diff --git a/.gitmodules b/.gitmodules new file mode 100644 index 0000000..6c3e94c --- /dev/null +++ b/.gitmodules @@ -0,0 +1,3 @@ +[submodule "is-wordpress"] + path = is-wordpress + url = git@github.com:tanrax/is-wordpress.git diff --git a/calculate-percentage.sh b/calculate-percentage.sh index d39bf9e..18a9cf3 100644 --- a/calculate-percentage.sh +++ b/calculate-percentage.sh @@ -5,4 +5,4 @@ NUM_NOT_WORDPRESS=$(cat $CSV | grep ,false$ | wc -l) NUM_TIMEOUT=$(cat $CSV | grep ,timeout$ | wc -l) TOTAL=$(cat $CSV | wc -l) -echo "($NUM_WORDPRESS * 100) / ($TOTAL - $NUM_TIMEOUT)" | bc +echo "($NUM_WORDPRESS * 100) / ($TOTAL - $NUM_TIMEOUT)" | bc | awk '{print $1"%"}' diff --git a/is-wordpress b/is-wordpress new file mode 160000 index 0000000..b64770e --- /dev/null +++ b/is-wordpress @@ -0,0 +1 @@ +Subproject commit b64770eb63dcb3924957bed824e79d380089e2c2 diff --git a/src/wordpress_used/core.clj b/src/wordpress_used/core.clj index 6763b73..45664ce 100644 --- a/src/wordpress_used/core.clj +++ b/src/wordpress_used/core.clj @@ -5,10 +5,10 @@ [clojure.java.shell :as shell] ) (:gen-class)) -(defn request - "Make a request by means of curl" +(defn wordpress? + "Check if a web page is generated with WordPress" [url] - (shell/sh "curl" "-L" "-m" "5" "-H" "User-Agent: Firefox" url)) + (= (clojure.string/trim-newline (:out (shell/sh "bash" "./is-wordpress/is-wordpress" url))) "true")) (defn read-csv-domains @@ -17,11 +17,6 @@ (with-open [reader (io/reader (io/resource url))] (doall (csv/read-csv reader)))) -(defn wordpress? - "Check if a web page is generated with WordPress" - [url] - (let [response (request url)] - (every? identity [(re-find (re-pattern "meta.*generator.*WordPress") (:out response))]))) (defn -main [& args]