Aho-Corasick

forked from BobuSumisu/aho-corasick

Fix the bug of missing pattern field in the return value (Match) of MatchFirst function

---

Implementation of the Aho-Corasick string-search algorithm in Go.

Licensed under MIT License.

Documentation

Can be found at pkg.go.dev.

Example Usage

Use a TrieBuilder to build a Trie:

trie := NewTrieBuilder().
    AddStrings([]string{"or", "amet"}).
    Build()

Then go and match something interesting:

matches := trie.MatchString("Lorem ipsum dolor sit amet, consectetur adipiscing elit.")
fmt.Printf("Got %d matches.\n", len(matches))

// => Got 3 matches.

What did we match?

for _, match := range matches {
    fmt.Printf("Matched pattern %d %q at position %d.\n", match.Match(),
        match.Pattern(), match.Pos())
}

// => Matched pattern 0 "or" at position 1.
// => Matched pattern 0 "or" at position 15.
// => Matched patterh 1 "amet" at position 22.

Building

You can easily load patterns from file:

builder := NewTrieBuilder()
builder.LoadPatterns("patterns.txt")
builder.LoadStrings("strings.txt")

Both functions expects a text file with one pattern per line. LoadPatterns expects the pattern to be in hexadecimal form.

Storing

Use Encode to store a Trie in gzip compressed binary format:

f, err := os.Create("trie.gz")
err := Encode(f, trie)

And Decode to load it from binary format:

f, err := os.Open("trie.gz")
trie, err := Decode(f)