z/m6H0eRH31S8p8bHENANkPrNy4tkPytwPyh2jTytxPythPytNdPytDgYyF");
    for (String item : invalidPayloads) {
        Assert.assertTrue("item '" + item + "' should have failed validation", !matcher.isValidHash(item));
    }
    // Now test with a valid string
    Assert.assertTrue(matcher.isValidHash(ssdeepInput));
}