--- cdli/cdli_files.py 2006/06/20 13:13:55 1.30 +++ cdli/cdli_files.py 2007/08/31 14:22:52 1.79 @@ -1,11 +1,13 @@ """CDLI extensions of the filearchive""" -from Products.versionedFile.versionedFile import * +from Products.versionedFile.extVersionedFile import * from Products.ZCatalog.CatalogPathAwareness import CatalogAware from tempfile import mkstemp,mkdtemp import os.path import os from types import * import urlparse +import urllib +import cgi from OFS.OrderedFolder import OrderedFolder from OFS.SimpleItem import SimpleItem import time @@ -13,14 +15,166 @@ from OFS.Folder import manage_addFolder import re from AccessControl import ClassSecurityInfo from Acquisition import Implicit +from Globals import Persistent from threading import Thread from ZPublisher.HTTPRequest import HTTPRequest from ZPublisher.HTTPResponse import HTTPResponse from ZPublisher.BaseRequest import RequestContainer import threading +from BTrees.OOBTree import OOBTree +import logging +import transaction +import copy +import codecs +import sys + +def unicodify(s): + """decode str (utf-8 or latin-1 representation) into unicode object""" + if not s: + return u"" + if isinstance(s, str): + try: + return s.decode('utf-8') + except: + return s.decode('latin-1') + else: + return s + +def utf8ify(s): + """encode unicode object or string into byte string in utf-8 representation. + assumes string objects to be utf-8""" + if not s: + return "" + if isinstance(s, str): + return s + else: + return s.encode('utf-8') + + +def generateXMLReturn(hash): + """erzeugt das xml file als returnwert fuer uploadATFRPC""" + + ret="" + + ret+="" + for error in hash['errors']: + ret+="""%s"""%error + + ret+="" + + ret+="" + for changed in hash['changed']: + ret+="""%s"""%changed + ret+="" + + ret+="" + for new in hash['newPs']: + ret+=""""""%new + ret+="" + + ret+="" + return ret + + + + + + + + + +def unique(s): + """Return a list of the elements in s, but without duplicates. + + For example, unique([1,2,3,1,2,3]) is some permutation of [1,2,3], + unique("abcabc") some permutation of ["a", "b", "c"], and + unique(([1, 2], [2, 3], [1, 2])) some permutation of + [[2, 3], [1, 2]]. + + For best speed, all sequence elements should be hashable. Then + unique() will usually work in linear time. + + If not possible, the sequence elements should enjoy a total + ordering, and if list(s).sort() doesn't raise TypeError it's + assumed that they do enjoy a total ordering. Then unique() will + usually work in O(N*log2(N)) time. + If that's not possible either, the sequence elements must support + equality-testing. Then unique() will usually work in quadratic + time. + (from the python cookbook) + """ + + n = len(s) + if n == 0: + return [] + + # Try using a dict first, as that's the fastest and will usually + # work. If it doesn't work, it will usually fail quickly, so it + # usually doesn't cost much to *try* it. It requires that all the + # sequence elements be hashable, and support equality comparison. + u = {} + try: + for x in s: + u[x] = 1 + except TypeError: + del u # move on to the next method + else: + return u.keys() + + # We can't hash all the elements. Second fastest is to sort, + # which brings the equal elements together; then duplicates are + # easy to weed out in a single pass. + # NOTE: Python's list.sort() was designed to be efficient in the + # presence of many duplicate elements. This isn't true of all + # sort functions in all languages or libraries, so this approach + # is more effective in Python than it may be elsewhere. + try: + t = list(s) + t.sort() + except TypeError: + del t # move on to the next method + else: + assert n > 0 + last = t[0] + lasti = i = 1 + while i < n: + if t[i] != last: + t[lasti] = last = t[i] + lasti += 1 + i += 1 + return t[:lasti] + + # Brute force is all that's left. + u = [] + for x in s: + if x not in u: + u.append(x) + return u + +class BasketContent(SimpleItem): + """classe fuer den Inhalt eines Baskets""" + + def __init__(self,content=[]): + """content""" + self.contentList=content[0:] + + def getContent(self): + """get content""" + + return self.contentList + + def setContent(self,content): + self.contentList=content[0:] + + def numberOfItems(self): + """number""" + + return len(self.getContent()) + + class uploadATFfinallyThread(Thread): """class for adding uploaded filed (temporarily stored in the staging area at /tmp""" @@ -40,6 +194,7 @@ class uploadATFfinallyThread(Thread): self.SESSION=SESSION self.username=username self.serverport=serverport + def __call__(self): """call of the thread (equals run)""" @@ -73,10 +228,11 @@ class uploadATFfinallyThread(Thread): #add the files self.uploadATFfinallyThread(ctx,self.procedure,comment=self.comment,basketname=self.basketname,unlock=self.unlock,SESSION=self.SESSION,username=self.username) #commit the transactions - get_transaction().commit() + transaction.get().commit() conn.close() #set flag for end of this method self.end=True + logging.info("ended") return True def __del__(self): @@ -97,15 +253,16 @@ class uploadATFfinallyThread(Thread): self.result+="

Start processing

" #shall I only upload the changed files? + logging.info("uploadATFfinally procedure: %s"%procedure) if procedure=="uploadchanged": - - uploadFns=SESSION.get('changed',[])+SESSION.get('newPs',[]) + changed=[x[0] for x in SESSION.get('changed',[])] + uploadFns=changed+SESSION.get('newPs',[]) #or all elif procedure=="uploadAll": uploadFns=[] for x in os.listdir(SESSION['tmpdir']): - if not x in SESSION['errors']: + if not x in SESSION['lockerrors']: uploadFns.append(x) #or maybe nothing @@ -115,46 +272,64 @@ class uploadATFfinallyThread(Thread): uploadFns=[] #do first the changed files + i=0 for fn in uploadFns: + i+=1 founds=ctx2.CDLICatalog.search({'title':fn}) if len(founds)>0: SESSION['author']=str(username) - self.result+="

Changing : %s"%fn - founds[0].getObject().manage_addCDLIFileObject('',comment,SESSION['author'],file=file(os.path.join(SESSION['tmpdir'],fn))) - + self.result="

Changing : %s"%fn+self.result + founds[0].getObject().manage_addCDLIFileObject('',comment,SESSION['author'],file=os.path.join(SESSION['tmpdir'],fn),from_tmp=True) + if i==200: + i=0 + transaction.get().commit() + logging.info("changing: do commit") + transaction.get().commit() + logging.info("changing: last commit") + #now add the new files newPs=SESSION['newPs'] if len(newPs)>0: tmpDir=SESSION['tmpdir'] - self.result+="

Adding files

" + logging.info("adding start") + self.result="

Adding files

"+self.result #TODO: make this configurable, at the moment base folder for the files has to be cdli_main ctx2.importFiles(comment=comment,author=str(username) ,folderName=tmpDir, files=newPs,ext=self) - + logging.info("adding finished") #unlock locked files? if unlock: - self.result+="

Unlock files

" + logging.info("unlocking start") + self.result="

