Changes for next data load

• Add author_index field so author and author2 don't need to be searched simultaneously
• Add tokenized title field
  

• Duplicate subject headings removed
• Do not load electronic journals with no 856 (identified by "Electronic journals" in 655, Ebsco in 710, [electronic resource] in GMD)
• Do not load records with invalid OCLC in 001
• Do not load items with blank cat date
• Fix spacing problems between 245|a and |b
• Fix invalid unicode problems
  

• Boost documents with many holdings
  
• Boost title keywords
• Boost 651|a