CJK kinsoku line breaking: push-out chain · case 01

"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(x):
    text, L = x
    NO_START = '、。，．）」』！？ーぁぃぅぇぉっゃゅょ'
    NO_END = '（「『'
    HANG = '、。，．'
    lines = []
    pos = 0
    n = len(text)
    while pos < n:
        end = pos + L
        if end >= n:
            lines.append(text[pos:])
            break
        if text[end] in HANG and (end + 1 >= n or text[end + 1] not in NO_START):
            end += 1
        else:
            if end > pos + 1 and text[end] in NO_START:
                end -= 1
            while end > pos + 1 and text[end - 1] in NO_END:
                end -= 1
        lines.append(text[pos:end])
        pos = end
    return lines
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
fixtures = [[('regression: push-out chain', ['、本おう）こきかえ文日おいーゃ。（け', 7], ['、本おう）こき', 'かえ文日お', 'いーゃ。（け']), ('regression: push-out chain', ['文ーう文おっ日いあうあ）いゃ）』日『字', 5], ['文ーう文', 'おっ日いあ', 'うあ）', 'いゃ）』日', '『字']), ('partial-repair probe', ['字う『、え組こ、文、ーえ」、』語う。', 8], ['字う『、え組こ、', '文、ーえ」、』語', 'う。']), ('partial-repair probe', ['本』。。くゃ（か『「', 6], ['本』。。くゃ', '（か『「']), ('closing bracket pushes out', ['日本語文」字組版', 4], ['日本語', '文」字組', '版']), ('period followed by bracket cannot hang', ['日本語文。」字組', 4], ['日本語', '文。」字', '組']), ('control layout', ['あっけあ『（き版', 4], ['あっけあ', '『（き版']), ('control layout', ['語あ』文か（おく本」', 6], ['語あ』文か', '（おく本」'])], [('regression: push-out chain', ['ゃ『くかこけう」字（く、っ『語ゃ組', 6], ['ゃ『くかこけ', 'う」字', '（く、っ', '『語ゃ組']), ('regression: push-out chain', ['けけーうけゃ）（っ』』）ゃっ組）お', 4], ['けけーう', 'けゃ）', '（', 'っ', '』', '』）ゃっ', '組）お']), ('partial-repair probe', ['か語組ゃっうこうくきくゃ日', 5], ['か語組ゃっ', 'うこうくき', 'くゃ日']), ('partial-repair probe', ['お。）語うきき文』く「日本本組『ゃ文日', 4], ['お。）語', 'うきき', '文』く', '「日本本', '組『ゃ文', '日']), ('small kana prohibited at line start', ['日本語かっこいい', 4], ['日本語', 'かっこい', 'い']), ('period hangs past the measure', ['日本語文。字組版', 4], ['日本語文。', '字組版']), ('control layout', ['文字くい「「いおい字こく字っーー、かお本', 6], ['文字くい', '「「いおい字', 'こく字っーー、', 'かお本']), ('control layout', ['お語）日版』「か文え日日こ（ー（お', 7], ['お語）日版』', '「か文え日日こ', '（ー（お'])], [('regression: push-out chain', ['い。（）。けう組。き文文日文文えっ、』あ』字組本', 6], ['い。（）。け', 'う組。き文文', '日文文', 'えっ、』あ』', '字組本']), ('regression: push-out chain', ['う版あう。、っ』、）「字。あかか、『。いおきゃ', 4], ['う版あ', 'う', '。', '、', 'っ』、）', '「字。あ', 'かか、『。', 'いおきゃ']), ('partial-repair probe', ['こあ）ー、あ）っこ、「（あ「けけっ。えかきー文', 8], ['こあ）ー、あ）っ', 'こ、「（あ「け', 'けっ。えかきー文']), ('partial-repair probe', ['。け』かあ『組け』、字', 8], ['。け』かあ『組', 'け』、字']), ('period followed by bracket cannot hang', ['日本語文。」字組', 4], ['日本語', '文。」字', '組']), ('small kana prohibited at line start', ['日本語かっこいい', 4], ['日本語', 'かっこい', 'い']), ('control layout', ['お語）日版』「か文え日日こ（ー（お', 7], ['お語）日版』', '「か文え日日こ', '（ー（お']), ('control layout', ['日うああ版字お日。日お日', 4], ['日うああ', '版字お日。', '日お日'])], [('regression: push-out chain', ['」』こ）「。きゃ」語文うう『', 4], ['」』こ）', '「。', 'きゃ」語', '文うう『']), ('regression: push-out chain', ['語け本日ゃ」）「字、字字', 5], ['語け本', '日ゃ」）', '「字、字字']), ('regression: push-out chain', ['（ゃえい。く、ああ（おあ）「文）（か語ゃ。ーけき', 3], ['（ゃえ', 'い。く、', 'ああ', '（お', 'あ）', '「文）', '（か', '語', 'ゃ。ー', 'けき']), ('partial-repair probe', ['（）本）『組こ日本）版、（かけ本', 3], ['（）', '本）', '『組こ', '日本）', '版、', '（かけ', '本']), ('small kana prohibited at line start', ['日本語かっこいい', 4], ['日本語', 'かっこい', 'い']), ('closing bracket pushes out', ['日本語文」字組版', 4], ['日本語', '文」字組', '版']), ('control layout', ['ー』版）本うかえお日', 8], ['ー』版）本うかえ', 'お日']), ('control layout', ['こう）あ、く。日語', 8], ['こう）あ、く。日', '語'])], [('regression: push-out chain', ['（『あ（日字（字「い語。字『うく版文ーゃーあ', 7], ['（『あ（日字', '（字「い語。字', '『うく版', '文ーゃーあ']), ('regression: push-out chain', ['こっ（『おあゃお字」。。組うう組、こえけ）語', 5], ['こっ（『お', 'あゃお', '字」。。組', 'うう組、こ', 'えけ）語']), ('regression: push-out chain', ['、けゃ語え語字「か語く文（、字）っ。おか、ー', 7], ['、けゃ語え語字', '「か語く文（、', '字）っ。お', 'か、ー']), ('partial-repair probe', ['かあ語、本ゃっけ」', 8], ['かあ語、本ゃっ', 'け」']), ('period followed by bracket cannot hang', ['日本語文。」字組', 4], ['日本語', '文。」字', '組']), ('opening bracket at line end', ['日本語「文字」組版', 4], ['日本語', '「文字」', '組版']), ('control layout', ['け本こ日ゃき、こう」版（', 8], ['け本こ日ゃき、こ', 'う」版（']), ('control layout', ['日」』版、。字文うえこ日', 5], ['日」』版、。', '字文うえこ', '日'])]]
for label, args, expected in fixtures[N - 1]:
    check(label, solve(args), expected)
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)

SHA-256 / 965c68a6cf2820abb116838466ad9fc6cbe6478289ee44f611313cd4983f858e

"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(x):
    text, L = x
    NO_START = '、。，．）」』！？ーぁぃぅぇぉっゃゅょ'
    NO_END = '（「『'
    HANG = '、。，．'
    lines = []
    pos = 0
    n = len(text)
    while pos < n:
        end = pos + L
        if end >= n:
            lines.append(text[pos:])
            break
        if text[end] in HANG and (end + 1 >= n or text[end + 1] not in NO_START):
            end += 1
        else:
            while end > pos + 1 and text[end - 1] in NO_START:
                end -= 1
            while end > pos + 1 and text[end - 1] in NO_END:
                end -= 1
        lines.append(text[pos:end])
        pos = end
    return lines
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
fixtures = [[('regression: push-out chain', ['、本おう）こきかえ文日おいーゃ。（け', 7], ['、本おう）こき', 'かえ文日お', 'いーゃ。（け']), ('regression: push-out chain', ['文ーう文おっ日いあうあ）いゃ）』日『字', 5], ['文ーう文', 'おっ日いあ', 'うあ）', 'いゃ）』日', '『字']), ('partial-repair probe', ['字う『、え組こ、文、ーえ」、』語う。', 8], ['字う『、え組こ、', '文、ーえ」、』語', 'う。']), ('partial-repair probe', ['本』。。くゃ（か『「', 6], ['本』。。くゃ', '（か『「']), ('closing bracket pushes out', ['日本語文」字組版', 4], ['日本語', '文」字組', '版']), ('period followed by bracket cannot hang', ['日本語文。」字組', 4], ['日本語', '文。」字', '組']), ('control layout', ['あっけあ『（き版', 4], ['あっけあ', '『（き版']), ('control layout', ['語あ』文か（おく本」', 6], ['語あ』文か', '（おく本」'])], [('regression: push-out chain', ['ゃ『くかこけう」字（く、っ『語ゃ組', 6], ['ゃ『くかこけ', 'う」字', '（く、っ', '『語ゃ組']), ('regression: push-out chain', ['けけーうけゃ）（っ』』）ゃっ組）お', 4], ['けけーう', 'けゃ）', '（', 'っ', '』', '』）ゃっ', '組）お']), ('partial-repair probe', ['か語組ゃっうこうくきくゃ日', 5], ['か語組ゃっ', 'うこうくき', 'くゃ日']), ('partial-repair probe', ['お。）語うきき文』く「日本本組『ゃ文日', 4], ['お。）語', 'うきき', '文』く', '「日本本', '組『ゃ文', '日']), ('small kana prohibited at line start', ['日本語かっこいい', 4], ['日本語', 'かっこい', 'い']), ('period hangs past the measure', ['日本語文。字組版', 4], ['日本語文。', '字組版']), ('control layout', ['文字くい「「いおい字こく字っーー、かお本', 6], ['文字くい', '「「いおい字', 'こく字っーー、', 'かお本']), ('control layout', ['お語）日版』「か文え日日こ（ー（お', 7], ['お語）日版』', '「か文え日日こ', '（ー（お'])], [('regression: push-out chain', ['い。（）。けう組。き文文日文文えっ、』あ』字組本', 6], ['い。（）。け', 'う組。き文文', '日文文', 'えっ、』あ』', '字組本']), ('regression: push-out chain', ['う版あう。、っ』、）「字。あかか、『。いおきゃ', 4], ['う版あ', 'う', '。', '、', 'っ』、）', '「字。あ', 'かか、『。', 'いおきゃ']), ('partial-repair probe', ['こあ）ー、あ）っこ、「（あ「けけっ。えかきー文', 8], ['こあ）ー、あ）っ', 'こ、「（あ「け', 'けっ。えかきー文']), ('partial-repair probe', ['。け』かあ『組け』、字', 8], ['。け』かあ『組', 'け』、字']), ('period followed by bracket cannot hang', ['日本語文。」字組', 4], ['日本語', '文。」字', '組']), ('small kana prohibited at line start', ['日本語かっこいい', 4], ['日本語', 'かっこい', 'い']), ('control layout', ['お語）日版』「か文え日日こ（ー（お', 7], ['お語）日版』', '「か文え日日こ', '（ー（お']), ('control layout', ['日うああ版字お日。日お日', 4], ['日うああ', '版字お日。', '日お日'])], [('regression: push-out chain', ['」』こ）「。きゃ」語文うう『', 4], ['」』こ）', '「。', 'きゃ」語', '文うう『']), ('regression: push-out chain', ['語け本日ゃ」）「字、字字', 5], ['語け本', '日ゃ」）', '「字、字字']), ('regression: push-out chain', ['（ゃえい。く、ああ（おあ）「文）（か語ゃ。ーけき', 3], ['（ゃえ', 'い。く、', 'ああ', '（お', 'あ）', '「文）', '（か', '語', 'ゃ。ー', 'けき']), ('partial-repair probe', ['（）本）『組こ日本）版、（かけ本', 3], ['（）', '本）', '『組こ', '日本）', '版、', '（かけ', '本']), ('small kana prohibited at line start', ['日本語かっこいい', 4], ['日本語', 'かっこい', 'い']), ('closing bracket pushes out', ['日本語文」字組版', 4], ['日本語', '文」字組', '版']), ('control layout', ['ー』版）本うかえお日', 8], ['ー』版）本うかえ', 'お日']), ('control layout', ['こう）あ、く。日語', 8], ['こう）あ、く。日', '語'])], [('regression: push-out chain', ['（『あ（日字（字「い語。字『うく版文ーゃーあ', 7], ['（『あ（日字', '（字「い語。字', '『うく版', '文ーゃーあ']), ('regression: push-out chain', ['こっ（『おあゃお字」。。組うう組、こえけ）語', 5], ['こっ（『お', 'あゃお', '字」。。組', 'うう組、こ', 'えけ）語']), ('regression: push-out chain', ['、けゃ語え語字「か語く文（、字）っ。おか、ー', 7], ['、けゃ語え語字', '「か語く文（、', '字）っ。お', 'か、ー']), ('partial-repair probe', ['かあ語、本ゃっけ」', 8], ['かあ語、本ゃっ', 'け」']), ('period followed by bracket cannot hang', ['日本語文。」字組', 4], ['日本語', '文。」字', '組']), ('opening bracket at line end', ['日本語「文字」組版', 4], ['日本語', '「文字」', '組版']), ('control layout', ['け本こ日ゃき、こう」版（', 8], ['け本こ日ゃき、こ', 'う」版（']), ('control layout', ['日」』版、。字文うえこ日', 5], ['日」』版、。', '字文うえこ', '日'])]]
for label, args, expected in fixtures[N - 1]:
    check(label, solve(args), expected)
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)

SHA-256 / 310e181a42ed22e686ec71523b6adfccdfd589292d6a7fb75a9aede1b271361c

"""Failure Map reference implementation. Python standard library only."""
import json

N = 1
observations = []
def solve(x):
    text, L = x
    NO_START = '、。，．）」』！？ーぁぃぅぇぉっゃゅょ'
    NO_END = '（「『'
    HANG = '、。，．'
    lines = []
    pos = 0
    n = len(text)
    while pos < n:
        end = pos + L
        if end >= n:
            lines.append(text[pos:])
            break
        if text[end] in HANG and (end + 1 >= n or text[end + 1] not in NO_START):
            end += 1
        else:
            while end > pos + 1 and text[end] in NO_START:
                end -= 1
            while end > pos + 1 and text[end - 1] in NO_END:
                end -= 1
        lines.append(text[pos:end])
        pos = end
    return lines
def check(label, actual, expected):
    observations.append({"check": label, "actual": actual, "expected": expected, "passed": actual == expected})
fixtures = [[('regression: push-out chain', ['、本おう）こきかえ文日おいーゃ。（け', 7], ['、本おう）こき', 'かえ文日お', 'いーゃ。（け']), ('regression: push-out chain', ['文ーう文おっ日いあうあ）いゃ）』日『字', 5], ['文ーう文', 'おっ日いあ', 'うあ）', 'いゃ）』日', '『字']), ('partial-repair probe', ['字う『、え組こ、文、ーえ」、』語う。', 8], ['字う『、え組こ、', '文、ーえ」、』語', 'う。']), ('partial-repair probe', ['本』。。くゃ（か『「', 6], ['本』。。くゃ', '（か『「']), ('closing bracket pushes out', ['日本語文」字組版', 4], ['日本語', '文」字組', '版']), ('period followed by bracket cannot hang', ['日本語文。」字組', 4], ['日本語', '文。」字', '組']), ('control layout', ['あっけあ『（き版', 4], ['あっけあ', '『（き版']), ('control layout', ['語あ』文か（おく本」', 6], ['語あ』文か', '（おく本」'])], [('regression: push-out chain', ['ゃ『くかこけう」字（く、っ『語ゃ組', 6], ['ゃ『くかこけ', 'う」字', '（く、っ', '『語ゃ組']), ('regression: push-out chain', ['けけーうけゃ）（っ』』）ゃっ組）お', 4], ['けけーう', 'けゃ）', '（', 'っ', '』', '』）ゃっ', '組）お']), ('partial-repair probe', ['か語組ゃっうこうくきくゃ日', 5], ['か語組ゃっ', 'うこうくき', 'くゃ日']), ('partial-repair probe', ['お。）語うきき文』く「日本本組『ゃ文日', 4], ['お。）語', 'うきき', '文』く', '「日本本', '組『ゃ文', '日']), ('small kana prohibited at line start', ['日本語かっこいい', 4], ['日本語', 'かっこい', 'い']), ('period hangs past the measure', ['日本語文。字組版', 4], ['日本語文。', '字組版']), ('control layout', ['文字くい「「いおい字こく字っーー、かお本', 6], ['文字くい', '「「いおい字', 'こく字っーー、', 'かお本']), ('control layout', ['お語）日版』「か文え日日こ（ー（お', 7], ['お語）日版』', '「か文え日日こ', '（ー（お'])], [('regression: push-out chain', ['い。（）。けう組。き文文日文文えっ、』あ』字組本', 6], ['い。（）。け', 'う組。き文文', '日文文', 'えっ、』あ』', '字組本']), ('regression: push-out chain', ['う版あう。、っ』、）「字。あかか、『。いおきゃ', 4], ['う版あ', 'う', '。', '、', 'っ』、）', '「字。あ', 'かか、『。', 'いおきゃ']), ('partial-repair probe', ['こあ）ー、あ）っこ、「（あ「けけっ。えかきー文', 8], ['こあ）ー、あ）っ', 'こ、「（あ「け', 'けっ。えかきー文']), ('partial-repair probe', ['。け』かあ『組け』、字', 8], ['。け』かあ『組', 'け』、字']), ('period followed by bracket cannot hang', ['日本語文。」字組', 4], ['日本語', '文。」字', '組']), ('small kana prohibited at line start', ['日本語かっこいい', 4], ['日本語', 'かっこい', 'い']), ('control layout', ['お語）日版』「か文え日日こ（ー（お', 7], ['お語）日版』', '「か文え日日こ', '（ー（お']), ('control layout', ['日うああ版字お日。日お日', 4], ['日うああ', '版字お日。', '日お日'])], [('regression: push-out chain', ['」』こ）「。きゃ」語文うう『', 4], ['」』こ）', '「。', 'きゃ」語', '文うう『']), ('regression: push-out chain', ['語け本日ゃ」）「字、字字', 5], ['語け本', '日ゃ」）', '「字、字字']), ('regression: push-out chain', ['（ゃえい。く、ああ（おあ）「文）（か語ゃ。ーけき', 3], ['（ゃえ', 'い。く、', 'ああ', '（お', 'あ）', '「文）', '（か', '語', 'ゃ。ー', 'けき']), ('partial-repair probe', ['（）本）『組こ日本）版、（かけ本', 3], ['（）', '本）', '『組こ', '日本）', '版、', '（かけ', '本']), ('small kana prohibited at line start', ['日本語かっこいい', 4], ['日本語', 'かっこい', 'い']), ('closing bracket pushes out', ['日本語文」字組版', 4], ['日本語', '文」字組', '版']), ('control layout', ['ー』版）本うかえお日', 8], ['ー』版）本うかえ', 'お日']), ('control layout', ['こう）あ、く。日語', 8], ['こう）あ、く。日', '語'])], [('regression: push-out chain', ['（『あ（日字（字「い語。字『うく版文ーゃーあ', 7], ['（『あ（日字', '（字「い語。字', '『うく版', '文ーゃーあ']), ('regression: push-out chain', ['こっ（『おあゃお字」。。組うう組、こえけ）語', 5], ['こっ（『お', 'あゃお', '字」。。組', 'うう組、こ', 'えけ）語']), ('regression: push-out chain', ['、けゃ語え語字「か語く文（、字）っ。おか、ー', 7], ['、けゃ語え語字', '「か語く文（、', '字）っ。お', 'か、ー']), ('partial-repair probe', ['かあ語、本ゃっけ」', 8], ['かあ語、本ゃっ', 'け」']), ('period followed by bracket cannot hang', ['日本語文。」字組', 4], ['日本語', '文。」字', '組']), ('opening bracket at line end', ['日本語「文字」組版', 4], ['日本語', '「文字」', '組版']), ('control layout', ['け本こ日ゃき、こう」版（', 8], ['け本こ日ゃき、こ', 'う」版（']), ('control layout', ['日」』版、。字文うえこ日', 5], ['日」』版、。', '字文うえこ', '日'])]]
for label, args, expected in fixtures[N - 1]:
    check(label, solve(args), expected)
print(json.dumps({"observations": observations, "passed": all(x["passed"] for x in observations)}, ensure_ascii=False))
raise SystemExit(0 if all(x["passed"] for x in observations) else 1)

SHA-256 / f3bf45f9ac7273e966603d40fb41a530770df9f391615b89847442bca6ba2ac8