Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Original file line number Diff line number Diff line change
Expand Up @@ -162,11 +162,43 @@ public void execute(Tuple tuple) {
// check that the mimetype is in the whitelist
if (!mimeTypeWhiteList.isEmpty()) {
boolean mt_match = false;
// see if a mimetype was guessed in JSOUPBolt
// parse.Content-Type is assumed byte-detected (JSoupParserBolt uses Tika detection,
// not the raw server header). A custom upstream writing a header-copied value bypasses
// this check — that is a caller responsibility.
String mimeType = metadata.getFirstValue("parse.Content-Type");
Comment thread
dpol1 marked this conversation as resolved.
// otherwise rely on what could have been obtained from HTTP
if (mimeType == null) {
mimeType = metadata.getFirstValue(HttpHeaders.CONTENT_TYPE, this.protocolMDprefix);
// parse.Content-Type is absent: detect from content bytes so that
// the whitelist is evaluated against the same type Tika will use
// to select a parser, not the server-declared HTTP header which is
// untrusted and may differ from what the bytes actually are.
String httpCTHint =
metadata.getFirstValue(HttpHeaders.CONTENT_TYPE, this.protocolMDprefix);
org.apache.tika.metadata.Metadata detectionMd =
new org.apache.tika.metadata.Metadata();
if (StringUtils.isNotBlank(httpCTHint)) {
// pass the header as a hint only — detect() weighs it but
// content bytes take precedence

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not quite: for bytes with no magic match Tika returns application/octet-stream and applyHint swaps in any hint that specialises it, which every type does. So on unrecognised content the server header wins, and the "undetectable content is now rejected" note in the description only holds when no Content-Type header was sent. The result still matches what AutoDetectParser dispatches on, so the change is fine, but say "the hint narrows the magic result" rather than "bytes take precedence".

detectionMd.set(org.apache.tika.metadata.Metadata.CONTENT_TYPE, httpCTHint);
}
// pass the filename so detection matches what the parser dispatches on;
// without it, an ambiguous byte sequence (e.g. plain text with a .html
// extension) can resolve differently here than at parse time
try {
URL _url = URLUtil.toURL(url);
detectionMd.set(TikaCoreProperties.RESOURCE_NAME_KEY, _url.getFile());
} catch (MalformedURLException e1) {
throw new IllegalStateException("Malformed URL", e1);
}
try {
mimeType = tika.detect(new ByteArrayInputStream(content), detectionMd);
Comment thread
dpol1 marked this conversation as resolved.
} catch (IOException e) {
LOG.warn("Failed to detect MIME type for {}: {}", url, e.getMessage());
}
if (mimeType != null) {
// write back so downstream code and metadata consumers see
// the same value (avoids a second detection pass)

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The parse path never reads parse.Content-Type, AutoDetectParser detects again, and on success the parse. copy loop overwrites this. The write-back only survives on the rejected tuple, which is useful when debugging a rejection, so keep it but fix the comment.

metadata.setValue("parse.Content-Type", mimeType);
}
}
if (mimeType != null) {
for (Pattern mt : mimeTypeWhiteList) {
Expand Down
Original file line number Diff line number Diff line change
@@ -0,0 +1,145 @@
/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to you under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/

package org.apache.stormcrawler.tika;

import java.io.IOException;
import java.nio.charset.StandardCharsets;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import org.apache.http.HttpHeaders;
import org.apache.storm.task.OutputCollector;
import org.apache.stormcrawler.Constants;
import org.apache.stormcrawler.Metadata;
import org.apache.stormcrawler.TestUtil;
import org.apache.stormcrawler.parse.ParsingTester;
import org.apache.stormcrawler.persistence.Status;
import org.apache.stormcrawler.protocol.ProtocolResponse;
import org.junit.jupiter.api.Assertions;
import org.junit.jupiter.api.BeforeEach;
import org.junit.jupiter.api.Test;

/**
* Regression test for: when no parse.Content-Type is present, ParserBolt must evaluate
* parser.mimetype.whitelist against the byte-detected MIME type (via tika.detect()), not the
* server-declared HTTP Content-Type header. Previously the whitelist checked the header while
* Tika's AutoDetectParser dispatched on the bytes, allowing a server to claim a whitelisted type
* while serving arbitrary content.
*/
class ParserBoltWhitelistDetectionTest extends ParsingTester {

@BeforeEach
void setupParserBolt() {
bolt = new ParserBolt();
setupParserBolt(bolt);
}

