lee/heritrix3
1
0
Fork 0
mirror of https://github.com/internetarchive/heritrix3.git synced 2026-09-24 14:45:56 +00:00
Code Issues Packages Projects Releases Wiki Activity

Attempt to filter out embedded images.

Browse Source 
(cherry picked from commit aa5ff1dbdaefd04652a9c66506d20f1a6ae01dc3)
This commit is contained in:
Colin Rosenthal
2019-11-19 15:04:50 +01:00
parent 7a83aecb28
commit 204491c357
1 changed files with 13 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files
modules/src/main/java/org/archive/modules/extractor/ExtractorHTML.java
+13 -11
View File
@@ -445,18 +445,20 @@ public class ExtractorHTML extends ContentExtractor implements InitializingBean
} else if (attr.start(5) > -1) {
// SRC etc.
CharSequence context = elementContext(element, attr.group(5));
// true, if we expect another HTML page instead of an image etc.
final Hop hop;
if(!framesAsEmbeds
&& (elementStr.equalsIgnoreCase(FRAME) || elementStr
.equalsIgnoreCase(IFRAME))) {
hop = Hop.NAVLINK;
} else {
hop = Hop.EMBED;
if (!context.toString().toLowerCase().startsWith("data:")) {
// true, if we expect another HTML page instead of an image etc.
final Hop hop;
if (!framesAsEmbeds
&& (elementStr.equalsIgnoreCase(FRAME) || elementStr
.equalsIgnoreCase(IFRAME))) {
hop = Hop.NAVLINK;
} else {
hop = Hop.EMBED;
}
processEmbed(curi, value, context, hop);
}
processEmbed(curi, value, context, hop);
} else if (attr.start(6) > -1) {
// CODEBASE
codebase = (value instanceof String)?