find_plex_unscanned.py

#!/usr/bin/python
import sqlite3
import os
import urllib
import unicodedata
import getopt
import sys
import codecs

# Find unscanned files not in Plex database but on filesystem
# Version 0.03
# - 0.01 - Initial release
# - 0.02 - Go through all sections and verify DB exists and optionally write results to file
# - 0.03 - Add option to generate lists of matched files for each section (-l)

VIDEO_EXTENSIONS = ['.m4v', '.3gp', '.nsv', '.ts', '.ty', '.strm', '.rm', '.rmvb', '.m3u', '.mov', '.qt', '.divx', '.xvid', '.bivx', '.vob', '.nrg', '.img', '.iso', '.pva', '.wmv', '.asf', '.asx', '.ogm', '.m2v', '.avi', '.bin', '.dat', '.dvr-ms', '.mpg', '.mpeg', '.mp4', '.mkv', '.avc', '.vp3', '.svq3', '.nuv', '.viv', '.dv', '.fli', '.flv', '.rar', '.001', '.wpl', '.zip', '.jpg']
OK_FILES = ['imdb.nfo', '.DS_Store', '__db.001']
OTHER_EXTENSIONS = ['.srt', '.sub', '.idx', '.jpg']

def usage():
    print "Usage: python %s [OPTION]" % (__file__)
    print "Options:"
    print "-f, --file <filename>        Write out results to <filename>"
    print "-h, --help   This message"
    print "-i, --ignores        Display list of files that are within paths that aren't video extensions or subtitles"
    print "-l, --list	Create files listing all matched files in plex"
    return

def listTree(top, files=list()):
    r = files[:]
    if not os.path.exists(top):
        print 'Not mounted: ' + top
        return r
    for f in os.listdir(top):
        pathname = os.path.join(top, f)
        if os.path.isdir(pathname):
            r = listTree(pathname, r)
        elif os.path.isfile(pathname):
            r.append(unicodedata.normalize('NFC', pathname))
        else:
            print 'Skipping %s' % pathname
    return r

def main():
    display_ignores = False
    fh = None
    display_matches = False

    try:
        opts, args = getopt.getopt(sys.argv[1:], "f:ihl", ["file=", "ignores", "help", "list"])
    except getopt.GetoptError,exc:
        print exc.msg
        sys.exit(2)
    for opt, arg in opts:
        if opt in ("-f", "--file"):
            filename = arg
            fh = codecs.open(filename, "w", "utf-8")
        if opt in ("-i", "--ignores"):
            display_ignores = True
        if opt in ("-h", "--help"):
            usage()
            exit()
        if opt in ("-l", "--list"):
            display_matches = True

    dbPath = os.path.expanduser('~/Library/Application Support/Plex Media Server/Plug-in Support/Databases/com.plexapp.plugins.library.db')

    #Linux DB path - TODO (detect OS or allow specification via command line)
    #dbPath = '/var/lib/plexmediaserver/Library/Application Support/Plex Media Server/Plug-in Support/Databases/com.plexapp.plugins.library.db'

    if (not os.path.isfile(dbPath)):
        print "Cannot find database at %s. Running PMS locally on this Mac?\nExiting." % (dbPath)
        exit()

    conn = sqlite3.connect(dbPath)
    rows = conn.cursor().execute('select L.id, L.name from library_sections L').fetchall()
    for row in rows:
        section_id = row[0]
        section_name = row[1]
        print "\n\nScanning %s[%s]" % (section_name, section_id)
        root_paths = conn.cursor().execute('select root_path from section_locations where library_section_id=?', [section_id,]).fetchall()

        basePaths = map(lambda p: p[0], root_paths)

        print "Composed of %s" % (", ".join(basePaths))

        paths = list()
        rows = conn.cursor().execute('SELECT mp.file, md.title from media_parts mp JOIN media_items mi ON mp.media_item_id=mi.id JOIN metadata_items md ON mi.metadata_item_id=md.id  WHERE mi.library_section_id=?', [section_id,]).fetchall()
        paths = map(lambda p: unicodedata.normalize('NFC', p[0]), rows)

        if (display_matches):
            match_fn = "plex_matches_%s.txt" % (section_id)
            print "\tWriting matches to %s..." % (match_fn)
            match_fh = codecs.open(match_fn, "w", "utf-8")
            rows.sort(key=lambda x: x[1])
            for row in rows:
                match_fh.write("%s -- %s\n" % (row[1], row[0]))
            match_fh.close()


        missing = list()
        for basePath in basePaths:
            print "\tHandling path %s ..." % basePath
            filePaths = listTree(basePath)
            for filePath in filePaths:
                if filePath not in paths:
                    cext = os.path.splitext(filePath)[1].lower()
                    fname = os.path.split(filePath)[1]
                    if (fname in OK_FILES):
                        #Don't do anything for acceptable files
                        pass
                    elif (cext in OTHER_EXTENSIONS):
                        #ignore images and subtitles
                        pass
                    elif (cext not in VIDEO_EXTENSIONS):
                        #these shouldn't be here
                        if (display_ignores):
                            print 'Ignoring %s' % filePath
                        pass
                    else:
                        missing.append(filePath)

            print '\tfinished %s' % basePath
        print '\nMissing %i items:' % len(missing)

        if (fh):
            fh.write("\nSection %s:\n" % (section_name))
            fh.write("Missing %i items\n" % len(missing))

        for item in missing:
            print item
            if (fh):
                fh.write(item+"\n")



    if (fh):
        fh.close()

    conn.close()
    exit()

if __name__ == "__main__":
    main()