--- /dev/null
+# coding: utf-8
+from __future__ import unicode_literals
+
+import re
+
+from .common import InfoExtractor
+from ..utils import (
+ decode_packed_codes,
+ js_to_json,
+)
+
+
+class FileMoonIE(InfoExtractor):
+ _VALID_URL = r'https?://(?:www\.)?filemoon\.sx/./(?P<id>\w+)'
+ _TEST = {
+ 'url': 'https://filemoon.sx/e/dw40rxrzruqz',
+ 'md5': '5a713742f57ac4aef29b74733e8dda01',
+ 'info_dict': {
+ 'id': 'dw40rxrzruqz',
+ 'title': 'dw40rxrzruqz',
+ 'ext': 'mp4'
+ }
+ }
+
+ def _real_extract(self, url):
+ video_id = self._match_id(url)
+
+ webpage = self._download_webpage(url, video_id)
+ matches = re.findall(r'(?s)(eval.*?)</script>', webpage)
+ packed = matches[-1]
+ unpacked = decode_packed_codes(packed)
+ jwplayer_sources = self._parse_json(
+ self._search_regex(
+ r'(?s)player\s*\.\s*setup\s*\(\s*\{\s*sources\s*:\s*(.*?])', unpacked, 'jwplayer sources'),
+ video_id, transform_source=js_to_json)
+
+ formats = self._parse_jwplayer_formats(jwplayer_sources, video_id)
+
+ return {
+ 'id': video_id,
+ 'title': self._generic_title(url) or video_id,
+ 'formats': formats
+ }