Permalink
Browse files

Added comments

  • Loading branch information...
1 parent d6d7d54 commit 6af79b6d2aac7706218bd6988abfbdc74c483fde @julosaure committed Aug 17, 2012
Showing with 11 additions and 3 deletions.
  1. +11 −3 src/SduDataImporter/core.clj
@@ -12,22 +12,27 @@
(def columnNames ["activity row id" "created date" "primary contact id" "contact integration id" "case" "member" "parish data" "start time" "call type" "type" "inin id" "parish" "interaction time" "resolution" "notes" "division" "program" "task status" "contact external unique id" "subject" "description" "contact type" "caller id" "crmit inin id" "validation status" "worker office number"])
+(def dirToProcess "/Users/julien/Documents/Output/")
+
+(def patFilesToProcess #"LAActivity_AIMA_Part_(.*)\.txt")
+
+;; UTILS
+
(def connection (clojure-solr/connect "http://127.0.0.1:8983/solr"))
(def patDateTime #"(\d{2})/(\d{2})/(\d{4}) (\d{2}:\d{2}:\d{2})")
-(def dirToProcess "/Users/julien/Documents/Output/")
-
-(def patFilesToProcess #"LAActivity_AIMA_Part_(.*)\.txt")
;; FUNCTIONS
(defn changeDateFormat [date]
+ "Change date format from '23-11-2012 03:34:51' to solr format '2012-11-23T03:34:51Z'"
(let [match (re-matches patDateTime date)]
(if match
(str (nth match 3) "-" (nth match 2) "-" (nth match 1) "T" (nth match 4) "Z"))))
(defn lineToMap [line]
+ "Transforms the sequence of fields into a map, and change date fields format"
(let [mapLine (zipmap columnNames (clojure.string/split line columnSeparator))
keys ["created date" "start time"]]
;(println (str mapLine))
@@ -36,13 +41,15 @@
(defn indexLine [line]
+ "Index a line"
(let [mapLine (lineToMap line)]
(if-not (= 0 (compare firstColName (mapLine "activity row id")))
(clojure-solr/with-connection connection
(clojure-solr/add-document! mapLine)))))
(defn parseFile [fileName]
+ "Parse a file"
(println (str "Reading file " (.getName fileName)))
(with-open [rdr (clojure.java.io/reader fileName)]
(doseq [line (line-seq rdr)]
@@ -53,6 +60,7 @@
(clojure-solr/commit!)))
(defn parseDir [dirName]
+ "Parse a directory"
(println (str "Reading directory " dirName))
(doseq [f (.listFiles (File. dirName))]
(if (re-matches patFilesToProcess (.getName f))

0 comments on commit 6af79b6

Please sign in to comment.