abhin4v/marvin
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Added Url skipping features

Browse Source
This commit is contained in:
Abhinav Sarkar 2011-04-26 18:52:01 +05:30
parent 5b901e51a9
commit 68602f3942
1 changed files with 23 additions and 17 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

40
src/marvin/core.clj
View File

@ -11,6 +11,10 @@
(defn parse-int [s] (Integer/parseInt s)) (defn parse-int [s] (Integer/parseInt s))
(def url-pattern
#"(?i)\b((?:https?://|www\d{0,3}[.]|[a-z0-9.\-]+[.][a-z]{2,4}/)(?:[^\s()<>]+|\(([^\s()<>]+|(\([^\s()<>]+\)))*\))+(?:\(([^\s()<>]+|(\([^\s()<>]+\)))*\)|[^\s`!()\[\]{};:'\".,<>?«»“”‘’]))")
(defn tokenize-line [line] (defn tokenize-line [line]
(let [non-char-pattern (re-pattern "[\\p{Z}\\p{C}\\p{P}]+") (let [non-char-pattern (re-pattern "[\\p{Z}\\p{C}\\p{P}]+")
tokenize tokenize
@ -211,7 +215,7 @@
min-sentence-length min-sentence-length
max-sentence-length)] max-sentence-length)]
(println ">> Sending message:" sentence) (println ">> Sending message:" sentence)
(send-message bot channel sentence)))) ] (send-message bot channel sentence))))]
(try (try
(swap! msg-count inc) (swap! msg-count inc)
(cond (cond
@ -241,21 +245,24 @@
:else :else
(do (do
(doseq [line (sentencize-text message)] (doseq [line (sentencize-text message)]
(when-not (= 1 (count (tokenize-line line))) (let [urls (map first (re-seq url-pattern line))]
(do (if (not (empty? urls))
(println ">" sender ":" line) (doseq [url urls] (println "Url Found >" url))
(process-line (when-not (= 1 (count (tokenize-line line)))
line (do
trained-map-atom (println ">" sender ":" line)
startphrase-list-atom (process-line
endphrase-set-atom line
line-list-atom trained-map-atom
key-size startphrase-list-atom
history-size) endphrase-set-atom
(when (and line-list-atom
@bot-talking? key-size
(<= (rand) (/ 1 speak-interval))) history-size)
(create-statement-and-send))))) (when (and
@bot-talking?
(<= (rand) (/ 1 speak-interval)))
(create-statement-and-send)))))))
(when (zero? (mod @msg-count save-interval)) (when (zero? (mod @msg-count save-interval))
(println "Saving memory") (println "Saving memory")
(save-memory bot @line-list-atom)))) (save-memory bot @line-list-atom))))
@ -334,6 +341,5 @@
(parse-int min-sentence-length) (parse-int min-sentence-length)
(parse-int max-sentence-length))) (parse-int max-sentence-length)))
;;filter out links
;;switch to pircbotx ;;switch to pircbotx
;;pronoun substitution ;;pronoun substitution