Grammalecte  Diff

dDEFINITIONS = {}
dDECLENSIONS = {}
lFUNCTIONS = []

aRULESET = set()     # set of rule-ids to check if there is several rules with the same id
nRULEWITHOUTNAME = 0

dJSREGEXES = {}

sWORDLIMITLEFT  = r"(?<![\w.,–-])"   # r"(?<![-.,—])\b"  seems slower
sWORDLIMITRIGHT = r"(?![\w–-])"      # r"\b(?!-—)"       seems slower

    return dColorType


def prepareFunction (s):
    "convert simple rule syntax to a string of Python code"
    s = s.replace("__also__", "bCondMemo")
    s = s.replace("__else__", "not bCondMemo")
    s = s.replace("sContext", "_sAppContext")
    s = re.sub(r"isStart *\(\)", 'before("^ *$|, *$")', s)
    s = re.sub(r"isRealStart *\(\)", 'before("^ *$")', s)
    s = re.sub(r"isStart0 *\(\)", 'before0("^ *$|, *$")', s)
    s = re.sub(r"isRealStart0 *\(\)", 'before0("^ *$")', s)
    s = re.sub(r"isEnd *\(\)", 'after("^ *$|^,")', s)
    s = re.sub(r"isRealEnd *\(\)", 'after("^ *$")', s)
    s = re.sub(r"isEnd0 *\(\)", 'after0("^ *$|^,")', s)

        print(sRegex)
    return 0


def createRule (s, nIdLine, sLang, bParagraph, dOptPriority):
    "returns rule as list [option name, regex, bCaseInsensitive, identifier, list of actions]"
    global dJSREGEXES
    global nRULEWITHOUTNAME

    sLineId = "#" + str(nIdLine) + ("p" if bParagraph else "s")
    sRuleId = sLineId

    #### GRAPH CALL
    if s.startswith("@@@@"):
        if bParagraph:
            print("Error. Graph call can be made only after the first pass (sentence by sentence)")
            exit()
        return ["@@@@", s[4:], sLineId]

    #### OPTIONS
    sOption = False         # False or [a-z0-9]+ name
    nPriority = 4           # Default is 4, value must be between 0 and 9
    tGroups = None          # code for groups positioning (only useful for JavaScript)
    cCaseMode = 'i'         # i: case insensitive,  s: case sensitive,  u: uppercasing allowed
    cWordLimitLeft = '['    # [: word limit, <: no specific limit
    cWordLimitRight = ']'   # ]: word limit, >: no specific limit
    m = re.match("^__(?P<borders_and_case>[\\[<]\\w[\\]>])(?P<option>/[a-zA-Z0-9]+|)(?P<ruleid>\\(\\w+\\)|)(?P<priority>![0-9]|)__ *", s)
    m = re.match("^__(?P<borders_and_case>[\\[<]\\w[\\]>])(?P<option>/[a-zA-Z0-9]+|)(?P<ruleid>\\(\\w+\\))(?P<priority>![0-9]|)__ *", s)
    if m:
        cWordLimitLeft = m.group('borders_and_case')[0]
        cCaseMode = m.group('borders_and_case')[1]
        cWordLimitRight = m.group('borders_and_case')[2]
        sOption = m.group('option')[1:]  if m.group('option')  else False
        if m.group('ruleid'):
            sRuleId =  m.group('ruleid')[1:-1]
            if sRuleId in aRULESET:
                print("# Error. Several rules have the same id: " + sRuleId)
                exit()
            aRULESET.add(sRuleId)
        sRuleId =  m.group('ruleid')[1:-1]
        if sRuleId in aRULESET:
            print("# Error. Several rules have the same id: " + sRuleId)
            exit()
        aRULESET.add(sRuleId)
        else:
            nRULEWITHOUTNAME += 1
        nPriority = dOptPriority.get(sOption, 4)
        if m.group('priority'):
            nPriority = int(m.group('priority')[1:])
        s = s[m.end(0):]
    else:
        print("# Warning. No option defined at line: " + sLineId)
        print("# Warning. Rule wrongly shaped at line: " + sLineId)
        exit()

    #### REGEX TRIGGER
    i = s.find(" <<-")
    if i == -1:
        print("# Error: no condition at line " + sLineId)
        return None
    sRegex = s[:i].strip()

            pass
        elif bGraph:
            lGraphRule.append([i, sLine])
        # Regex rules
        elif re.match("[  \t]*$", sLine):
            # empty line
            pass
        elif sLine.startswith(("    ", "\t")):
        elif sLine.startswith("    "):
            # rule (continuation)
            lRuleLine[-1][1] += " " + sLine.strip()
        else:
            # new rule
            lRuleLine.append([i, sLine.strip()])

    # generating options files

        # JavaScript
        sJSCallables += "    {}: function ({})".format(sFuncName, sParams) + " {\n"
        sJSCallables += "        return " + jsconv.py2js(sReturn) + ";\n"
        sJSCallables += "    },\n"

    displayStats(lParagraphRules, lSentenceRules)

    print("Unnamed rules: " + str(nRULEWITHOUTNAME))

    dVars = {
        "fBuildTime": fBuildTime,
        "callables": sPyCallables,
        "callablesJS": sJSCallables,
        "gctests": sGCTests,
        "gctestsJS": sGCTestsJS,
        "paragraph_rules": mergeRulesByOption(lParagraphRules),