Unlock files

"+self.result unlockFns=[] for x in os.listdir(SESSION['tmpdir']): if not x in SESSION['errors']: unlockFns.append(x) - + logging.info("unlocking have now what to unlock") + for fn in unlockFns: + #logging.info("will unlock: %s"%fn) founds=ctx2.CDLICatalog.search({'title':fn}) + #logging.info("found it: %s"%repr(founds)) if len(founds)>0: + #logging.info("unlock: %s"%founds[0].getObject().getId()) SESSION['author']=str(username) founds[0].getObject().lockedBy="" + logging.info("unlocking done") - #if a basketname is give, add files to the basket + #if a basketname is given, add files to the basket if not (basketname ==''): - self.result+="

Add basket

" + logging.info("add to basket %s"%basketname) + self.result="

Add to basket

"+self.result basketId=ctx2.basketContainer.getBasketIdfromName(basketname) if not basketId: # create new basket + logging.info("create basket %s"%basketname) + self.result="

Create a new basket

"+self.result ob=ctx2.basketContainer.addBasket(basketname) basketId=ob.getId() basket=getattr(ctx2.basketContainer,str(basketId)) @@ -165,9 +340,17 @@ class uploadATFfinallyThread(Thread): RESPONSE.redirect(self.aq_parent.absolute_url()) - + logging.info("uploadfinally done") return True + +class tmpStore(SimpleItem): + """simple item""" + meta_type="cdli_upload" + def __init__(self,id): + """init tmp""" + self.id=id + class uploadATFThread(Thread): """class for checking the files befor uploading""" @@ -180,13 +363,14 @@ class uploadATFThread(Thread): Thread.__init__(self) - def set(self,upload,basketId,username,serverport="8080"): + def set(self,upload,basketId,username,idTmp,serverport="8080"): """set start values for the thread""" self.result="" self.upload=upload self.basketId=basketId self.username=username self.serverport=serverport + self.idTmp=idTmp def __call__(self): """call method """ @@ -205,7 +389,7 @@ class uploadATFThread(Thread): return app.__of__(RequestContainer(REQUEST = req)) def run(self): - + idTmp=self.idTmp self.result="" #find context within ZODB from Zope import DB @@ -213,26 +397,36 @@ class uploadATFThread(Thread): root = conn.root() app = root['Application'] ctx = self.getContext(app,serverport=self.serverport) - self.uploadATFThread(ctx,self.upload,self.basketId) + logging.info("run intern") + try: + logging.info("created: %s"%idTmp) + ctx.temp_folder._setObject(idTmp,tmpStore(idTmp)) + except: + logging.error("thread upload: %s %s"%sys.exc_info()[0:2]) + + logging.info("call thread intern") + self.uploadATFThread(ctx,self.upload,idTmp,self.basketId) #ctx.cdliRoot.cdli_main.tmpStore2[self.getName()[0:]]=self.returnValue - - get_transaction().commit() - while self.continueVar: - pass + transaction.get().commit() + conn.close() + return getattr(ctx.temp_folder,idTmp) def getResult(self): """method for accessing result""" return self.result - def uploadATFThread(self,ctx,upload,basketId=0): + def uploadATFThread(self,ctx,upload,idTmp,basketId=0): """upload an atf file""" #TODO: add comments #TODO: finish uploadATF + + stObj=getattr(ctx.temp_folder,idTmp) + logging.info("start, upload thread") self.result="

I got your file, start now to split it into single atf-files!

" #make sure that id is a string and not an integer @@ -247,6 +441,8 @@ class uploadATFThread(Thread): changed=[] # changed files errors=[] # files with errors + lockerrors=[] # files with errors + newPs=[] # new p filed psNotInCatalog=[] # files not in the catalog @@ -261,7 +457,7 @@ class uploadATFThread(Thread): if basketObj: basketId=basketObj.getId() - #if there is no active baske and no basketid given, id is empty, else get besketname and length + #if there is no active basket and no basketid given, id is empty, else get besketname and length if basketId == '0': basketNameFromId="" basketLen=0 @@ -269,14 +465,14 @@ class uploadATFThread(Thread): basketNameFromId=getattr(ctx2.basketContainer,basketId).title basketLen=getattr(ctx2.basketContainer,basketId).getLastVersion().numberOfItems() - + logging.info("got the file, upload thread") self.result+="""

I got the files

< p>I am computing the differences to the exisiting files

""" #start to check the files for fn in os.listdir(dir): - self.result+="

process:%s

"%fn + self.result="

process:%s

"%fn+self.result # check if file is in the catalog #TODO: checkCatalog is not implemented yet @@ -287,272 +483,55 @@ class uploadATFThread(Thread): founds=ctx2.CDLICatalog.search({'title':fn}) #if not than add filename to the list of newfiles - if len(founds)==0: - newPs.append(fn) - #if p file alread at the server - for found in founds: - #analyse the differences to the actual file - obj=found.getObject() - - if (not obj.lockedBy=='') and (not obj.lockedBy==self.username): - errors.append(obj) - else: - data=file(os.path.join(dir,fn)).read() - diffs=obj.diff(data) - if diffs[0]>0: - changed.append((obj,diffs)) - #hochladen - - #ready, set the returnValues - self.result+="

