Automatic analysis of repository structure to facilitate natural language queries

#Repositories hold data such as emails, text documents, executable files, and the like. This data may be retrieved from the repository in response to a query. The process of retrieving this data in a fast, precise, and accurate way can become more difficult based, in part, upon how much data may be stored in the repository.

#According to one embodiment of the present invention, a method for analyzing a repository includes obtaining a list of known persons in a repository based on objects, users, and groups retrieved from the repository. The method may further select one of the objects having a field and a value, and determine whether the field of the selected is searchable, contains special values, or is a facet, where the field is defined as a facet based on a probability that the field of the selected object has a limited number of possible values. The method may also count a number of documents to contain a property for each property retrieved from a repository and generate a repository information archive. The repository information archive may include a persons matrix, the relationship between the selected object and at least one other object, statistics and counts related to properties in the selected objects, and whether or not the field of the selected object is a facet and a distribution of values in a facet.

#FIG. 1 shows a block diagram of an example computing system to analyze a repository;

#FIG. 2 shows a block diagram of an example repository to be analyzed;

#FIG. 3 shows a block diagram of an example object of a repository;

#The subject matter disclosed herein relates to techniques for comprehensive analysis of a repository structure. Retrieving data from a large document repository based on queries can be a serious challenge.

#For example, a growing number of queries on databases and repositories are Natural Language (NL) queries, and may be made with human speech and phrasing. A request originating from a speech based NL query may have a large number of possible meanings for each of its words. This results in a huge number of combinations that could map to a large number of queries. Retrieving results based on identifying a collection of the words, as interpreted from the NL query, may be prone to errors and also may be very imprecise. A paradigm based on a collection of words for query access to repositories may lead to large recall and low precision.

#In the present disclosure, a repository may be analyzed to improve precision of query results. In one example, static analysis may be performed on the repository in order to generate an information structure to be used in interpreting queries and query retrieval processes. Static analysis may be static in its performance prior to the execution of a set of queries on the repository.