AfflictionNetworks/depot
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Oops. We were not taking the size of the matched-against column into account at all. That's not intuitive. The more precise the match, the higher we want to rank it.

Browse Source
This commit is contained in:
Par Winzell
2009-03-29 19:55:24 +00:00
parent 34599bc760
commit b2bd4b51f7
1 changed files with 4 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/java/com/samskivert/depot/impl/PostgreSQLBuilder.java
+4 -1
View File
@@ -67,7 +67,10 @@ public class PostgreSQLBuilder
appendIdentifier("ftsCol_" + rank.getDefinition().getName()); appendIdentifier("ftsCol_" + rank.getDefinition().getName());
_builder.append(", to_tsquery('"). _builder.append(", to_tsquery('").
append(translateFTConfig(getFTIndex(rank.getDefinition()).configuration())). append(translateFTConfig(getFTIndex(rank.getDefinition()).configuration())).
append("', ?), 32)"); // TODO: The normalization parameter is really quite important, and should
// TODO: perhaps be configurable, but for the moment we hard-code it to 1:
// TODO: "divides the rank by the 1 + logarithm of the document length"
append("', ?), 1)");
} }
@Override public void visit (EpochSeconds epochSeconds) { @Override public void visit (EpochSeconds epochSeconds) {