Done

" - - self.returnValue={} - self.returnValue['changed']=changed - self.returnValue['errors']=errors - self.returnValue['newPs']=newPs - self.returnValue['tmpdir']=dir - self.returnValue['basketLen']=basketLen - self.returnValue['numberOfFiles']=numberOfFiles - self.returnValue['basketNameFromId']=basketNameFromId - self.returnValue['basketNameFromFile']=basketNameFromFile - self.returnValue['basketId']=basketId - self.returnValue['dir']=dir - - #ctx2.cdli_main.setTemp('v_uploadATF_returnValue',True) - - -class Basket_old(Folder): - """shopping basket - alte fassung """ - - meta_type="Basket" - _v_stack={} - - def getObjUrl(self,objId): - """getUrl""" - founds=self.CDLICatalog.search({'title':objId}) - if len(founds)>0: - return founds[0].getObject().absolute_url() - - else: #assume version number - splitted=objId.split("_") - founds=self.CDLICatalog.search({'title':splitted[1]}) - return founds[0].getObject().absolute_url()+'/'+objId - - def storeAllLink(self,results): - """erzeuge link zum speicher aller results""" - nr=self.REQUEST['_ZopeId'] - - if results: - self._v_stack[nr]=[x.getObject().getId() for x in results] - - return self.absolute_url()+"/storeAll?id="+nr - - def storeAll(self,id): - """store all""" - try: - results=self._v_stack[id] - except: - #TODO: write expired page - return "expired" - - return self.storeInBasketForm(results) - - def storeInBasketForm(self,ids): - """ store an object form""" - - if type(ids) is not ListType: - ids=[ids] - self.REQUEST.SESSION['ids']=ids[0:] - - self.REQUEST.SESSION['BACKLINK']=self.REQUEST['HTTP_REFERER'] - - pt=PageTemplateFile(os.path.join(package_home(globals()),'zpt','storeBasketObject.zpt')).__of__(self) - return pt() - - def storeInBasket(self,username,ids=None,RESPONSE=None,REQUEST=None): - """store it""" - - if not ids: - ids=REQUEST.SESSION['ids'] + data=file(os.path.join(dir,fn)).read() + status,msg=checkFile(fn,data,dir) + #status=True - self.REQUEST.SESSION['basketUser']=username - - baskets=self.ZopeFind(self,obj_ids=[username]) - if len(baskets)>0: - basket=baskets[0][1] - else: - manage_addBasketObject(self,username) - basket=self._getOb(username) - - - basket.addObjects(ids) - back=self.REQUEST.SESSION.get('BACKLINK', None) - - if RESPONSE: - RESPONSE.redirect(back) - - - def showBasket(self,user=None,set=None,RESPONSE=None): - """show the basket""" - - if user: - self.REQUEST.SESSION['basketUser']=user - - if not user and not set: - user=self.REQUEST.SESSION.get('basketUser',None) - - if not user: - pt=PageTemplateFile(os.path.join(package_home(globals()),'zpt','orizeBasketUser.zpt')).__of__(self) - return pt() - else: - baskets=self.ZopeFind(self,obj_ids=[user]) - - - if len(baskets)>0: - RESPONSE.redirect(baskets[0][1].absolute_url()) - return True - else: - pt=PageTemplateFile(os.path.join(package_home(globals()),'zpt','emptyBasket.zpt')).__of__(self) - return pt() - - -def manage_addBasket_oldForm(self): - """add the basket form""" - pt=PageTemplateFile(os.path.join(package_home(globals()),'zpt','addBasket.zpt')).__of__(self) - return pt() - -def manage_addBasket_old(self,id,title,RESPONSE=None): - """add the basket""" - ob=Basket() - - ob.id=str(id) - ob.title=title - self._setObject(id, ob) - ob=self._getOb(id) - - if RESPONSE is not None: - RESPONSE.redirect('manage_main') - - -class BasketObject_old(Folder): - """Basket Object - alte fassung""" - - meta_type="basketObject" - def __init__(self): - """init basket object""" - self.contents=[] - - def numberOfItems(self): - """return anzahl der elemente im basket""" - return len(self.contents) - - def addObjects(self,ids): - """addObjects""" - - for id in ids: - founds=self.CDLICatalog.search({'title':id}) - for found in founds: - if found.getObject() not in self.contents: - tm=self.contents[0:] - tm.append(found.getObject()) - self.contents=tm[0:] - - return True - - def index_html(self): - """view the basket""" - pt=PageTemplateFile(os.path.join(package_home(globals()),'zpt','basketObject_index_html.zpt')).__of__(self) - return pt() - - def deleteObjects(self,ids,RESPONSE=None): - """delete objects""" - list = self.contents[0:] - for content in list: - - if content.getId() in ids: - self.contents.remove(content) - - - if RESPONSE: - RESPONSE.redirect(self.absolute_url()) - - - def unlockTest(self): - """unlock all files of the testuser for debuggin""" - for object in self.contents: - - if str(object.lockedBy)=="test": - object.lockedBy="" + if not status: # error + errors.append((fn,msg)) - def downloadObjectsAsOneFile(self,lock=None,procedure=None,REQUEST=None): - """download all selected files in one file""" - - ret="" - lockedObjects={} - + else: + if len(founds)==0: + newPs.append(fn) - if lock: - - if str(self.REQUEST['AUTHENTICATED_USER'])=='Anonymous User': + #if p file alread at the server + for found in founds: + #analyse the differences to the actual file + obj=found.getObject() + + if (not (str(obj.lockedBy))=='') and (not (str(obj.lockedBy)==str(self.username))): + lockerrors.append((fn,str(obj.lockedBy))) + else: - return "please login first" + diffs=obj.diff(data) + if diffs[0]>0: + changed.append((obj,diffs)) #hochladen - #check if a locked object exist in the basket. - lockedObjects={} - for object in self.contents: - - if not object.lockedBy=="": - lockedObjects[object.title]=repr(object.lockedBy) - - - keys=lockedObjects.keys() - - - if len(keys)>0 and (not procedure): - self.REQUEST.SESSION['lockedObjects']=lockedObjects - pt=PageTemplateFile(os.path.join(package_home(globals()),'zpt','lockedObjects.zpt')).__of__(self) - return pt() - - elif not procedure: #keine fails gesperrt dann alle donwloaden - procedure="downloadAll" + #ready, set the returnValues + self.result+="

