kmp-search.core — kmp-search 0.1.2

functions to search a byte stream for a byte pattern

• create a context with the byte pattern

• call the search function repeatedly to process each sequential buffer of stream contents

• each call to search returns a new context when it:

  • finds a match within a buffer, or
  • exhausts a buffer without finding a match

• retrieve the search result from the new context using the match function. The result is either:

  • if a match was found, a Long containing the offset of the match within the stream of bytes processed, or

  • if no match was not found, nil

• each returned context also contains enough state to allow matching across buffer boundaries

• to find all matches, call search on each buffer repeatedly until match returns nil, then continue searching the next buffer

  • the context for each call to search will be the context returned by the previous call

• search-file is useful both in its own right and as an example

reference: http://www.inf.fh-flensburg.de/lang/algorithmen/pattern/kmpen.htm

context clj

(context pattern)

Search clj protocol

match clj

(match this)

search clj

(search this buffer)

(search this buffer limit)

search-file clj

(search-file pattern file & {:keys [buffer-size] :or {buffer-size 1024}})

returns a vector containing all the offsets where a byte pattern appears within a file