Almost done with this stupidness

2 files changed, 59 insertions, 0 deletions

diff --git a/old/html-cleaners/clean_all.scm b/old/html-cleaners/clean_all.scm
new file mode 100755
index 0000000..82091d9
--- /dev/null
+++ b/old/html-cleaners/clean_all.scm
@@ -0,0 +1,14 @@
+#! /usr/bin/guile \
+-e main -s
+!#
+
+(define (main args)
+  (for-each
+   (lambda (file)
+     (system (string-append "./clean_up_html.scm " file ".html"
+                            " | tidy | pandoc -f html -t org > "
+                            "org/" file ".org")))
+   '("index" "linux_room" "music_room" "nathans_philosophy"
+     "truth" "ttr" "tv_corner" "distro_guides/arch"
+     "distro_guides/fedora" "distro_guides/xfce"
+     "error/404")))
diff --git a/old/html-cleaners/clean_up_html.scm b/old/html-cleaners/clean_up_html.scm
new file mode 100755
index 0000000..6ee81c5
--- /dev/null
+++ b/old/html-cleaners/clean_up_html.scm
@@ -0,0 +1,45 @@
+#! /usr/bin/guile \
+-e main -s
+!#
+
+;; Loads an entire text file into a string
+(define (read-entire-file-port fp)
+  (letrec ((read-characters
+            (lambda (fp)
+              (let ((character (read-char fp)))
+                (if (eof-object? character)
+                    ""
+                    (string-append (string character)
+                                   (read-characters fp)))))))
+    (read-characters fp)))
+
+(define (trim-words-all content word)
+  (let ((index (string-contains content word)))
+    (if index
+        (trim-words-all (string-append
+                         (substring content 0 index)
+                         (substring content
+                                    (+ index (string-length word))
+                                    (string-length content)))
+                        word)
+        content)))
+
+(define (clean-up-html file-name)
+  (call-with-input-file file-name
+    (lambda (fp)
+      (let ((content (read-entire-file-port fp))
+            (remove-list '("<table border=\"1\" width=\"60%\">"
+                           "<table>"
+                           "</table>"
+                           "<tr>"
+                           "<td>"
+                           "</tr>"
+                           "</td>")))
+        (for-each
+         (lambda (item)
+           (set! content (trim-words-all content item)))
+         remove-list)
+        (display content)))))
+
+(define (main args)
+  (clean-up-html (cadr args)))