Done

" - - for object in self.contents: - - if (procedure=="downloadAll") or (object.lockedBy=='') or (object.lockedBy==self.REQUEST['AUTHENTICATED_USER']): - ret+=object.getLastVersion().data - - if lock and object.lockedBy=='': - object.lockedBy=self.REQUEST['AUTHENTICATED_USER'] - - - self.REQUEST.RESPONSE.setHeader("Content-Disposition","""attachement; filename="basket_%s.atf" """%self.getId()) - self.REQUEST.RESPONSE.setHeader("Content-Type","application/octet-stream") - length=len(ret) - self.REQUEST.RESPONSE.setHeader("Content-Length",length) - self.REQUEST.RESPONSE.write(ret) + stObj.returnValue={} + stObj.returnValue['errors']=errors -def manage_addBasket_oldObjectForm(self): - """add form""" - pass - -def manage_addBasket_oldObject(self,id,title='',RESPONSE=None): - """add""" - - ob=BasketObject() - - ob.id=str(id) - ob.title=title - self._setObject(id, ob) - ob=self._getOb(id) + stObj.returnValue['newPs']=newPs + stObj.returnValue['tmpdir']=dir + stObj.returnValue['basketLen']=basketLen + stObj.returnValue['numberOfFiles']=numberOfFiles + stObj.returnValue['basketNameFromId']=basketNameFromId + stObj.returnValue['basketNameFromFile']=basketNameFromFile + stObj.returnValue['basketId']=basketId + stObj.returnValue['dir']=dir + #stObj.returnValue['changed']=copy.copy(changed) + stObj.returnValue['changed']=[(x[0].getId(),x[1][0]) for x in changed] + #stObj.returnValue['lockerrors']=[x[0].getId() for x in lockerrors] + stObj.returnValue['lockerrors']=[x for x in lockerrors] + self.returnValue=True + #ctx2.cdli_main.setTemp('v_uploadATF_returnValue',True) - if RESPONSE is not None: - RESPONSE.redirect('manage_main') - - + class CDLIBasketContainer(OrderedFolder): """contains the baskets""" @@ -560,6 +539,55 @@ class CDLIBasketContainer(OrderedFolder) security=ClassSecurityInfo() meta_type="CDLIBasketContainer" + def getPNumbersOfBasket(self,basketName): + """get all pnumbers of a basket as a list, returns an empty list if basket not found + @param basketName: name of the basket + """ + ret=[] + basketId=self.getBasketIdfromName(basketName) + if not basketId: + return [] + + ob=getattr(self,basketId).getContent() + + ret=[x[0].split(".")[0] for x in ob] + + return ret + + security.declareProtected('manage','getBasketAsOneFile') + def getBasketAsOneFile(self,basketName,current="no"): + """returns all files of the basket combined in one file + @param basketName: Name of the basket + @param current: (optional) if current is set to "yes" then the most current version of + all files are downloaded and not the versions of the files as stored in the basket + """ + ret="" + basketId=self.getBasketIdfromName(basketName) + if not basketId: + return "" + + ob=getattr(self,basketId).getLastVersion() + for object in ob.getContent(): + if current=="no": #version as they are in the basket + ret+=str(object[0].getData())+"\n" + elif current=="yes": + #search current object + logging.info("crrent: %s"%object[1].getId().split(".")[0]) + founds=self.CDLICatalog.search({'title':object[1].getId().split(".")[0]}) + if len(founds)>0: + ret+=str(founds[0].getObject().getLastVersion().getData())+"\n" + return ret + + security.declareProtected('manage','upDateBaskets') + def upDateBaskets(self): + """update content in to objects""" + + founds=self.ZopeFind(self,obj_metatypes=['CDLIBasketVersion'],search_sub=1) + + for found in founds: + found[1].updateBasket() + + security.declareProtected('manage','deleteBaskets') def deleteBaskets(self,ids=None): """delete baskets, i.e. move them into trash folder""" @@ -577,15 +605,25 @@ class CDLIBasketContainer(OrderedFolder) cut=self.manage_cutObjects(ids) trash.manage_pasteObjects(cut) - def manageBaskets(self,ids,submit,REQUEST=None,RESPONSE=None): + security.declareProtected('manage','manageBaskets') + def manageBaskets(self,submit,ids=None,basket1="",basket2="",joinBasket="",subtractBasket="",REQUEST=None,RESPONSE=None): """manage baskets, delete or copy""" if submit=="delete": self.deleteBaskets(ids) - + elif submit=="join": + flag,msg=self.joinBasket(joinBasket, ids) + logging.info("joining %s %s"%(flag,msg)) + + elif submit=="subtract": + logging.info("BBBb %s %s"%(basket1,basket2)) + flag,msg=self.subtractBasket(subtractBasket, basket1,basket2) + logging.info("subtract %s %s"%(flag,msg)) if RESPONSE: RESPONSE.redirect(self.absolute_url()) + + security.declareProtected('View','getBasketIdfromName') def getBasketIdfromName(self,basketname): """get id from name""" @@ -611,7 +649,7 @@ class CDLIBasketContainer(OrderedFolder) return pt(basketId=basketId,basketName=basketName) - + security.declareProtected('manage','index_html') def index_html(self): """stanadard ansicht""" @@ -633,7 +671,19 @@ class CDLIBasketContainer(OrderedFolder) self.id=id self.title=title - + + def getBasketsId(self): + """get all baskets als klartext""" + + ret="" + baskets=self.ZopeFind(self,obj_metatypes=['CDLIBasket']) + for basket in baskets: + com,user,time,values = basket[1].getContentIds() + ret+= "BASKET:"+com+"\t"+user+"\t"+time+"\n" + for x in values: + ret+= x[0]+"\t"+x[1]+"\n" + return ret + def getBaskets(self,sortField='title'): """get all baskets files""" @@ -691,9 +741,83 @@ class CDLIBasketContainer(OrderedFolder) baskets.sort(sortComment) return baskets + + + def subtractBasket(self,newBasket,basket1,basket2): + """subtract basket2 from basket1 + (i.e. newbasket will contain alle elements of basket1 which are not in basket2), + if basket2 contains files which are not in basket1, then theses files fill be ignored + + @param newbasket: name of the new basket + @param basket1: basket where basket2 will be subtracted from + @param basket2: see above + + """ + logging.info("CCCCC %s %s"%(basket1,basket2)) + + try: + newB=self.addBasket(newBasket) + except: + return False, "cannot create the new basket" + + - - + + + bas2= getattr(self,basket2) + bas2content=bas2.getContent() + bas2ids=[x[0] for x in bas2content] + + + + bas1= getattr(self,basket1) + bas1content=bas1.getContent() + + + newBasketContent={} + + for id,version in bas1content: + if not (id in bas2ids): + newBasketContent[id]=version + + username=self.getActualUserName() + + logging.info("sbc %s"%newBasketContent) + newB.addObjectsWithVersion(newBasketContent,username=username,catalog=self.CDLICatalog) + + return True, "" + + + def joinBasket(self,newBasket,oldBaskets): + """join two baskets + @param newbasket: name of the new basket + @param oldbaskets: list of baskets to be joined + """ + try: + newB=self.addBasket(newBasket) + except: + return False, "cannot create the new basket" + + newBasketContent={} + for ob in oldBaskets: + x= getattr(self,ob,None) + if x is None: + return False, "cannot find basket: %s"%ob + + ids=x.getContent() # hole den Inhalt + + for id,version in ids: + if newBasketContent.has_key(id): # p number gibt's schon + newBasketContent[id]=max(newBasketContent[id],version) # speichere die groessere Versionsnumber + else: + newBasketContent[id]=version + username=self.getActualUserName() + + logging.info("nbc %s"%newBasketContent) + newB.addObjectsWithVersion(newBasketContent,username=username,catalog=self.CDLICatalog) + + return True, "" + def getNewId(self): """createIds""" last=getattr(self,'last',0) @@ -707,9 +831,13 @@ class CDLIBasketContainer(OrderedFolder) def setActiveBasket(self,basketId,REQUEST=None): """store active basketId in a cookie""" self.REQUEST.RESPONSE.setCookie("CDLIActiveBasket",basketId,path="/") - + try: + qs=cgi.parse_qs(REQUEST['QUERY_STRING']) + del(qs['basketId']) + except: + qs={} if REQUEST: - REQUEST.RESPONSE.redirect(REQUEST['URL1']+'?'