如何使用 Pyparsing 解析字符串，其中涉及忽略异常并移至下一个分隔符？

Question

I have a program that requires a list of effects, followed by start time, and end time. So I have this string that you acquire from user input (it can be faulty), and I'm trying to parse the relevant information and ignore faulty information, while moving to the next effect, after each ";". However I'm not quite sure how to use the Pyparsing library to do this, and I'm wondering whether this can be done purely with the library. The comments within the code signify what it should return, and the output below is what it actually returns.

import pyparsing as pp

testcase = "bounce, 5, 10; shutter, 12, 14" # returns [[bounce, 5, 10], [shutter, 12, 14]]
testcase2= "bounce, 5, 10; shutter, 12, 14; low_effort, 2, 23" # returns [[bounce, 5, 10], [shutter, 12, 14], [low_effort, 2, 23]]
testcase3= "_lolw, a, 2; effect, 6;" # returns [[effect, 6, None]]
testcase4= "bounce, 1, 10; effect, 5, a; bounce, 2, 10" # returns [[bounce, 1, 10], [bounce, 2, 10]]
testcase5= ";;;effect, 10; bounce, a, 1; bounce, 3, 10" # returns [[effect, 10, None], [bounce, 3, 10]]
testcase6= "effect, b, a; 9, 10, 11; max9, 10, 11; here, 2, 3; !b, 1, 2;;;" # returns [[here, 2, 3]]

def parseKeyframes(string: str):
    comma = pp.Suppress(",")
    pattern = pp.Word(pp.alphas + "_") + comma + pp.Word(pp.nums) + pp.Optional(comma + pp.Word(pp.nums), default=None)
    # parse pattern seperated by ";"
    pattern = pattern | pp.SkipTo(pp.Literal(";"))
    parsed = pp.delimitedList(pp.Group(pattern), ";")
    parsed = parsed.parseString(string)
    return parsed

print(parseKeyframes(testcase))
print(parseKeyframes(testcase2))
print(parseKeyframes(testcase3))
print(parseKeyframes(testcase4))
print(parseKeyframes(testcase5))
print(parseKeyframes(testcase6))

Output:

[['bounce', '5', '10'], ['shutter', '12', '14']]
[['bounce', '5', '10'], ['shutter', '12', '14'], ['low_effort', '2', '23']]
[['_lolw, a, 2'], ['effect', '6', None]]
[['bounce', '1', '10'], ['effect', '5', None]]
[[''], [''], [''], ['effect', '10', None], ['bounce, a, 1'], ['bounce', '3', '10']]
[['effect, b, a'], ['9, 10, 11'], ['max9, 10, 11'], ['here', '2', '3'], ['!b, 1, 2'], [''], ['']]

Answer 1

You have the right idea, just need to add a few additional terms, and suppress the errors.

1. Add '.suppress()' to the SkipTo term so that the skipped invalid text gets suppressed from the output.
2. To catch the case where a partial match is followed by invalid text (such as 'effect, 5, a' ), you must add a FollowedBy term so that a match is followed by either a ';' or the end of the string.
3. To handle invalid match at the end of the string, Skip to ";" or the end of the string

Since ";" | end_of_string ";" | end_of_string ";" | end_of_string occurs twice for very similar purposes, I created a next_delim term for that. This makes the lookaheads a little clearer. I also defined word and integer terms, which made pattern easier for me to follow.

I think this slightly modified version of your parser will give you your expected results:

comma = pp.Suppress(",")

# ';'-delimited list of  word , int [, int]
word = pp.Word(pp.alphas + "_")
integer = pp.Word(pp.nums)
pattern = (word
           + comma 
           + integer 
           + pp.Optional(comma + integer, default=None))


# parse pattern seperated by ";"
next_delim = ";" | pp.StringEnd()
skip_invalid = pp.SkipTo(next_delim).suppress()
pattern_or_skip = (pp.Group(pattern + pp.FollowedBy(next_delim))
                   | skip_invalid)

parser = pp.delimitedList(pattern_or_skip, delim=";")