andros/calculate-wordpress-usage
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Add cores

Browse Source
This commit is contained in:
Andros Fenollosa 2021-09-27 18:37:34 +02:00
parent ffa15e4178
commit cfb7f1dcb9
1 changed files with 25 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
src/wordpress_used/core.clj
View File

@ -2,36 +2,51 @@
(:require
[clojure.data.csv :as csv]
[clojure.java.io :as io]
[clojure.java.shell :as shell]
) (:gen-class))
[clojure.java.shell :as shell]) (:gen-class))
(defn wordpress?
"Check if a web page is generated with WordPress"
[url]
(= (clojure.string/trim-newline (:out (shell/sh "bash" "./is-wordpress/is-wordpress" url))) "true"))
(defn read-csv-domains
"Read CSV file with all domains"
[url]
(with-open [reader (io/reader (io/resource url))]
(doall (csv/read-csv reader))))
(defn get-chunk-list
"Cuts a list by the maximum number of fragments and returns the selected fragment."
[items-list chunk max-chunks]
(let [list-size (count items-list)
chunk-size (Math/ceil (/ list-size max-chunks))
chunk-start (int (* chunk chunk-size))
chunk-end (int (+ (* chunk chunk-size) chunk-size))]
(subvec items-list chunk-start (if (>= chunk-end list-size) list-size chunk-end))))
(defn analyse-list-chunk
"Analyse only the given list one chunck"
[items-list chunk max-chunks]
(doseq [domain-data (get-chunk-list items-list chunk max-chunks)] (let [line (get domain-data 0)
url (get domain-data 1)]
;; Show info
(prn (str line " " url))
;; Edit domains-csv with check WordPress
(shell/sh "sed" "-i" (str line "s/nil/" (wordpress? url) "/g") (str "resources/" file-csv)))))
(defn -main
[& args]
(let [;; Name of the file containing the CSV with the domains
file-csv "top-1m.csv"
file-csv (first args)
;; Number of threads to be executed.
number-of-threads (second args)
;; Get domains from CSV
domains-csv (vec (read-csv-domains file-csv))
;; Filters leaving those that have not been checked
domains-unchecked (vec (filter #(= (get % 2) "nil") domains-csv))]
;; List with domains with a boolean indicating if it is generate or not in WordPress
(prn "Start")
(doseq [domain-data domains-unchecked] (let [line (get domain-data 0)
url (get domain-data 1)]
;; Show info
(prn (str line " " url))
;; Edit domains-csv with check WordPress
(shell/sh "sed" "-i" (str line "s/nil/" (wordpress? url) "/g") (str "resources/" file-csv))))
(dotimes [i number-of-threads] (.start (Thread. (fn []
(analyse-list-chunk domains-uncheked i number-of-threads)))))
(prn "Complete")))