flokati/modules/markov.go

// vi:ts=4:sts=4:sw=4:noet:tw=72

package modules

// This Markov chain code is taken from the "Generating arbitrary text"
// codewalk: http://golang.org/doc/codewalk/markov/
//
// Minor modifications have been made to make it easier to integrate
// with a webserver and to save/load state

import (
	"bufio"
	"encoding/gob"
	"flag"
	"fmt"
	"math/rand"
	"os"
	"regexp"
	"strings"
	"sync"
	"time"

	"git.dnix.de/an/xlog"
)

var (
	markovPrefixLen      = flag.Int("markov_prefix_len", 3, "markov: prefix len")
	markovAnswerLen      = flag.Int("markov_answer_len", 10, "markov: answer len")
	markovResponseChance = flag.Int("markov_response_chance", 10, "markov: chance to get an answer (percent)")
	markovStateFile      = flag.String("markov_state_file", "state.dat", "markov: state file")
	markovTrainFile      = flag.String("markov_train_file", "train.txt", "markov: training file")
)

var markovChain *MarkovChain

func init() {
	MsgFuncs["markov"] = markovHandleMessage
	RunFuncs["markov"] = markovRun
}

func markovHandleMessage(m *Message) {

	tok := strings.Split(m.Text, " ")
	if tok[0] == "!talk" {
		if len(tok) < 2 {
			SayCh <- m.Channel + "\npfffrrz ... worüber denn?"
		} else {
			talk := tok[1]
			for {
				time.Sleep(time.Duration(rand.Intn(8)+2) * time.Second)
				talk = markovChain.Generate(*markovAnswerLen, talk)
				SayCh <- m.Channel + "\n" + talk
				if rand.Intn(100) < 10 {
					break
				}
			}
		}
		return
	}

	text := m.Text
	if text == "" {
		return
	}
	text = markovParseText(text)

	if rand.Intn(100) <= *markovResponseChance || strings.Index(text, BotNick) != -1 {
		responseText := markovChain.Generate(*markovAnswerLen, text)
		if responseText != "" {
			go func() {
				time.Sleep(time.Duration(rand.Intn(8)+2) * time.Second)
				SayCh <- m.Channel + "\n" + responseText
			}()
		}
	}

	if m.From != BotName {
		markovChain.Write(text)
	}
}

func markovRun() {
	markovChain = markovNewChain(*markovPrefixLen)
	err := markovChain.Load(*markovStateFile)
	if err != nil {
		xlog.Error(err.Error())
	}
	filepath := *markovTrainFile
	if filepath != "-" {
		file, _ := os.Open(filepath)
		scanner := bufio.NewScanner(file)
		for scanner.Scan() {
			text := scanner.Text()
			text = markovParseText(text)
			if text != "" {
				markovChain.Write(text)
			}
		}
	}
	go func() {
		for {
			time.Sleep(60 * time.Second)
			markovChain.Save(*markovStateFile)
		}
	}()
}

func markovParseText(text string) string {
	messageRegex := regexp.MustCompile(`<([^>]+)>`)
	matches := messageRegex.FindAllStringSubmatch(text, -1)
	for _, matches2 := range matches {
		if strings.HasPrefix(matches2[1], "http") || strings.HasPrefix(matches2[1], "mailto") {
			text = strings.Replace(text, matches2[0], "", -1)
		} else if strings.HasPrefix(matches2[1], "@U") {
			parts := strings.SplitN(matches2[1], "|", 2)
			if len(parts) == 2 {
				text = strings.Replace(text, matches2[0], "@"+parts[1], -1)
			} else {
				text = strings.Replace(text, matches2[0], "", -1)
			}
		} else if strings.HasPrefix(matches2[1], "@") {
			text = strings.Replace(text, matches2[0], matches2[1], -1)
		} else if strings.HasPrefix(matches2[1], "#") {
			parts := strings.SplitN(matches2[1], "|", 2)
			if len(parts) == 2 {
				text = strings.Replace(text, matches2[0], "#"+parts[1], -1)
			} else {
				text = strings.Replace(text, matches2[0], "", -1)
			}
		}
	}
	text = strings.TrimSpace(text)
	text = strings.Replace(text, "&lt;", "<", -1)
	text = strings.Replace(text, "&gt;", ">", -1)
	text = strings.Replace(text, "&amp;", "&", -1)
	//text = strings.Replace(text, ",", " ", -1)
	//reg := regexp.MustCompile("[^a-zA-Z0-9 ]+")
	//delText := reg.ReplaceAllString(text, "")
	return strings.ToLower(text)
}

