Roll your own Media Database

From techdaze

Jump to: navigation, search

1 Roll your own Media Database
- 1.1 Metadata
- 1.2 BashAndGrep
  - 1.2.1 Resources

Roll your own Media Database

Example: Brxlbravo

Example: Metadata parser

Example: File Uploader

DoingSomethingWithAllFilesInaDirectory

Metadata

Metadata is simply information about a particular (media) file that's considered secondary to its primary content. For instance, consider a JPEG file: in addition to the actual pixels of image data, the image has dimensions, and a particular color depth. Also, a JPEG file produced by a digital camera often has information about the time the photos was taken, and various camera settings.

Examples of command line tools that display (and possibly allow manipulation of) a file's metadata:

identify (from ImageMagick)
exiftool (EXIF data from images)

+ ffmpeg (by default displays information about a movie such as frame rate, duration, image size)

BashAndGrep

#!python numbers=off
#

import os



def bashandgrep (cmdtodo, patterns):
	"""
	executes the bash command CMDTODO
: and scans its output for patterns, 
	if they match, captures (groups /
: parentheses in the pattern) are stored
	and returned in a new dictionary with
: names matching the keys of the patterns dict.
	"""
	ret = {}
	(child_stdin, child_stdout) = os.popen4(cmdtodo)

	output = child_stdout.read()
	for (name, pattern) in patterns.items():
		# print "scanning for %s" % name
		m = pattern.search(output)
		if m:
			caps = m.groups()
			if (len(caps) == 1): 
				ret[name] = caps[0]
			else:
				ret[name] = caps
	
	return ret

Example using ffmpeg to grab some moviedata:

#!python numbers=off
from bashandgrep import *
import re

patterns = {
	'fps': re.compile(r"([0-9.]*) fps"),
	'size' : re.compile(r"([0-9]+)x([0-9]+)"),
	'duration' : re.compile(r"Duration: ([0-9]+):([0-9]+):([0-9]+).([0-9]+)")
}

command = "ffmpeg -i ~/movies/anniehall.avi"
data = bashandgrep(command, patterns)
print data

#!python numbers=off
def sqlinsert (dbconn, tablename, data):
	sql = "INSERT INTO " + tablename
	sql += " SET "
	assignpairs = [ x+"=%s" for x in data.keys() ] 
	sql += ", ".join(assignpairs)
	
	# DEBUG
	# print "sqlinsert debug (dbconn is None)"
	# vlist = ["'"+str(x)+"'" for x in data.values()]
	# print "\t", sql % tuple(vlist)
	
	if dbconn:
		cursor = dbconn.cursor()
		cursor.execute(sql, data.values())
		cursor.close()