Add initial code

2 files changed, 52 insertions, 1 deletions

diff --git a/__init__.py b/__init__.py
new file mode 100644
index 0000000..0104d8f
--- /dev/null
+++ b/__init__.py
@@ -0,0 +1,48 @@
+
+from pelican import signals
+
+import multiprocessing
+import htmlmin
+import os
+import re
+
+import logging
+logger = logging.getLogger(__name__)
+
+
+def run(pelican):
+    options = pelican.settings.get('MINIFY', {})
+    htmlfile = re.compile(r'.html?$')
+    pool = multiprocessing.Pool()
+    
+    for base, dirs, files in os.walk(pelican.settings['OUTPUT_PATH']):
+        for f in files:
+            # filter only html files
+            if not htmlfile.search(f):
+                continue
+            # add full html path to wooker pool
+            filepath = os.path.join(base, f)
+            pool.apply_async(worker, (filepath, options))
+
+    pool.close()
+    pool.join()
+
+
+def worker(filepath, options):
+    rawhtml = open(filepath, encoding='utf-8').read()
+    with open(filepath, 'w', encoding='utf-8') as f:
+        try:
+            logger.debug('Minifying: %s' % filepath)
+            compressed = htmlmin.minify(rawhtml, **options)
+            f.write(compressed)
+        except Exception as e:
+            logger.critical('Minification failed: {}'.format(e))
+        finally:
+            f.close()
+
+
+def register():
+    """
+    Minify HTML at the end of the processing loop
+    """
+    signals.finalized.connect(run)
diff --git a/readme.md b/readme.md
index 04f95e6..898c80c 100644
--- a/readme.md
+++ b/readme.md
@@ -1,2 +1,5 @@
 # pelican-minify
-An Pelican plugin to use htmlmin to remove extra whitespace.
+An [Pelican] plugin to use htmlmin to remove extra whitespace from webpages.
+
+[Pelican]: "http://pelican.readthedocs.org/en/latest/"
+