Features

• Configurable and extensible (though underlying API currently quite rough around the edges); pluggable crawler steps, tokenizer and even indexer.
• Works for multi-module projects (those that correctly use the <distributionManagement> pom section)
• Generates an autocomplete list
• Generates a url-title-map (allowing presentation of the title in search results)
• Includes a copy UI search script (should work with maven skins; reflow, MSB3 and io.earcam.maven.site.skin - requires bootstrap and jquery)
• Process HTML and PDFs
• By default Javadoc and Xref generated content are minimally indexed to preserve index quality and minimize JSON size

Limitations

1. Does not support “snippet” highlighting - have experimented here, but performance issues ensue - future work to do this dynamically and lazily planned
2. Does not handle building multiple search indices from projects containing multiple independent sites - the index will be generated in the root project declaring a reporting directory.
3. Does not (yet) handle file formats other than HTML and PDF, although extensible with custom code