Mirror/MarginaliaSearch
1
0
Fork 0
mirror of https://github.com/MarginaliaSearch/MarginaliaSearch.git synced 2025-02-23 13:09:00 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Added support for <base href>-style tags.

Browse Source
This commit is contained in:
vlofgren 2022-06-14 17:55:14 +02:00
parent 389818c6c3
commit 1de63f225d
6 changed files with 81 additions and 28 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

17
marginalia_nu/src/main/java/nu/marginalia/wmsa/edge/converting/processor/DocumentProcessor.java
View File

@ -185,26 +185,25 @@ public class DocumentProcessor {
}
private void getLinks(EdgeUrl baseUrl, ProcessedDocumentDetails ret, Document doc, EdgePageWordSet words) {
var links = doc.getElementsByTag("a");
var frames = doc.getElementsByTag("frame");
var feeds = doc.select("link[rel=alternate]");
LinkProcessor lp = new LinkProcessor(ret, baseUrl);
final LinkProcessor lp = new LinkProcessor(ret, baseUrl);
for (var atag : links) {
baseUrl = linkParser.getBaseLink(doc, baseUrl);
for (var atag : doc.getElementsByTag("a")) {
linkParser.parseLink(baseUrl, atag).ifPresent(lp::accept);
}
for (var frame : frames) {
for (var frame : doc.getElementsByTag("frame")) {
linkParser.parseFrame(baseUrl, frame).ifPresent(lp::accept);
}
for (var link : feeds) {
for (var link : doc.select("link[rel=alternate]")) {
feedExtractor
.getFeedFromAlternateTag(baseUrl, link)
.getFeedFromAlternateTag(baseUrl, link)
.ifPresent(lp::acceptFeed);
}
Set<String> linkTerms = new HashSet<>();
final Set<String> linkTerms = new HashSet<>();
for (var domain : lp.getForeignDomains()) {
linkTerms.add("links:"+domain.toString().toLowerCase());

34
marginalia_nu/src/main/java/nu/marginalia/wmsa/edge/converting/processor/logic/LinkParser.java
View File

@ -1,9 +1,12 @@
package nu.marginalia.wmsa.edge.converting.processor.logic;
import com.google.common.base.CharMatcher;
import com.google.common.base.Strings;
import lombok.SneakyThrows;
import nu.marginalia.wmsa.edge.model.EdgeUrl;
import org.jetbrains.annotations.Contract;
import org.jetbrains.annotations.Nullable;
import org.jsoup.nodes.Document;
import org.jsoup.nodes.Element;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
@ -26,11 +29,11 @@ public class LinkParser {
".gz", ".asc", ".md5", ".asf", ".mov", ".sig", ".pub", ".iso");
@Contract(pure=true)
public Optional<EdgeUrl> parseLink(EdgeUrl baseUrl, Element l) {
public Optional<EdgeUrl> parseLink(EdgeUrl relativeBaseUrl, Element l) {
return Optional.of(l)
.filter(this::shouldIndexLink)
.map(this::getUrl)
.map(link -> resolveUrl(baseUrl, link))
.map(link -> resolveUrl(relativeBaseUrl, link))
.flatMap(this::createURI)
.map(URI::normalize)
.map(this::renormalize)
@ -100,6 +103,8 @@ public class LinkParser {
}
private static final Pattern paramRegex = Pattern.compile("\\?.*$");
private static final Pattern spaceRegex = Pattern.compile(" ");
@SneakyThrows
private String resolveUrl(EdgeUrl baseUrl, String s) {
s = paramRegex.matcher(s).replaceAll("");
@ -111,10 +116,12 @@ public class LinkParser {
// url looks like /my-page
if (s.startsWith("/")) {
return baseUrl.sibling(s).toString();
return baseUrl.withPath(s).toString();
}
return baseUrl.sibling(relativeNavigation(baseUrl) + s.replaceAll(" ", "%20")).toString();
final String partFromNewLink = spaceRegex.matcher(s).replaceAll("%20");
return baseUrl.withPath(relativeNavigation(baseUrl) + partFromNewLink).toString();
}
// for a relative url that looks like /foo or /foo/bar; return / or /foo
@ -162,4 +169,23 @@ public class LinkParser {
}
return true;
}
@Nullable
public EdgeUrl getBaseLink(Document parsed, EdgeUrl documentUrl) {
var baseTags = parsed.getElementsByTag("base");
try {
for (var tag : baseTags) {
String href = tag.attr("href");
if (!Strings.isNullOrEmpty(href)) {
return new EdgeUrl(resolveUrl(documentUrl, href));
}
}
}
catch (Exception ex) {
logger.warn("Failed to parse <base href=...>, falling back to document url");
}
return documentUrl;
}
}

14
marginalia_nu/src/main/java/nu/marginalia/wmsa/edge/crawling/retreival/CrawlerRetreiver.java
View File

