"""
filter.py: utility functions for filtering data through a command, with
optional persistent caching.
Adam Sampson <ats@offog.org>
"""

def filter_through(command, data):
	"""Filter data through a command (an execvp-style list), and return the
	result. Die if the command fails."""

	(childi, src) = os.pipe()
	(dest, childo) = os.pipe()

	pid = os.fork()
	if pid == 0:
		os.close(src)
		os.close(dest)
		os.dup2(childi, 0)
		os.dup2(childo, 1)
		os.execvp(command[0], command)
		os._exit(1)
	os.close(childi)
	os.close(childo)

	output = []
	pos = 0
	max = len(data)
	chunk = 4096
	while 1:
		if src is None:
			ofs = []
		else:
			ofs = [src]
		(ifs, ofs, efs) = select.select([dest], ofs, [])
		if dest in ifs:
			d = os.read(dest, chunk)
			if d == "":
				os.close(dest)
				break
			output.append(d)
		if src in ofs:
			n = os.write(src, data[pos:pos + chunk])
			pos += n
			if pos == max:
				os.close(src)
				src = None
	if src is not None:
		die("output EOF before input EOF:", command)

	(pid, estat) = os.waitpid(pid, 0)
	if not os.WIFEXITED(estat) and os.WEXITSTATUS(estat) == 0:
		die("command failed:", command)

	return "".join(output)

class FilterCache:
	"""Provide caching for filter_through in a given file, expiring items
	that haven't been used."""

	def __init__(self, file):
		self.file = file
		try:
			f = open(file)
			self.cache = pickle.load(f)
			f.close()
		except IOError:
			self.cache = {}
		for k in self.cache:
			self.cache[k][0] = False

	def close(self):
		for k, v in self.cache.items():
			if not v[0]:
				del self.cache[k]
		tempname = self.file + ".new-%d" % (os.getpid(),)
		f = open(tempname, "w")
		pickle.dump(self.cache, f)
		f.close()
		os.rename(tempname, self.file)

	def filter(self, command, data):
		key = (command, data)
		if key in self.cache:
			self.cache[key][0] = True
		else:
			self.cache[key] = [True, filter_through(command, data)]
		return self.cache[key][1]