/**
* The whitelist allows Word documents (application/.+word.*). The server header claims Word,
* but the body bytes are plain HTML. After the fix, detection on bytes yields text/html which
* does NOT match the whitelist, so the document must be rejected with ERROR.
*/
@Test
void whitelistAppliesToTheDetectedType() throws IOException {
Map<String, Object> conf = new HashMap<>();
// the whitelist shipped by the archetypes
conf.put("parser.mimetype.whitelist", "application/.+word.*");
conf.put(ProtocolResponse.PROTOCOL_MD_PREFIX_PARAM, "http.");
bolt.prepare(conf, TestUtil.getMockedTopologyContext(), new OutputCollector(output));

// no parse.Content-Type: no JSoupParserBolt upstream, or detect.mimetype disabled
Metadata metadata = new Metadata();
metadata.addValue(
"http." + HttpHeaders.CONTENT_TYPE,
"application/vnd.openxmlformats-officedocument.wordprocessingml.document");

// the body is NOT a word document
byte[] content =
"<html><body><p>not a word document</p></body></html>"
.getBytes(StandardCharsets.UTF_8);
parse("https://example.org/doc.docx", content, metadata);

System.out.println("detected type: " + metadata.getFirstValue("parse.Content-Type"));

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Leftover println, and this is where the assert belongs: check parse.Content-Type starts with text/html here and in filenameHintInfluencesDetection, otherwise the tests pass for any reason the whitelist rejects or accepts. whitelistUsesPreexistingParsedContentType has the same gap: HTML bytes pass whether or not the bolt re-detects; plain text at a .txt URL with parse.Content-Type preset to text/html would prove the preset value is trusted.

System.out.println("emitted documents: " + output.getEmitted().size());

List<List<Object>> status = output.getEmitted(Constants.StatusStreamName);
Assertions.assertEquals(
1, status.size(), "content not matching the whitelist should be rejected");
Assertions.assertEquals(
Status.ERROR,
status.get(0).get(2),
"status should be ERROR for mismatched content type");
}

/**
* Sanity check: when parse.Content-Type IS already present (e.g. set by JSoupParserBolt), the
* whitelist check must still use it directly and not re-detect.
*/
@Test
void whitelistUsesPreexistingParsedContentType() throws IOException {
Map<String, Object> conf = new HashMap<>();
conf.put("parser.mimetype.whitelist", "text/html.*");
conf.put(ProtocolResponse.PROTOCOL_MD_PREFIX_PARAM, "http.");
bolt.prepare(conf, TestUtil.getMockedTopologyContext(), new OutputCollector(output));

Metadata metadata = new Metadata();
// simulate JSoupParserBolt having detected the type already
metadata.addValue("parse.Content-Type", "text/html; charset=UTF-8");
metadata.addValue("http." + HttpHeaders.CONTENT_TYPE, "text/html; charset=UTF-8");

byte[] content = "<html><body><p>hello</p></body></html>".getBytes(StandardCharsets.UTF_8);
parse("https://example.org/index.html", content, metadata);

// document should pass the whitelist and be emitted (no ERROR on status stream)
List<List<Object>> status = output.getEmitted(Constants.StatusStreamName);
boolean hasError =
status != null && status.stream().anyMatch(row -> Status.ERROR.equals(row.get(2)));
Assertions.assertFalse(hasError, "whitelisted HTML document should not be rejected");
}

/**
* Plain-text bytes with a .html URL extension. Without the filename hint, Tika resolves the
* ambiguous bytes as text/plain; with it, the extension pushes detection to text/html. The
* whitelist is set to text/html.*, so the document must be accepted — verifying that the same
* RESOURCE_NAME_KEY hint is passed to both the whitelist check and the parser dispatch.
*/
@Test
void filenameHintInfluencesDetection() throws IOException {
Map<String, Object> conf = new HashMap<>();
conf.put("parser.mimetype.whitelist", "text/html.*");
conf.put(ProtocolResponse.PROTOCOL_MD_PREFIX_PARAM, "http.");
bolt.prepare(conf, TestUtil.getMockedTopologyContext(), new OutputCollector(output));

// no parse.Content-Type, no Content-Type header — detection relies on bytes + filename
Metadata metadata = new Metadata();

// plain text bytes: no HTML magic, ambiguous without the filename hint
byte[] content = "just some plain text, no html tags".getBytes(StandardCharsets.UTF_8);

// .html extension should push detection to text/html
parse("https://example.org/page.html", content, metadata);

System.out.println("detected type: " + metadata.getFirstValue("parse.Content-Type"));

List<List<Object>> status = output.getEmitted(Constants.StatusStreamName);
boolean hasError =
status != null && status.stream().anyMatch(row -> Status.ERROR.equals(row.get(2)));
Assertions.assertFalse(
hasError, "document with .html URL should be accepted by text/html.* whitelist");
}
}