text.clj

(ns littlesis.names
  (:require [net.cgrand.enlive-html :as html]))

(def *base-url* "http://blog.littlesis.org/2010/02/18/sources-allege-goldman-sachs-and-john-paulson-are-partnering-in-speculative-attacks-o\
n-greece/")

(defn fetch-url [url]
  (html/html-resource (java.net.URL. url)))

(defn entry [node]
  (first (html/select node [:div.entry])))

;; could probably be better expressed as reduce                                                                                             
(defn text-nodes [node]
  (let [content (:content node)]
    (loop [subnode (first content) nodes (rest content) result []]
      (if (nil? subnode)
        result
        (if (= (type subnode) java.lang.String)
          (recur (first nodes) (rest nodes) (conj result subnode))
          (recur (first nodes) (rest nodes) (into result (text-nodes subnode))))))))