' % (self.steps,)
def reset(self):
"""t.reset() restores a pipeline template to its initial state."""
self.steps = []
def clone(self):
"""t.clone() returns a new pipeline template with identical
initial state as the current one."""
t = Template()
t.steps = self.steps[:]
t.debugging = self.debugging
return t
def debug(self, flag):
"""t.debug(flag) turns debugging on or off."""
self.debugging = flag
def append(self, cmd, kind):
"""t.append(cmd, kind) adds a new step at the end."""
if type(cmd) is not type(''):
raise TypeError, \
'Template.append: cmd must be a string'
if kind not in stepkinds:
raise ValueError, \
'Template.append: bad kind %r' % (kind,)
if kind == SOURCE:
raise ValueError, \
'Template.append: SOURCE can only be prepended'
if self.steps and self.steps[-1][1] == SINK:
raise ValueError, \
'Template.append: already ends with SINK'
if kind[0] == 'f' and not re.search(r'\$IN\b', cmd):
raise ValueError, \
'Template.append: missing $IN in cmd'
if kind[1] == 'f' and not re.search(r'\$OUT\b', cmd):
raise ValueError, \
'Template.append: missing $OUT in cmd'
self.steps.append((cmd, kind))
def prepend(self, cmd, kind):
"""t.prepend(cmd, kind) adds a new step at the front."""
if type(cmd) is not type(''):
raise TypeError, \
'Template.prepend: cmd must be a string'
if kind not in stepkinds:
raise ValueError, \
'Template.prepend: bad kind %r' % (kind,)
if kind == SINK:
raise ValueError, \
'Template.prepend: SINK can only be appended'
if self.steps and self.steps[0][1] == SOURCE:
raise ValueError, \
'Template.prepend: already begins with SOURCE'
if kind[0] == 'f' and not re.search(r'\$IN\b', cmd):
raise ValueError, \
'Template.prepend: missing $IN in cmd'
if kind[1] == 'f' and not re.search(r'\$OUT\b', cmd):
raise ValueError, \
'Template.prepend: missing $OUT in cmd'
self.steps.insert(0, (cmd, kind))
def open(self, file, rw):
"""t.open(file, rw) returns a pipe or file object open for
reading or writing; the file is the other end of the pipeline."""
if rw == 'r':
return self.open_r(file)
if rw == 'w':
return self.open_w(file)
raise ValueError, \
'Template.open: rw must be \'r\' or \'w\', not %r' % (rw,)
def open_r(self, file):
"""t.open_r(file) and t.open_w(file) implement
t.open(file, 'r') and t.open(file, 'w') respectively."""
if not self.steps:
return open(file, 'r')
if self.steps[-1][1] == SINK:
raise ValueError, \
'Template.open_r: pipeline ends width SINK'
cmd = self.makepipeline(file, '')
return os.popen(cmd, 'r')
def open_w(self, file):
if not self.steps:
return open(file, 'w')
if self.steps[0][1] == SOURCE:
raise ValueError, \
'Template.open_w: pipeline begins with SOURCE'
cmd = self.makepipeline('', file)
return os.popen(cmd, 'w')
def copy(self, infile, outfile):
return os.system(self.makepipeline(infile, outfile))
def makepipeline(self, infile, outfile):
cmd = makepipeline(infile, self.steps, outfile)
if self.debugging:
print cmd
cmd = 'set -x; ' + cmd
return cmd
def makepipeline(infile, steps, outfile):
# Build a list with for each command:
# [input filename or '', command string, kind, output filename or '']
list = []
for cmd, kind in steps:
list.append(['', cmd, kind, ''])
#
# Make sure there is at least one step
#
if not list:
list.append(['', 'cat', '--', ''])
#
# Take care of the input and output ends
#
[cmd, kind] = list[0][1:3]
if kind[0] == 'f' and not infile:
list.insert(0, ['', 'cat', '--', ''])
list[0][0] = infile
#
[cmd, kind] = list[-1][1:3]
if kind[1] == 'f' and not outfile:
list.append(['', 'cat', '--', ''])
list[-1][-1] = outfile
#
# Invent temporary files to connect stages that need files
#
garbage = []
for i in range(1, len(list)):
lkind = list[i-1][2]
rkind = list[i][2]
if lkind[1] == 'f' or rkind[0] == 'f':
(fd, temp) = tempfile.mkstemp()
os.close(fd)
garbage.append(temp)
list[i-1][-1] = list[i][0] = temp
#
for item in list:
[inf, cmd, kind, outf] = item
if kind[1] == 'f':
cmd = 'OUT=' + quote(outf) + '; ' + cmd
if kind[0] == 'f':
cmd = 'IN=' + quote(inf) + '; ' + cmd
if kind[0] == '-' and inf:
cmd = cmd + ' <' + quote(inf)
if kind[1] == '-' and outf:
cmd = cmd + ' >' + quote(outf)
item[1] = cmd
#
cmdlist = list[0][1]
for item in list[1:]:
[cmd, kind] = item[1:3]
if item[0] == '':
if 'f' in kind:
cmd = '{ ' + cmd + '; }'
cmdlist = cmdlist + ' |\n' + cmd
else:
cmdlist = cmdlist + '\n' + cmd
#
if garbage:
rmcmd = 'rm -f'
for file in garbage:
rmcmd = rmcmd + ' ' + quote(file)
trapcmd = 'trap ' + quote(rmcmd + '; exit') + ' 1 2 3 13 14 15'
cmdlist = trapcmd + '\n' + cmdlist + '\n' + rmcmd
#
return cmdlist
# Reliably quote a string as a single argument for /bin/sh
# Safe unquoted
_safechars = frozenset(string.ascii_letters + string.digits + '@%_-+=:,./')
def quote(file):
"""Return a shell-escaped version of the file string."""
for c in file:
if c not in _safechars:
break
else:
if not file:
return "''"
return file
# use single quotes, and put single quotes into double quotes
# the string $'b is then quoted as '$'"'"'b'
return "'" + file.replace("'", "'\"'\"'") + "'"