Grammalecte  Diff

        return re.compile(sRegex).groups
    except re.error:
        traceback.print_exc()
        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

    sLineId = f"#{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()

    for _ in re.finditer(r"\(\?:[^)]*\([\[\w -]", sRegex):
        print(f"# Warning: groups inside non grouping parenthesis in regex at line {sLineId}")

    #### PARSE ACTIONS
    lActions = []
    nAction = 1
    for sAction in s.split(" <<- "):
        t = createAction(sRuleId + "_" + str(nAction), sAction, nGroup)
        nAction += 1
        if t:
            lActions.append(t)
    if not lActions:
        return None

    return [sOption, sRegex, bCaseInsensitive, sLineId, sRuleId, nPriority, lActions, tGroups]


def checkReferenceNumbers (sText, sActionId, nToken):
    "check if token references in <sText> greater than <nToken> (debugging)"
    for x in re.finditer(r"\\(\d+)", sText):
        if int(x.group(1)) > nToken:
            print(f"# Error in token index at line {sActionId} ({nToken} tokens only)")
            print(sText)


def checkIfThereIsCode (sText, sActionId):
    "check if there is code in <sText> (debugging)"
    if re.search("[.]\\w+[(]|sugg\\w+[(]|\\([0-9]|\\[[0-9]", sText):
        print(f"# Warning at line {sActionId}:  This message looks like code. Line should probably begin with =")
        print(sText)


def createAction (sIdAction, sAction, nGroup):
    "returns an action to perform as a tuple (condition, action type, action[, iGroup [, message, URL ]])"
    m = re.search(r"([-~=>])(\d*|)>>", sAction)
    if not m:
        print(f"# No action at line {sIdAction}")
        return None

    #### CONDITION
    sCondition = sAction[:m.start()].strip()
    if sCondition:
        sCondition = prepareFunction(sCondition)
        lFUNCTIONS.append(("_c_"+sIdAction, sCondition))
        checkReferenceNumbers(sCondition, sIdAction, nGroup)
        if ".match" in sCondition:
            print("# Error. JS compatibility. Don't use .match() in condition, use .search()")
        sCondition = "_c_"+sIdAction
    else:
        sCondition = None

    #### iGroup / positioning
    iGroup = int(m.group(2)) if m.group(2) else 0
    if iGroup > nGroup:
        print(f"# Selected group > group number in regex at line {sIdAction}")

    #### ACTION
    sAction = sAction[m.end():].strip()
    cAction = m.group(1)
    if cAction == "-":
        ## error
        iMsg = sAction.find(" && ")
        if iMsg == -1:
            sMsg = "# Error. Error message not found."
            sURL = ""
            print(f"# No message. Action id: {sIdAction}")
        else:
            sMsg = sAction[iMsg+4:].strip()
            sAction = sAction[:iMsg].strip()
            sURL = ""
            mURL = re.search("[|] *(https?://.*)", sMsg)
            if mURL:
                sURL = mURL.group(1).strip()
                sMsg = sMsg[:mURL.start(0)].strip()
            checkReferenceNumbers(sMsg, sIdAction, nGroup)
            if sMsg[0:1] == "=":
                sMsg = prepareFunction(sMsg[1:])
                lFUNCTIONS.append(("_m_"+sIdAction, sMsg))
                sMsg = "=_m_"+sIdAction
            else:
                checkIfThereIsCode(sMsg, sIdAction)

    checkReferenceNumbers(sAction, sIdAction, nGroup)
    if sAction[0:1] == "=" or cAction == "=":
        sAction = prepareFunction(sAction)
        sAction = sAction.replace("m.group(i[4])", "m.group("+str(iGroup)+")")
    else:
        checkIfThereIsCode(sAction, sIdAction)

    if cAction == ">":
        ## no action, break loop if condition is False
        return [sCondition, cAction, ""]

    if not sAction:
        print(f"# Error in action at line {sIdAction}:  This action is empty.")
        return None

    if cAction == "-":
        ## error detected --> suggestion
        if sAction[0:1] == "=":
            lFUNCTIONS.append(("_s_"+sIdAction, sAction[1:]))
            sAction = "=_s_"+sIdAction
        elif sAction.startswith('"') and sAction.endswith('"'):
            sAction = sAction[1:-1]
        if not sMsg:
            print(f"# Error in action at line {sIdAction}:  the message is empty.")
        return [sCondition, cAction, sAction, iGroup, sMsg, sURL]
    if cAction == "~":
        ## text processor
        if sAction[0:1] == "=":
            lFUNCTIONS.append(("_p_"+sIdAction, sAction[1:]))
            sAction = "=_p_"+sIdAction
        elif sAction.startswith('"') and sAction.endswith('"'):
            sAction = sAction[1:-1]
        return [sCondition, cAction, sAction, iGroup]
    if cAction == "=":
        ## disambiguator
        if sAction[0:1] == "=":
            sAction = sAction[1:]
        if "define" in sAction and not re.search(r"define\(dTokenPos, *m\.start.*, \[.*\] *\)", sAction):
            print(f"# Error in action at line {sIdAction}: second argument for define must be a list of strings")
            print(sAction)
        lFUNCTIONS.append(("_d_"+sIdAction, sAction))
        sAction = "_d_"+sIdAction
        return [sCondition, cAction, sAction]
    print(f"# Unknown action at line {sIdAction}")
    return None


def _calcRulesStats (lRules):
    "count rules and actions"
    d = {'=':0, '~': 0, '-': 0, '>': 0}
    for aRule in lRules: