Class MinHashLSHInsertionSession

join/LSH/datasketch/lsh.py:349–379

Context manager for batch insertion of documents into a MinHashLSH.

class MinHashLSHInsertionSession:
    '''Context manager for batch insertion of documents into a MinHashLSH.
    '''

    def __init__(self, lsh, buffer_size):
        self.lsh = lsh
        self.lsh.buffer_size = buffer_size

    def __enter__(self):
        return self

    def __exit__(self, exc_type, exc_val, exc_tb):
        self.close()

    def close(self):
        self.lsh.keys.empty_buffer()
        for hashtable in self.lsh.hashtables:
            hashtable.empty_buffer()

    def insert(self, key, minhash, check_duplication=True):
        '''
        Insert a unique key to the index, together
        with a MinHash (or weighted MinHash) of the set referenced by
        the key.

        Args:
            key (hashable): The unique identifier of the set.
            minhash (datasketch.MinHash): The MinHash of the set.
        '''
        self.lsh._insert(key, minhash, check_duplication=check_duplication,
                         buffer=True)