Multi-user search system with methodology for instant indexing

摘要

A multi-user search system with methodology for instant indexing. In one embodiment, for example, a system for instant indexing includes a token store storing sets of tokens for current versions of documents. The system further includes a tokenizer server configured to tokenize new versions of the documents and to generate sets of tokens for the new versions of the documents, an instant indexer configured to determine tokens to use to index the documents based on identified differences between the sets of tokens for the new versions of the documents and the sets of tokens for the current versions of the documents, and to generate index mutations including the tokens to use to index the documents, an index mutation journal configured to store the generated index mutations in association with timestamps, and an index mutation server configured to provide, to index servers, from the index mutation journal, generated index mutations for the index servers that are associated with timestamps that are newer than specified timestamps.

主权项

1. A system comprising:
one or more hardware processors; a token store storing a set of tokens for a current version of a document; a tokenizer server configured to tokenize, via at least one of the hardware processors, a new version of the document and to generate, via at least one of the hardware processors, a set of tokens for the new version of the document; an instant indexer server configured to determine, via at least one of the hardware processors, one or more tokens to use to index the document based on identified differences between the set of tokens for the new version of the document and the set of tokens for the current version of the document, and to generate, via at least one of the hardware processors, an index mutation comprising the one or more tokens to use to index the document; an index mutation journal configured to store the generated index mutation in association with a timestamp; an index mutation server configured to provide, via at least one of the hardware processors, from the generated index mutation, to an index server, the one or more tokens to use to index the document, if the timestamp associated with the generated index mutation in the index mutation journal is newer than a timestamp specified by the index server; the index server configured to receive, via at least one of the hardware processors, from the index mutation server, the one or more tokens to use to index the document, and configured to update an index of a plurality of documents at the index server to index the document by the one or more tokens to use to index the document; and wherein, at a time prior to the update, the index at the index server indexes the plurality of documents but does not index the document by the one or more tokens to use to index the document.