Grammalecte  Check-in [1d0bea98d6]

"""
Grammalecte: compile rules
Create a Direct Acyclic Rule Graphs (DARGs)
"""

import re
import traceback
import json

import darg
import compile_rules_js_convert as jsconv


dACTIONS = {}
dFUNCTIONS = {}
dFUNCNAME = {}


def createFunction (sType, sActionId, sCode, bStartWithEqual=False):
def createFunction (sType, sCode, bStartWithEqual=False):
    "create a function (stored in <dFUNCTIONS>) and return function name"
    sCode = prepareFunction(sCode)
    if sType not in dFUNCNAME:
        dFUNCNAME[sType] = {}
    if sCode not in dFUNCNAME[sType]:
        dFUNCNAME[sType][sCode] = len(dFUNCNAME[sType])+1
    sFuncName = "_g_" + sType + "_" + str(dFUNCNAME[sType][sCode])
    dFUNCTIONS[sFuncName] = sCode
    return sFuncName  if not bStartWithEqual  else "="+sFuncName


def storeAction (sActionId, aAction):
    "store <aAction> in <dACTIONS> avoiding duplicates"
    nVar = 0
    while True:
        sActionName = sActionId + "_" + str(nVar)
        if sActionName not in dACTIONS:
            dACTIONS[sActionName] = aAction
            return sActionName
        elif aAction == dACTIONS[sActionName]:
        if aAction == dACTIONS[sActionName]:
            return sActionName
        nVar += 1


def prepareFunction (sCode):
    "convert simple rule syntax to a string of Python code"
    if sCode[0:1] == "=":

    sCode = re.sub(r"<1", 'lToken[nTokenOffset]["sValue"]', sCode)
    return sCode


def genTokenLines (sTokenLine, dDef):
    "tokenize a string and return a list of lines of tokens"
    lToken = sTokenLine.split()
    lTokenLines = None
    lTokenLines = []
    for sToken in lToken:
        # replace merger characters by spaces
        if "␣" in sToken:
            sToken = sToken.replace("␣", " ")
        # optional token?
        bNullPossible = sToken.startswith("?") and sToken.endswith("¿")
        if bNullPossible:

        print(" # Error. No action found at: ", sActionId)
        return None

    # Condition
    sCondition = sAction[:m.start()].strip()
    if sCondition:
        sCondition = changeReferenceToken(sCondition, dPos)
        sCondition = createFunction("cond", sActionId, sCondition)
        sCondition = createFunction("cond", sCondition)
    else:
        sCondition = ""

    # Case sensitivity
    bCaseSensitivity = False if m.group("casing") == ":" else True
    bCaseSensitivity = not bool(m.group("casing"))

    # Action
    cAction = m.group("action")
    sAction = sAction[m.end():].strip()
    sAction = changeReferenceToken(sAction, dPos)
    # target
    cStartLimit = "<"

        if not m.group("end"):
            iEndAction = iStartAction
        else:
            if m.group("end").startswith(":."):
                cEndLimit = "<"
            iEndAction = int(m.group("end").lstrip(":."))
    if dPos and m.group("start"):
        try:
            iStartAction = dPos.get(iStartAction, iStartAction)
            if iEndAction:
                iEndAction = dPos.get(iEndAction, iEndAction)
        iStartAction = dPos.get(iStartAction, iStartAction)
        if iEndAction:
            iEndAction = dPos.get(iEndAction, iEndAction)
        except:
            print("# Error. Wrong groups in: " + sActionId)
            print("  iStartAction:", iStartAction, "iEndAction:", iEndAction)
            print(" ", dPos)
    if iStartAction < 0:
        iStartAction += 1
    if iEndAction < 0:
        iEndAction += 1

    if cAction == "-":
        ## error

            sURL = ""
            mURL = re.search("[|] *(https?://.*)", sMsg)
            if mURL:
                sURL = mURL.group(1).strip()
                sMsg = sMsg[:mURL.start(0)].strip()
            checkTokenNumbers(sMsg, sActionId, nToken)
            if sMsg[0:1] == "=":
                sMsg = createFunction("msg", sActionId, sMsg, True)
                sMsg = createFunction("msg", sMsg, True)
            else:
                checkIfThereIsCode(sMsg, sActionId)

    # checking consistancy
    checkTokenNumbers(sAction, sActionId, nToken)

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

    if not sAction and cAction != "%":
        print("# Error in action at line " + sActionId + ":  This action is empty.")

    if sAction[0:1] != "=" and cAction != "=":
        checkIfThereIsCode(sAction, sActionId)

    if cAction == "-":
        ## error detected --> suggestion
        if sAction[0:1] == "=":
            sAction = createFunction("sugg", sActionId, sAction, True)
            sAction = createFunction("sugg", sAction, True)
        elif sAction.startswith('"') and sAction.endswith('"'):
            sAction = sAction[1:-1]
        if not sMsg:
            print("# Error in action at line " + sActionId + ":  The message is empty.")
        return [sOption, sCondition, cAction, sAction, iStartAction, iEndAction, cStartLimit, cEndLimit, bCaseSensitivity, nPriority, sMsg, sURL]
    elif cAction == "~":
    if cAction == "~":
        ## text processor
        if sAction[0:1] == "=":
            sAction = createFunction("tp", sActionId, sAction, True)
            sAction = createFunction("tp", sAction, True)
        elif sAction.startswith('"') and sAction.endswith('"'):
            sAction = sAction[1:-1]
        return [sOption, sCondition, cAction, sAction, iStartAction, iEndAction, bCaseSensitivity]
    elif cAction == "%" or cAction == "/":
    if cAction in "%/":
        ## tags
        return [sOption, sCondition, cAction, sAction, iStartAction, iEndAction]
    elif cAction == "=":
    if cAction == "=":
        ## disambiguator
        if "define(" in sAction and not re.search(r"define\(\\-?\d+ *, *\[.*\] *\)", sAction):
            print("# Error in action at line " + sActionId + ": second argument for <define> must be a list of strings")
        sAction = createFunction("da", sActionId, sAction)
        sAction = createFunction("da", sAction)
        return [sOption, sCondition, cAction, sAction]
    else:
        print(" # Unknown action.", sActionId)
        return None
    print(" # Unknown action.", sActionId)
    return None


def make (lRule, dDef, sLang, dOptPriority):
    "compile rules, returns a dictionary of values"
    # for clarity purpose, don’t create any file here

    # removing comments, zeroing empty lines, creating definitions, storing tests, merging rule lines

    # processing rules
    print("  preparing rules...")
    for sGraphName, lRuleLine in dAllGraph.items():
        print("{:>8,} rules in {:<24} ".format(len(lRuleLine), "<"+sGraphName+">"), end="")
        lPreparedRule = []
        for i, sRuleGroup, sTokenLine, iActionBlock, sActions, nPriority in lRuleLine:
            for lRule in createRule(i, sRuleGroup, sTokenLine, iActionBlock, sActions, nPriority, dOptPriority, dDef):
                lPreparedRule.append(lRule)
            for aRule in createRule(i, sRuleGroup, sTokenLine, iActionBlock, sActions, nPriority, dOptPriority, dDef):
                lPreparedRule.append(aRule)
        # Graph creation
        oDARG = darg.DARG(lPreparedRule, sLang)
        dAllGraph[sGraphName] = oDARG.createGraph()
        # Debugging
        if False:
            print("\nRULES:")
            for e in lPreparedRule: