raek

(ns my-file-utils
  (:import java.io.File))

;; Java methods cannot be used as Clojure functions,
;; so you have to wrap them in functions.

;; anonymous function, map, and filter version
(defn list-files [d]
  (map (fn [f] (.getName f))
       (filter (fn [f] (.isDirectory f))

raek

(def pattern_graph
     [[0 1 0 1 1 1 0 1 0]
      [1 0 1 1 1 1 1 0 1]
      [0 1 0 1 1 1 0 1 0]
      [1 1 1 0 1 0 1 1 1]
      [1 1 1 1 0 1 1 1 1]
      [1 1 1 0 1 0 1 1 1]
      [0 1 0 1 1 1 0 1 0]
      [1 0 1 1 1 1 1 0 1]
      [0 1 0 1 1 1 0 1 0]])

raek

(ns ignacio.tfidf (:require [clojure.contrib.string  :as string])) ;; Simple tfidf in clojure, for fun.

(def *stopwords* (set (string/split #"\n" (slurp "./stopwords.txt"))))

(defn tokenize [raw-text] ;; Lowercases and splits on non-letters, non-numbers.
  (remove *stopwords* (string/split #"[^a-z0-9äöüáéíóúãâêîôûàèìòùçñ]+" (string/lower-case raw-text))))

(defn idf2 [n-docs match] (Math/pow (Math/log (/ n-docs (count (keys match)))) 2))

(defn index-one [fname] ;; Index for one file. Given an fname, returns a map of token -> map of (fname, count)

raek

(defmacro create-object [name props funcs]
  (let [m (reduce #(conj %1 [%2 nil]) {} props)
        d (symbol (str "dispatcher-" name))
        msg-sym (gensym "msg")
        params-sym (gensym "params")]
    `(defn ~d [~msg-sym &~params-sym]
       (cond
         ~@(for [f funcs]
             [(= ~msg-sym ~(first f))
              (apply ~(second f) ~m ~params-sym)])))))