Indexing for regular expressions in text-centric applications

A computer system comprising:

• a processor unit in communication with memory;
• an index manager in communication with the memory, the index manager to build an index to support efficient evaluation of complex regular expressions, the index manager comprising:
• an index feature manager to specify index feature classes to extract index terms from a collection of documents and to identify metadata associated with the index terms to be stored in an index of the collection of documents, wherein each index term is a sequence of characters and character classes and each character class is a variable representing characters from a pre-determined set of characters;
• a feature class manager in communication with the index feature manager to apply the feature classes over the collection of documents to generate the index storing the metadata;
• a run time manager in communication with the index manager, the run time manager to extract constraints from a query regular expression and to transform said constraints into a new query over the index wherein each constraint is associated with one of the feature classes;
• responsive to receipt of a new query regular expression, an execution manager to execute the new query over the index to identify a subset of documents in the collection of documents over which the regular expression query can be evaluated;
• and
• the identified subset of documents returned for evaluation by the regular expression responsive to execution of the query over the index.

An article configured to build an index to support efficient evaluation of complex regular expressions, the article comprising:

• a storage device storing computer program instructions when executed to build the index, the instructions comprising:
• instructions to specify index feature classes to extract index terms from a collection of documents and to identify metadata associated with the index terms to be stored in an index of the collection of documents, wherein each index term is a sequence of characters and character classes and each character class is a variable representing characters from a pre-determined set of characters;
• instructions to apply the feature classes over the collection of documents to generate the index storing the metadata;
• instructions to receive a query regular expression, wherein the query regular expression is composed of multiple constraints and wherein each constraint is associated with one of the feature classes;
• instructions to extract the constraints from the query regular expression and to transform said constraints into a new query over the index;
• instructions to execute the new query over the index to identify a subset of documents in the collection of documents over which the regular expression query can be evaluated;
• and
• the identified subset of documents returned for evaluation by the regular expression responsive to execution of the query over the index.