+REQUEST['QUERY_STRING']) + REQUEST.RESPONSE.redirect(REQUEST['URL1']+'?'+urllib.urlencode(qs)) def getActiveBasket(self): """get active basket from cookie""" @@ -725,13 +853,13 @@ class CDLIBasketContainer(OrderedFolder) """get name of the actualuser""" return str(self.REQUEST['AUTHENTICATED_USER']) - + security.declareProtected('manage','addBasket') def addBasket(self,newBasketName): """add a new basket""" ob=manage_addCDLIBasket(self,newBasketName) return ob - + def storeInBasket(self,submit,ids=None,newBasketName=None,fromFileList=None,RESPONSE=None,REQUEST=None): """store it""" if not ids: @@ -781,6 +909,43 @@ class CDLIBasket(Folder,CatalogAware): meta_type="CDLIBasket" default_catalog="CDLIBasketCatalog" + def searchInBasket(self,indexName,searchStr,regExp=False): + """searchInBasket""" + + lst=self.searchInLineIndexDocs(indexName,searchStr,uniq=True,regExp=regExp) + ret={} + + lv=self.getLastVersion() + + + for obj in lv.content.getContent(): + id=obj[1].getId().split(".")[0] + if id in lst: + + ret[id]=self.showWordInFile(id,searchStr,lineList=self.getLinesFromIndex(indexName,searchStr,id,regExp=regExp),regExp=regExp,indexName=indexName) + + + pt=PageTemplateFile(os.path.join(package_home(globals()),'zpt','searchResultsInBasket')).__of__(self) + return pt(result=ret,indexName=indexName,regExp=regExp,word=searchStr) + + + + + def searchInBasket_v1(self,searchStr): + """search occurences of searchStr in files im basket""" + ret=[] + lv=self.getLastVersion() + logging.info("searching") + for obj in lv.content.getContent(): + txt=obj[0].getData() + for x in txt.split("\n"): + logging.info("search %s"%x) + if re.match(searchStr,x): + ret.append(x) + + return "\n".join(ret) + + def getFile(self,obj): return obj[1] @@ -792,12 +957,16 @@ class CDLIBasket(Folder,CatalogAware): return [x[1].getId() for x in self.getLastVersion().getContent()] + def isActual(self,obj): """teste ob im basket die aktuelle version ist""" actualNo=obj[1].getLastVersion().getVersionNumber() storedNo=obj[0].getVersionNumber() founds=self.CDLICatalog.search({'title':obj[0].getId()}) + if len(founds)>0: + actualNo=founds[0].getObject().getLastVersion().getVersionNumber() + if len(founds)>0 and founds[0].getObject().aq_parent.getId()==".trash": return False, -1 @@ -828,16 +997,30 @@ class CDLIBasket(Folder,CatalogAware): self.shortDescription=shortDescription self.comment=comment - + def getActualUserName(self): + """get name of the actualuser""" + + return str(self.REQUEST['AUTHENTICATED_USER']) + def getLastVersion(self): """hole letzte version""" - ids=[int(x[0]) for x in self.ZopeFind(self,obj_metatypes=["CDLIBasketVersion"])] + + ids=[] + idsTmp= self.objectIds() + for x in idsTmp: + try: + ids.append(int(x)) + except: + pass ids.sort() + if len(ids)==0: return None else: ob=getattr(self,str(ids[-1])) + + return ob def getVersions(self): @@ -846,16 +1029,101 @@ class CDLIBasket(Folder,CatalogAware): return versions + def updateObjects(self,ids,RESPONSE=None,REQUEST=None): + """update ids, ids not in the basket the add""" + if type(ids) is not ListType: + ids=[ids] + + lastVersion=self.getLastVersion() + oldContent=lastVersion.content.getContent() + newContent=[] + + #first copy the old + for obj in oldContent: + if obj[1].getId() not in ids: + newContent.append(obj) + #now add the new + + for id in ids: + founds=self.CDLICatalog.search({'title':id}) + + for found in founds: + if found.getObject() not in oldContent: + #TODO: was passiert wenn, man eine Object dazufŸgt, das schon da ist aber eine neuere version + newContent.append((found.getObject().getLastVersion(),found.getObject())) + + + content=newContent + user=self.getActualUserName() + + ob=manage_addCDLIBasketVersion(self,user,comment="",basketContent=newContent) + + obj=self._getOb(ob.getId()) + if RESPONSE: + + RESPONSE.redirect(obj.absolute_url()) + + return obj + + def addObjectsWithVersion(self,ids,deleteOld=None,username=None,catalog=None): + """generate a new version of the basket with objects added, + hier wird jedoch nicht die letzte Version jedes Files hinzugefuegt, s + ondern ids is ein Tupel mit der Id (d.h. der p-number) und der Versionsnummer. + """ + logging.info("add to basket (%s)"%(self.getId())) + lastVersion=self.getLastVersion() + + if not catalog: + catalog=self.CDLICatalog + + if lastVersion is None: + oldContent=[] + else: + oldContent=lastVersion.content.getContent() + + if deleteOld: + oldContent=[] + + newContent=[] + added=0 + + for id,version in ids.iteritems(): + logging.info("adding %s %s"%(id,version)) + id=id.split(".")[0] # title nur die pnumber ohne atf + + try: + founds=catalog.search({'title':id}) + except: + founds=[] + logging.info(" found %s "%(founds)) + for found in founds: + if found.getObject() not in oldContent: + + #TODO: was passiert wenn, man eine Object dazufŸgt, das schon da ist aber eine neuere version + newContent.append((found.getObject().getVersions()[version-1][1],found.getObject())) + added+=1 + + content=oldContent+newContent + if not username: + logging.error("XXXXXXXXXXX %s"%repr(self)) + user=self.getActualUserName() + else: + user = username + + ob=manage_addCDLIBasketVersion(self,user,comment="",basketContent=content) + logging.info("add to basket (%s) done"%(self.getId())) + return added + def addObjects(self,ids,deleteOld=None,username=None): """generate a new version of the basket with objects added""" - + logging.info("add to basket (%s)"%(self.getId())) lastVersion=self.getLastVersion() if lastVersion is None: oldContent=[] else: - oldContent=lastVersion.basketContent[0:] + oldContent=lastVersion.content.getContent() if deleteOld: oldContent=[] @@ -863,8 +1131,11 @@ class CDLIBasket(Folder,CatalogAware): newContent=[] added=0 for id in ids: - founds=self.CDLICatalog.search({'title':id}) - + try: + founds=self.CDLICatalog.search({'title':id}) + except: + founds=[] + for found in founds: if found.getObject() not in oldContent: #TODO: was passiert wenn, man eine Object dazufŸgt, das schon da ist aber eine neuere version @@ -878,9 +1149,39 @@ class CDLIBasket(Folder,CatalogAware): user = username ob=manage_addCDLIBasketVersion(self,user,comment="",basketContent=content) - + logging.info("add to basket (%s) done"%(self.getId())) return added + + + def getContent(self): + """print content""" + ret=[] + + lv=self.getLastVersion() + for obj in lv.content.getContent(): + logging.info("XXXXXXXXXX %s"%repr(obj)) + ret.append((obj[1].getId(),obj[0].versionNumber)) + + return ret + + def getContentIds(self): + """print basket content""" + ret=[] + lv=self.getLastVersion() + for obj in lv.content.getContent(): + ret.append((obj[0].getId(),obj[1].getId())) + + + return lv.getComment(),lv.getUser(),lv.getTime(),ret + + def changeBasket(self,ids,submit,RESPONSE=None,REQUEST=None): + """change a basket""" + if submit=="update": + return self.updateObjects(ids,RESPONSE=RESPONSE,REQUEST=REQUEST) + elif submit=="delete": + return self.deleteObjects(ids,RESPONSE=RESPONSE,REQUEST=REQUEST) + def deleteObjects(self,ids,RESPONSE=None,REQUEST=None): """delete objects""" @@ -888,7 +1189,7 @@ class CDLIBasket(Folder,CatalogAware): ids=[ids] lastVersion=self.getLastVersion() - oldContent=lastVersion.basketContent[0:] + oldContent=lastVersion.content.getContent() newContent=[] for obj in oldContent: if obj[1].getId() not in ids: @@ -922,29 +1223,92 @@ def manage_addCDLIBasket(self,title,shor else: return ob -class CDLIBasketVersion(SimpleItem): +class CDLIBasketVersion(Implicit,Persistent,Folder): """version of a basket""" meta_type="CDLIBasketVersion" + security=ClassSecurityInfo() - def downloadObjectsAsOneFile(self,lock=None,procedure=None,REQUEST=None): + def updateBasket(self): + """update""" + try: + self._setObject('content',BasketContent(self.basketContent)) + except: + try: + if len(self.basketContent)>0: + self.content.setContent(self.basketContent) + except: + print "error",self.getId(),self.aq_parent.getId() + self.basketContent=[] + + + def containsNonActualFiles(self): + """returns True if basket contains one or more non current files""" + + objs=self.getContent() + for obj in objs: + if not self.isActual(obj)[0]: + return True + return False + + def downloadListOfPnumbers(self): + """download pnumbers of the basket as list""" + + basket_name=self.aq_parent.title + + ids=self.getContent() # get the list of objects + logging.error(ids) + ret="\n".join([x[1].getId().split(".")[0] for x in ids]) + + self.REQUEST.RESPONSE.setHeader("Content-Disposition","""attachement; filename="%s.txt" """%basket_name) + self.REQUEST.RESPONSE.setHeader("Content-Type","application/octet-stream") + length=len(ret) + self.REQUEST.RESPONSE.setHeader("Content-Length",length) + self.REQUEST.RESPONSE.write(ret) + + security.declareProtected('manage','downloadObjectsAsOneFile') + def downloadObjectsAsOneFile(self,lock=None,procedure=None,REQUEST=None,check="yes",current="no"): """download all selected files in one file""" + if self.temp_folder.downloadCounterBaskets > 10000: + return """I am sorry, currently the server has to many requests for downloads, please come back later!""" + + + if (check=="yes") and self.containsNonActualFiles(): + pt=PageTemplateFile(os.path.join(package_home(globals()),'zpt','downloadObjectAsOneFile_check.zpt')).__of__(self) + + return pt(lock=lock) + + else: + + return self.downloadObjectsAsOneFileFinally(lock=lock,procedure=procedure,REQUEST=REQUEST,current="no") + + def downloadObjectsAsOneFileFinally(self,lock=None,procedure=None,REQUEST=None,current="no"): + """print do the download""" + ret="" lockedObjects={} - + self.temp_folder.downloadCounterBaskets+=1 + self._p_changed=1 + transaction.get().commit() + if lock: if str(self.REQUEST['AUTHENTICATED_USER'])=='Anonymous User': - + self.temp_folder.downloadCounterBaskets-=1 + self._p_changed=1 + transaction.get().commit() + self.temp_folder.downloadCounterBaskets-=1 + self._p_changed=1 + transaction.get().commit() return "please login first" #check if a locked object exist in the basket. lockedObjects={} - for object in self.basketContent: + for object in self.content.getContent(): - if not object[1].lockedBy=="": + if (not str(object[1].lockedBy)=="") and (not (str(object[1].lockedBy)==str(self.REQUEST['AUTHENTICATED_USER']))): lockedObjects[object[1].title]=repr(object[1].lockedBy) @@ -954,35 +1318,51 @@ class CDLIBasketVersion(SimpleItem): if len(keys)>0 and (not procedure): self.REQUEST.SESSION['lockedObjects']=lockedObjects pt=PageTemplateFile(os.path.join(package_home(globals()),'zpt','lockedObjects.zpt')).__of__(self) + + self.temp_folder.downloadCounterBaskets-=1 + self._p_changed=1 + transaction.get().commit() + return pt() elif not procedure: #keine fails gesperrt dann alle donwloaden procedure="downloadAll" + - for object in self.basketContent: - + + for object in self.content.getContent(): + if (procedure=="downloadAll") or (object[1].lockedBy=='') or (object[1].lockedBy==self.REQUEST['AUTHENTICATED_USER']): - ret+=object[0].data - + if current=="no": #version as they are in the basket + ret+=str(object[0].getData())+"\n" + elif current=="yes": + #search current object + founds=self.CDLICatalog.search({'title':object[1].getId().split(".")[0]}) + if len(founds)>0: + ret+=str(founds[0].getObject().getLastVersion().getData())+"\n" + if lock and object[1].lockedBy=='': object[1].lockedBy=self.REQUEST['AUTHENTICATED_USER'] - basket_name=self.aq_parent.title+"_V"+self.getId() #write basketname to header of atf file - ret="#atf basket %s\n"%basket_name+ret + ret="#basket: %s\n"%basket_name+ret + + self.temp_folder.downloadCounterBaskets-=1 + self._p_changed=1 + transaction.get().commit() self.REQUEST.RESPONSE.setHeader("Content-Disposition","""attachement; filename="%s.atf" """%basket_name) self.REQUEST.RESPONSE.setHeader("Content-Type","application/octet-stream") length=len(ret) self.REQUEST.RESPONSE.setHeader("Content-Length",length) self.REQUEST.RESPONSE.write(ret) + return True - def numberOfItems(self): """return anzahl der elemente im basket""" - return len(self.basketContent) + return self.content.numberOfItems() def getTime(self): """getTime""" @@ -998,14 +1378,15 @@ class CDLIBasketVersion(SimpleItem): def getContent(self): """get Basket Content""" - return self.basketContent + return self.content.getContent() def __init__(self,id,user,comment="",basketContent=[]): """ init a basket version""" self.id=id - self.coment=comment - self.basketContent=basketContent[0:] + self.comment=comment + self._setObject('content',BasketContent(basketContent)) + #self.basketContent=basketContent[0:]a self.user=user self.time=time.localtime() @@ -1017,8 +1398,15 @@ class CDLIBasketVersion(SimpleItem): """get Comment""" return self.comment + security.declareProtected('manage','index_html') def index_html(self): """view the basket""" + + if self.REQUEST.get('change',False): + ob=self.aq_parent.updateObjects(self.REQUEST['change']) + + self.REQUEST.RESPONSE.redirect(ob.absolute_url())#go to new basket, because changing generates a new basket + pt=PageTemplateFile(os.path.join(package_home(globals()),'zpt','BasketVersionMain.zpt')).