[MAHOUT-580] Missing basePath in Bayes job / InMemoryDataStore

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.4
• Fix Version/s: 0.5
• Assignee: Robin Anil
• Reporter: Sean R. Owen
• Due: 21/Jan/11
• Created: 10/Jan/11 19:48
• Updated: 31/Jan/24 22:12
• Resolved: 11/Jan/11 20:53

Description

Reported on user@ by Pierre Mage:

I found the following issue in the trunk: If the dataSource option is "hdfs"
in org.apache.mahout.classifier.
Classify, there is a problem with InMemoryBayesDatastore because its
basePath param is not defined :
To fix it, I edited Classify like this:
122. if ("hdfs".equals(dataSource)) {
123. params.set("basePath", modelBasePath);

Yes, looks like InMemoryDataStore expects this to be set – so does HBaseDataStore. But that takes the base path (table) separately as a param. And InMemoryDataStore doesn't take it at all. I adjusted this since it seems correct. And then tried to centralized handling of Bayes params in BayesParameters since it seems to get half processed in this class only.

I have attached my own rather more elaborate patch that reflects this.