// Prefix is a Markov chain prefix of one or more words.
type MarkovPrefix []string

// String returns the Prefix as a string (for use as a map key).
func (p MarkovPrefix) String() string {
	return strings.Trim(strings.Join(p, " "), " ")
}

// Shift removes the first word from the Prefix and appends the given word.
func (p MarkovPrefix) Shift(word string) {
	copy(p, p[1:])
	p[len(p)-1] = word
}

// MarkovChain contains a map ("chain") of prefixes to a list of suffixes.
// A prefix is a string of prefixLen words joined with spaces.
// A suffix is a single word. A prefix can have multiple suffixes.
type MarkovChain struct {
	MarkovChain map[string][]string
	prefixLen   int
	mu          sync.Mutex
}

// NewMarkovChain returns a new MarkovChain with prefixes of prefixLen words.
func markovNewChain(prefixLen int) *MarkovChain {
	return &MarkovChain{
		MarkovChain: make(map[string][]string),
		prefixLen:   prefixLen,
	}
}

func markovContains(a []string, s string) bool {
	for _, v := range a {
		if v == s {
			return true
		}
	}
	return false
}

// Write parses the bytes into prefixes and suffixes that are stored in MarkovChain.
func (c *MarkovChain) Write(in string) (int, error) {
	in = strings.ToLower(in)
	if strings.HasPrefix(in, BotNick) {
		tok := strings.Split(in, " ")
		in = strings.Replace(in, tok[0]+" ", "", 1)
	}
	sr := strings.NewReader(in)
	p := make(MarkovPrefix, c.prefixLen)
	c.mu.Lock()
	defer c.mu.Unlock()
	for {
		var s string
		if _, err := fmt.Fscan(sr, &s); err != nil {
			break
		}
		key := p.String()
		if !markovContains(c.MarkovChain[key], s) {
			c.MarkovChain[key] = append(c.MarkovChain[key], s)
			xlog.Debug("Chain len: %d, learned [%s] [%s]", len(c.MarkovChain), key, s)
		} else {
			xlog.Debug("Chain len: %d, [%s] [%s] is already known", len(c.MarkovChain), key, s)
		}
		p.Shift(s)
	}
	return len(in), nil
}

// Generate returns a string of at most n words generated from MarkovChain.
func (c *MarkovChain) Generate(n int, in string) string {
	in = strings.ToLower(in)
	if strings.HasPrefix(in, BotNick) {
		tok := strings.Split(in, " ")
		in = strings.Replace(in, tok[0]+" ", "", 1)
	}
	c.mu.Lock()
	defer c.mu.Unlock()
	var p MarkovPrefix
	var words []string
	var start string
	for attempt := 0; attempt < 10; attempt++ {
		p = make(MarkovPrefix, c.prefixLen)
		inWords := strings.Split(in, " ")
		if attempt < 9 {
			start = inWords[rand.Intn(len(inWords))]
		} else {
			start = ""
		}
		p.Shift(start)
		xlog.Debug("Looking for answer on [%s]", start)
		for i := 0; i < n; i++ {
			choices := c.MarkovChain[p.String()]
			if len(choices) == 0 {
				break
			}
			next := choices[rand.Intn(len(choices))]
			words = append(words, next)
			if strings.HasSuffix(next, ".") || strings.HasSuffix(next, "!") || strings.HasSuffix(next, "?") {
				break
			}
			p.Shift(next)
		}
		if len(words) > 0 {
			break
		}
	}
	start = strings.Trim(start, " ")
	return start + " " + strings.Join(words, " ")
}

// Save the chain to a file
func (c *MarkovChain) Save(fileName string) error {
	// Open the file for writing
	fo, err := os.Create(fileName)
	if err != nil {
		return err
	}
	// close fo on exit and check for its returned error
	defer func() {
		if err := fo.Close(); err != nil {
			panic(err)
		}
	}()

	// Create an encoder and dump to it
	c.mu.Lock()
	defer c.mu.Unlock()

	enc := gob.NewEncoder(fo)
	err = enc.Encode(c)
	if err != nil {
		return err
	}

	return nil
}

// Load the chain from a file
func (c *MarkovChain) Load(fileName string) error {
	// Open the file for reading
	fi, err := os.Open(fileName)
	if err != nil {
		return err
	}
	// close fi on exit and check for its returned error
	defer func() {
		if err := fi.Close(); err != nil {
			panic(err)
		}
	}()

	// Create a decoder and read from it
	c.mu.Lock()
	defer c.mu.Unlock()

	dec := gob.NewDecoder(fi)
	err = dec.Decode(c)
	if err != nil {
		return err
	}

	return nil
}