__of__(self) return pt() @@ -1054,7 +1442,7 @@ def manage_addCDLIBasketVersion(self,use else: return ob -class CDLIFileObject(CatalogAware,versionedFileObject): +class CDLIFileObject(CatalogAware,extVersionedFileObject): """CDLI file object""" meta_type="CDLI File Object" @@ -1065,6 +1453,7 @@ class CDLIFileObject(CatalogAware,versio security.declarePublic('makeThisVersionCurrent') + security.declareProtected('manage','index_html') def PrincipiaSearchSource(self): """Return cataloguable key for ourselves.""" return str(self) @@ -1080,7 +1469,7 @@ class CDLIFileObject(CatalogAware,versio newversion=parent.manage_addCDLIFileObject('',comment,author) - newversion.data=self.data[0:] + newversion.manage_upload(self.getData()) if RESPONSE is not None: RESPONSE.redirect(self.aq_parent.absolute_url()+'/history') @@ -1089,19 +1478,39 @@ class CDLIFileObject(CatalogAware,versio return True security.declarePublic('view') - + + def getFormattedData(self): + """fromat text""" + data=self.getData() +# return re.sub("\s\#lem"," #lem",data) #remove return vor #lem + return re.sub("#lem"," #lem",data) #remove return vor #lem + def view(self): """view file""" pt=PageTemplateFile(os.path.join(package_home(globals()),'zpt','viewCDLIFile.zpt')).__of__(self) return pt() + security.declarePublic('getPNumber') + def getPNumber(self): + """get the pnumber""" + try: + txt=re.match("&[Pp](\d*)\s*=([^\r\n]*)",self.getData()[0:]) + except: + txt=self.getData()[0:] + + return "ERROR" + try: + return "P"+txt.group(1) + except: + return "ERROR" + security.declarePublic('getDesignation') def getDesignation(self): """get the designation out of the file""" try: - txt=re.match("&[Pp](\d*)\s*=([^\r\n]*)",self.data[0:]) + txt=re.match("&[Pp](\d*)\s*=([^\r\n]*)",self.getData()[0:]) except: - txt=self.data[0:] + txt=self.getData()[0:] return "ERROR" try: @@ -1112,11 +1521,11 @@ class CDLIFileObject(CatalogAware,versio manage_addCDLIFileObjectForm=DTMLFile('dtml/fileAdd', globals(),Kind='CDLIFileObject',kind='CDLIFileObject', version='1') def manage_addCDLIFileObject(self,id,vC='',author='', file='',title='',precondition='', content_type='', - REQUEST=None): + from_tmp=False,REQUEST=None): """Add a new File object. Creates a new File object 'id' with the contents of 'file'""" - + id=str(id) title=str(title) content_type=str(content_type) @@ -1133,23 +1542,42 @@ def manage_addCDLIFileObject(self,id,vC= setattr(self._getOb(id),'author',author) + # Now we "upload" the data. By doing this in two steps, we # can use a database trick to make the upload more efficient. - if file: + + if file and not from_tmp: self._getOb(id).manage_upload(file) + elif file and from_tmp: + self._getOb(id).manage_upload_from_tmp(file) if content_type: self._getOb(id).content_type=content_type + self.reindex_object() + self._getOb(id).reindex_object() + if REQUEST is not None: REQUEST['RESPONSE'].redirect(self.absolute_url()+'/manage_main') -class CDLIFile(versionedFile,CatalogAware): +class CDLIFile(extVersionedFile,CatalogAware): """CDLI file""" + security=ClassSecurityInfo() meta_type="CDLI file" default_catalog='CDLICatalog' - + security.declareProtected('manage','index_html') #security.declarePublic('history') + def getLastVersionData(self): + """get last version data""" + return self.getLastVersion().getData() + + def getLastVersionFormattedData(self): + """get last version data""" + return self.getLastVersion().getFormattedData() + + #security.declarePublic('history') + + def history(self): """history""" @@ -1161,6 +1589,18 @@ class CDLIFile(versionedFile,CatalogAwar return pt() + def getBasketFromId(self,basketid, context=None): + """get basket from id""" + + if not context: + context=self + + for basket in self.ZopeFind(context,obj_metatypes=["CDLIBasket"]): + if basket[0]==basketid: + return basket[1] + else: + None + def isContainedInBaskets(self,context=None): """check is this file is part of any basket @@ -1193,8 +1633,15 @@ class CDLIFile(versionedFile,CatalogAwar else: return "Sorry file is locked by somebody else" - def manage_addCDLIFileObject(self,id,vC,author,file='',title='',precondition='', content_type='',changeName='no',newName='', RESPONSE=None): + def manage_addCDLIFileObject(self,id,vC,author, + file='',title='', + precondition='', + content_type='', + changeName='no',newName='', + come_from=None, + from_tmp=False,RESPONSE=None): """add""" + try: #TODO: der ganze vC unsinn muss ueberarbeitet werden vC=self.REQUEST['vC'] except: @@ -1214,8 +1661,7 @@ class CDLIFile(versionedFile,CatalogAwar - - + positionVersionNum=getattr(self,'positionVersionNum','front') if positionVersionNum=='front': @@ -1227,15 +1673,18 @@ class CDLIFile(versionedFile,CatalogAwar else: id=tmp[0]+"_V%i"%self.getVersion() - - manage_addCDLIFileObject(self,id,vC,author,file,id,precondition, content_type) - objs=self.ZopeFind(self,obj_ids=[id])[0][1].setVersionNumber(int(self.getVersion())) + + manage_addCDLIFileObject(self,id,vC,author,file,id,precondition, content_type,from_tmp=from_tmp) + #objs=self.ZopeFind(self,obj_ids=[id])[0][1].setVersionNumber(int(self.getVersion())) + objs=getattr(self,id).setVersionNumber(int(self.getVersion())) try: #FIXME: wozu ist das gut? self.REQUEST.SESSION['objID_parent']=self.getId() except: pass + if RESPONSE: + obj=self.ZopeFind(self,obj_ids=[id])[0][1] if obj.getSize()==0: self.REQUEST.SESSION['objID']=obj.getId() @@ -1243,7 +1692,10 @@ class CDLIFile(versionedFile,CatalogAwar return pt() else: - RESPONSE.redirect(self.REQUEST['URL2']+'?uploaded=%s'%self.title) + if come_from and (come_from!=""): + RESPONSE.redirect(come_from+"?change="+self.getId()) + else: + RESPONSE.redirect(self.REQUEST['URL2']+'?uploaded=%s'%self.title) else: return self.ZopeFind(self,obj_ids=[id])[0][1] @@ -1264,50 +1716,90 @@ def manage_addCDLIFile(self,id,title,loc self._setObject(id,newObj) - + getattr(self,id).reindex_object() if RESPONSE is not None: RESPONSE.redirect('manage_main') +def checkUTF8(data): + """check utf 8""" + try: + data.encode('utf-8') + return True + except: + return False + - - +def checkFile(filename,data,folder): + """check the files""" + # first check the file name + fn=filename.split(".") # no extension + + if not fn[0][0]=="P": + return False,"P missing in the filename" + elif len(fn[0])!=7: + return False,"P number has not the right length 6" + elif not checkUTF8(data): + return False,"not utf-8" + else: + return True,"" + + def splitatf(fh,dir=None,ext=None): """split it""" ret=None nf=None i=0 - for line in fh.readlines(): - if ext: - i+=1 - if (i%100)==0: - ext.result+="." - if i==10000: - i=0 - ext.result+="
