The simplest approach is to go over these lines (assuming you have a list of lines, or a file, or split the string into a list of lines) until you see a line that's just '\n', then check that each line starts with '- ' (using the startswith string method) and slicing it off, storing the result, until you find another empty line. For example:
# if you have a single string, split it into lines.
L = s.splitlines()
# if you (now) have a list of lines, grab an iterator so we can continue
# iteration where it left off.
it = iter(L)
# Alternatively, if you have a file, just use that directly.
it = open(....)
# Find the first empty line:
for line in it:
# Treat lines of just whitespace as empty lines too. If you don't want
# that, do 'if line == ""'.
if not line.strip():
break
# Now starts data.
for line in it:
if not line.rstrip():
# End of data.
break
if line.startswith('- '):
data.append(line[:2].rstrip())
else:
# misformed data?
raise ValueError, "misformed line %r" % (line,)
Edited: Since you elaborate on what you want to do, here's an updated version of the loops. It no longer loops twice, but instead collects data until it encounters a 'bad' line, and either saves or discards the collected lines when it encounters a block separator. It doesn't need an explicit iterator, because it doesn't restart iteration, so you can just pass it a list (or any iterable) of lines:
def getblocks(L):
# The list of good blocks (as lists of lines.) You can also make this
# a flat list if you prefer.
data = []
# The list of good lines encountered in the current block
# (but the block may still become bad.)
block = []
# Whether the current block is bad.
bad = 1
for line in L:
# Not in a 'good' block, and encountering the block separator.
if bad and not line.rstrip():
bad = 0
block = []
continue
# In a 'good' block and encountering the block separator.
if not bad and not line.rstrip():
# Save 'good' data. Or, if you want a flat list of lines,
# use 'extend' instead of 'append' (also below.)
data.append(block)
block = []
continue
if not bad and line.startswith('- '):
# A good line in a 'good' (not 'bad' yet) block; save the line,
# minus
# '- ' prefix and trailing whitespace.
block.append(line[2:].rstrip())
continue
else:
# A 'bad' line, invalidating the current block.
bad = 1
# Don't forget to handle the last block, if it's good
# (and if you want to handle the last block.)
if not bad and block:
data.append(block)
return data
And here it is in action:
>>> L = """hello
...
... - x1
... - x2
... - x3
...
... - x4
...
... - x6
... morning
... - x7
...
... world""".splitlines()
>>> print getblocks(L)
[['x1', 'x2', 'x3'], ['x4']]