Save videos, fetch gdata, rename, convert to MP3, clean up

#!/usr/bin/python
#/home/clockworkpc/bin/tubemixingsave.py

# Released under a GPLv3 Licence by Clockwork PC 2013
# www.clockworkpc.com.au

# You are entitled to the following four freedoms:
# Freedom 0: To run this program for any purpose
# Freedom 1: To study how this program works and change it to make it do what you wish
# Freedom 2: To redistribute copies so you can help your neighbour
# Freedom 3: To distribute copies of your modified version to others

import os, datetime, re, shutil

now = datetime.datetime.now()
nowstring = str(now.strftime("%Y-%m-%d_%A_%H%M%S"))

#print nowstring

tubemixingFolder = "/tmp/tubemixing/"
videoFolder = os.getenv("HOME") + "/Videos/"
saveFolder = videoFolder + "tubemixing/"
originalVideosFolder = saveFolder + "OriginalVideos/"
namedVideosFolder = saveFolder + "NamedVideos/"
namedVideosFolderDateStamped = namedVideosFolder + nowstring + "/"
videoListsFolder = saveFolder + "VideoLists/"
videoListsFile = videoListsFolder + nowstring
videoDataFolder = saveFolder + "gdata_text_files/"
videoDataFolderDatestamped = saveFolder + "gdata_text_files/" + nowstring + "/"
mp3Folder = os.getenv("HOME") + "/Music/tubemixing/" + nowstring + "/"
gdataFeed = "http://gdata.youtube.com/feeds/api/videos/"
originalVideosList = originalVideosFolder + "original_videos_list"
originalVideosListBackup = originalVideosFolder + "original_videos_list_" + nowstring + ".backup"

#Check for new videos in Tubemixing /tmp folder
def check_for_video_list():
    running = True
    for names in os.listdir(originalVideosFolder):
        if names.startswith("original_videos_list"):
            if not names.endswith(".backup"):
                print "found original"
                os.system("cp -v " + originalVideosList + " " + originalVideosListBackup)
                running = False
check_for_video_list()

#Create date-stamped folders
os.system("mkdir " + namedVideosFolderDateStamped)
os.system("mkdir " + mp3Folder)
os.system("mkdir " + videoDataFolderDatestamped)

# Match only videos that haven't already been saved
a = os.listdir(tubemixingFolder)
b = os.listdir(originalVideosFolder)
c = open(originalVideosList, "r")
d = c.read()
e = [x for x in a if x not in d]
print e

#Replenish list
#f = open(originalVideosList, "a")
#f.write(str(os.listdir(originalVideosFolder)))
#f.close()

# Match only videos that haven't already been saved (replaced)
#a = os.listdir(tubemixingFolder)
#b = os.listdir(originalVideosFolder)
#c = [x for x in a if x not in b]

# File-only os.listdir
def saveFolder_files_list():
    for file in os.listdir(saveFolder):
        if os.path.isfile(os.path.join(saveFolder,file)):
            return file

def files_only(saveFolder):
    return [filename for filename in os.listdir(saveFolder) if os.path.isfile(os.path.join(saveFolder, filename))]

def files_only(videoDataFolderDatestamped):
    return [filename for filename in os.listdir(videoDataFolderDatestamped) if os.path.isfile(os.path.join(videoDataFolderDatestamped, filename))]

# Copy NEW videos from Tubemixing temp folder to working folder
#for file in c:
for file in e:
    shutil.copyfile(tubemixingFolder+file,saveFolder+file)
# Update list of saved videos
g = open(originalVideosList, "a")
g.write(str(e))
g.close()

# Copy NEW videos to archive folder for future reference (replaced)
#   shutil.copyfile(tubemixingFolder+file,originalVideosFolder+file)

# Remove prefix
# NOTE: I set Tubemixing to save by default to 360p.
# You'll have to tweak it if you save higher resolution by default.
preFix1 = "tubemixing-360p-"
preFix2 = "tubemixing--"

#os.listdir(saveFolder)
for filename in files_only(saveFolder):
    if filename.startswith(preFix1):
        os.rename(saveFolder+filename,saveFolder+filename[16:])

# Second pass (some files are saved with preFix2)
os.listdir(saveFolder)
for filename in files_only(saveFolder):
    if filename.startswith(preFix2):
        os.rename(saveFolder+filename,saveFolder+filename[12:])

# Use gdata to fetch video title from YouTube
for filename in files_only(saveFolder):
    if len(filename) == 11:
        os.system("cd " + videoDataFolderDatestamped + " && wget " + gdataFeed + filename + " " + "-O" + " " + filename)

# Parse video title from data fetched from YouTube
print files_only(videoDataFolderDatestamped)
for filename in files_only(videoDataFolderDatestamped):
    f = open(videoDataFolderDatestamped+filename)
    for line in f:
        if "<title type='text'>" in line:
            x = line
            r = re.compile("<title type='text'>(.*?)</title><content")
            m = r.search(x)
            if m:
                titleInfo = m.group(1)
                print titleInfo
# Using video title, save properly named and suffixed files in a new date-stamped folder
# Using os.system instead of shutil because shutil is a FUCKING SLOW MEMORY HOG
                os.system("cp -v " + saveFolder+filename + " " + '"' + namedVideosFolderDateStamped+titleInfo+".webm" + '"')
#                   shutil.copyfile(saveFolder+filename,namedVideosFolderDateStamped+titleInfo+".webm")

# Remove code-named video files, because originals have already been saved for reference
# Unfortunately, the best I can do is use the clumsy rm command, which tries, but fails, to delete folders.  This is not the way to go, but I can't find a more elegant way of clearing the working folder.

for file in os.listdir(saveFolder):
    os.system("rm -v " + saveFolder + file)

# Use AVconv (FFmpeg) to extract MP3 audio into the Music folder
for filename in os.listdir(namedVideosFolderDateStamped):
    (str, filename)
    namedVideoFile = namedVideosFolderDateStamped + filename
    print namedVideoFile

    os.system("avconv -i " + '"' + namedVideoFile + '"' + " " + '"' + mp3Folder + filename + '.mp3"')

# Remove the ".webm" that carries over to the MP3 file name during the AVconv conversion

print
pathiter = (os.path.join(root, filename)
    for root, _, filenames in os.walk(mp3Folder)
    for filename in filenames
)
for path in pathiter:
    newname =  path.replace('.webm', '')
    if newname != path:
        os.rename(path,newname)

# Clear empty folders from directories that tend to get cluttered
work_path1 = os.getenv("HOME") + "/Videos/tubemixing/NamedVideos/"
work_path2 = os.getenv("HOME") + "/Music/tubemixing/"

def clear_named_videos_folders():
    for file in os.listdir(work_path1):
        if not os.listdir(work_path1+file):
            print "empty"
            os.rmdir(work_path1+file)
        else:
            print "not empty"

def clear_avconv_mp3_folders():
    for file in os.listdir(work_path2):
        if not os.listdir(work_path2+file):
            print "empty"
            os.rmdir(work_path2+file)
        else:
            print "not empty"

clear_named_videos_folders()
clear_avconv_mp3_folders()