@ -202,10 +202,11 @@ public class CrawlerRetreiver {
return domain.equals(url.domain.toString().toLowerCase());
}
private void findLinks(EdgeUrl url, Document parsed) {
private void findLinks(EdgeUrl baseUrl, Document parsed) {
baseUrl = linkParser.getBaseLink(parsed, baseUrl);
for (var link : parsed.getElementsByTag("a")) {
linkParser.parseLink(url, link)
linkParser.parseLink(baseUrl, link)
.filter(this::isSameDomain)
.filter(u -> !urlBlocklist.isUrlBlocked(u))
.filter(u -> !urlBlocklist.isForumLink(u))
@ -213,7 +214,7 @@ public class CrawlerRetreiver {
.ifPresent(queue::addLast);
}
for (var link : parsed.getElementsByTag("frame")) {
linkParser.parseFrame(url, link)
linkParser.parseFrame(baseUrl, link)
.filter(this::isSameDomain)
.filter(u -> !urlBlocklist.isUrlBlocked(u))
.filter(u -> !urlBlocklist.isForumLink(u))
@ -221,7 +222,7 @@ public class CrawlerRetreiver {
.ifPresent(queue::addLast);
}
for (var link : parsed.getElementsByTag("iframe")) {
linkParser.parseFrame(url, link)
linkParser.parseFrame(baseUrl, link)
.filter(this::isSameDomain)
.filter(u -> !urlBlocklist.isUrlBlocked(u))
.filter(u -> !urlBlocklist.isForumLink(u))
@ -230,10 +231,11 @@ public class CrawlerRetreiver {
}
}
private Optional<EdgeUrl> findCanonicalUrl(EdgeUrl url, Document parsed) {
private Optional<EdgeUrl> findCanonicalUrl(EdgeUrl baseUrl, Document parsed) {
baseUrl = baseUrl.withPath("/");
for (var link : parsed.select("link[rel=canonical]")) {
return linkParser.parseLink(url, link);
return linkParser.parseLink(baseUrl, link);
}
return Optional.empty();

1
marginalia_nu/src/main/java/nu/marginalia/wmsa/edge/model/EdgeDomain.java
View File

@ -21,6 +21,7 @@ public class EdgeDomain implements WideHashable {
@SneakyThrows
public EdgeDomain(String host) {
Objects.requireNonNull(host, "domain name must not be null");
var dot = host.lastIndexOf('.');

9
marginalia_nu/src/main/java/nu/marginalia/wmsa/edge/model/EdgeUrl.java
View File

@ -79,11 +79,6 @@ public class EdgeUrl implements WideHashable {
this.port = port(URI.getPort(), proto);
}
public EdgeUrl sibling(String newPath) {
return new EdgeUrl(proto, domain, port, newPath);
}
private static Integer port(Integer port, String protocol) {
if (null == port || port < 1) {
return null;
@ -120,5 +115,7 @@ public class EdgeUrl implements WideHashable {
return (int) path.chars().filter(c -> c=='/').count();
}
public EdgeUrl withPath(String s) {
return new EdgeUrl(proto, domain, port, s);
}
}

34
marginalia_nu/src/test/java/nu/marginalia/wmsa/edge/crawling/LinkParserTest.java
View File

@ -11,9 +11,8 @@ import static org.junit.jupiter.api.Assertions.*;
class LinkParserTest {
private String parseLink(String href, String base) throws URISyntaxException {
var url = new EdgeUrl("http://www.marginalia.nu/" + base);
var domain = url.domain;
private String parseLink(String href, String relBase) throws URISyntaxException {
var url = new EdgeUrl("http://www.marginalia.nu/" + relBase);
var parser = new LinkParser();
var stuff = Jsoup.parseBodyFragment("<a href='"+href+"''>test</a>");
var lnk = parser.parseLink(
@ -43,6 +42,7 @@ class LinkParserTest {
void testAnchor() throws URISyntaxException {
assertNull(parseLink("#test", "/"));
}
@Test
void testRelative() throws URISyntaxException {
assertEquals("http://www.marginalia.nu/test", parseLink("../test", "/"));
@ -51,4 +51,32 @@ class LinkParserTest {
assertEquals("http://www.marginalia.nu/test", parseLink("../test", "/foo/index.html"));
assertEquals("http://www.marginalia.nu/test", parseLink("/test", "/foo/index.html"));
}
private EdgeUrl getBaseUrl(String href, EdgeUrl documentUrl) {
LinkParser lp = new LinkParser();
return lp.getBaseLink(Jsoup.parse("<base href=\"" + href + "\" />"), documentUrl);
}
@Test
public void getBaseUrlTest() throws URISyntaxException {
assertEquals(new EdgeUrl("https://www.marginalia.nu/base"),
getBaseUrl("/base", new EdgeUrl("https://www.marginalia.nu/test/foo.bar")));
assertEquals(new EdgeUrl("https://memex.marginalia.nu/base"),
getBaseUrl("https://memex.marginalia.nu/base", new EdgeUrl("https://www.marginalia.nu/test/foo.bar")));
assertEquals(new EdgeUrl("https://www.marginalia.nu/test/base"),
getBaseUrl("base", new EdgeUrl("https://www.marginalia.nu/test/foo.bar")));
}
@Test
public void testParseBadBaseLink() throws URISyntaxException {
LinkParser lp = new LinkParser();
var url = new EdgeUrl("https://memex.marginalia.nu/");
assertEquals(url, lp.getBaseLink(Jsoup.parse("<base href/>"), url));
assertEquals(url, lp.getBaseLink(Jsoup.parse("<base target=\"foo\"/>"), url));
assertEquals(url, lp.getBaseLink(Jsoup.parse("<base href=\"http://\"/>"), url));
}
}