gistfile1.clj

(ns mol
  (:use cascalog.api)
  (:require [clojure.string :as s]
            [cascalog.ops :as c]
            [clojure.java.io :as io]))

;; Tab (\t) separated text file:
;;
;; gbif-id         latitude        longitude       nub-id         name
;; 236499193       24.663458       121.61801       1986930        Abaciscus
;; 236499220       23.455833       120.89278       1986930        Abaciscus
;; 236499202       24.663458       121.61801       1986930        Abaciscus
;;
(def gbif-points-path
  "Returns our test data as a File object."
  (io/as-file (io/resource "testdata/mol-1000.txt")))

(defn to-floats [& xs]
  "Converts a sequence of numbers to floats."
  (map #(Float. %) xs))

(defn make-dwc-tuples [path]
  "Returns a Cascalog query that returns [lat, lng, name, count] where count is
  the number of times the name appears in the file."
  (let [src (hfs-textline path)
        count-q (<- [?name ?species-count]
                    (src ?textline) 
                    (s/split ?textline #"\t" :> _ _ _ _ ?name)
                    (c/count ?species-count))]
    (<- [?lat-float ?lon-float ?name ?species-count]
        (src ?textline)
        (s/split ?textline #"\t" :> _ ?lat ?lon _ ?name)
        (count-q ?name ?species-count)
        (to-floats ?lat ?lon :> ?lat-float ?lon-float)
        (:distinct false))))

(defn run-mol-job []
  (?- (stdout) (make-dwc-tuples gbif-points-path)))