" - #check if basket name is in the first line - if line.find("#atf basket")>=0: - ret=line.replace('#atf basket ','') - ret=ret.split('_')[0] - else: - if (len(line.lstrip())>0) and (line.lstrip()[0]=="&"): #newfile - if nf: - nf.close() #close last file - - - filename=line[1:].split("=")[0].rstrip()+".atf" - if dir: - filename=os.path.join(dir,filename) - nf=file(filename,"w") - if nf: - nf.write(line) + + if (type(fh) is StringType) or (type(fh) is UnicodeType): + iter=fh.split("\n") + else: + iter=fh.readlines() - nf.close() - fh.close() + for lineTmp in iter: + lineTmp=lineTmp.replace(codecs.BOM_UTF8,'') # make sure that all BOM are removed.. + for line in lineTmp.split("\r"): + #logging.log("Deal with: %s"%line) + if ext: + i+=1 + if (i%100)==0: + ext.result+="." + if i==10000: + i=0 + ext.result+="
" + #check if basket name is in the first line + if line.find("#atf basket")>=0: #old convention + ret=line.replace('#atf basket ','') + ret=ret.split('_')[0] + elif line.find("#basket:")>=0: #new convention + ret=line.replace('#basket: ','') + ret=ret.split('_')[0] + + else: + if (len(line.lstrip())>0) and (line.lstrip()[0]=="&"): #newfile + if nf: + nf.close() #close last file + + + filename=line[1:].split("=")[0].rstrip()+".atf" + if dir: + filename=os.path.join(dir,filename) + nf=file(filename,"w") + logging.info("open %s"%filename) + if nf: + nf.write(line.replace("\n","")+"\n") + + try: + nf.close() + except: + pass + + if not((type(fh) is StringType) or (type(fh) is UnicodeType)): + fh.close() return ret,len(os.listdir(dir)) -class CDLIFileFolder(versionedFileFolder): +class CDLIFileFolder(extVersionedFileFolder): """CDLI File Folder""" security=ClassSecurityInfo() @@ -1316,7 +1808,7 @@ class CDLIFileFolder(versionedFileFolder folderMetaType=['CDLI Folder'] default_catalog='CDLICatalog' defaultFileCatalog=default_catalog #wenn dieses definiert ist, wird beim hinzufŸgen einer neuen version eines files dieser catalog neuiniziert - + #downloadCounter=0 # counts how many download for all files currently run, be mehr als 5 wird verweigert. tmpStore2={} def setTemp(self,name,value): """set tmp""" @@ -1354,13 +1846,23 @@ class CDLIFileFolder(versionedFileFolder return ret + def getFile(self,fn): + """get the content of the file fn""" + founds=self.CDLICatalog.search({'title':fn}) + if not founds: + return "" + else: + obj=founds[0].getObject().getLastVersion() + + return obj.getData()[0:] + def checkCatalog(self,fn): """check if fn is in the catalog""" #TODO add checkCatalog - def findObjectsFromListWithVersion(self,list): + def findObjectsFromListWithVersion(self,list,author=None): """find objects from a list with versions @param list: list of tuples (cdliFile,version) """ @@ -1373,16 +1875,27 @@ class CDLIFileFolder(versionedFileFolder pt=getattr(self,'filelistVersioned.html') - return pt(search=list) + return pt(search=list,author=author) - - def findObjectsFromList(self,display=False,start=None,upload=None,list=None,basketName=None,numberOfObjects=None,RESPONSE=None): + def getAllPNumbers(self): + """get a list of all files (resp their p-numbers) stored""" + + ret=[x.getId for x in self.CDLICatalog()] + + return ret + + def findObjectsFromList(self,enterList=None,display=False,start=None,upload=None,list=None,basketName=None,numberOfObjects=None,RESPONSE=None): """findObjectsFromList (, TAB oder LINE separated)""" if upload: # list from file upload txt=upload.read() + + if enterList: + txt=enterList + + if upload or enterList: txt=txt.replace(",","\n") txt=txt.replace("\t","\n") txt=txt.replace("\r","\n") @@ -1409,9 +1922,11 @@ class CDLIFileFolder(versionedFileFolder if list is not None: # got already a list ret=[] for fileId in list: - if len(fileId.split("."))==1: + if fileId.find("*"): #check for wildcards + fileId=fileId + elif len(fileId.split("."))==1: fileId=fileId+".atf" - + ret+=self.CDLICatalog({'title':fileId}) #TODO: get rid of one of these.. ids=[x.getObject().getId() for x in ret] @@ -1431,7 +1946,6 @@ class CDLIFileFolder(versionedFileFolder RESPONSE.redirect("filelist.html?start:int="+str(start)) - security.declareProtected('Manage','createAllFilesAsSingleFile') def createAllFilesAsSingleFile(self,RESPONSE=None): """download all files""" @@ -1441,14 +1955,24 @@ class CDLIFileFolder(versionedFileFolder catalog=getattr(self,self.default_catalog) #tf,tfilename=mkstemp() - - + if not hasattr(self.temp_folder,'downloadCounter'): + self.temp_folder.downloadCounter=0 + + if getattr(self.temp_folder,'downloadCounter',0) > 5: + return """I am sorry, currently the server has to many requests for downloads, please come back later!""" + + self.temp_folder.downloadCounter+=1 + self._p_changed=1 + transaction.get().commit() + list=[(x.getId,x) for x in catalog()] list.sort(sortF) + + RESPONSE.setHeader("Content-Disposition","""attachement; filename=%s"""%"all.atf") RESPONSE.setHeader("Content-Type","application/octet-stream") - + tmp="" for l in list: obj=l[1].getObject() @@ -1456,7 +1980,11 @@ class CDLIFileFolder(versionedFileFolder #os.write(tf,obj.getLastVersion().data) if RESPONSE: - RESPONSE.write(obj.getLastVersion().data[0:]) + RESPONSE.write(obj.getLastVersion().getData()[0:]) + RESPONSE.write("\n") + self.temp_folder.downloadCounter-=1 + self._p_changed=1 + transaction.get().commit() #os.close(tf) #RESPONSE.redirect(self.absolute_url()+"/downloadFile?fn="%tfilename) return True @@ -1489,17 +2017,7 @@ class CDLIFileFolder(versionedFileFolder return ret - def getFolders_OLD(self): - """get all subfolders""" - ret=[] - folders=self.ZopeFind(self,obj_metatypes=self.folderMetaType) - for folder in folders: - ret.append((folder[1], - len(self.ZopeFind(folder[1],obj_metatypes=self.folderMetaType)), - len(getattr(self,self.default_catalog)({'path':folder[0]})) - )) - return ret - + security.declareProtected('manage','index_html') def index_html(self): """main""" ext=self.ZopeFind(self,obj_ids=["index.html"]) @@ -1546,18 +2064,272 @@ class CDLIRoot(Folder): """main folder for cdli""" meta_type="CDLIRoot" + downloadCounterBaskets=0# counts the current basket downloads if counter > 10 no downloads are possible + + def deleteFiles(self,ids): + """delete files (resp. move into .trash folder)""" + # find or generete trash folder + + found=self.ZopeFind(self,obj_ids=['.trash']) + + if len(found)<1: + manage_addCDLIFileFolder(self, '.trash',title="Trash") + trash=self._getOb('.trash') + else: + logging.info(found) + trash=found[0][1] + + + for id in ids: + founds=self.CDLICatalog.search({'title':id.split(".")[0]}) + if founds: + logging.info(founds) + folder=founds[0].getObject().aq_parent #get the parent folder of the object + logging.info(folder) + cut=folder.manage_cutObjects([founds[0].getId]) #cut it out + trash.manage_pasteObjects(cut) #paste it in the trash + + + def findWordRegExp(self,indexName,searchTerm): + """find all words in index which match regexp in SearchTerm + @param indexName: name of the index to be searched in + @param searchTerm: word to be searched""" + + ret=[] + for x in self.lineIndexes[indexName].iterkeys(): + if re.match(searchTerm,x): + ret.append(x) + return ret + + def searchRegExpInLineIndexDocs(self,indexName,searchTerm): + """search in inLineIndex with regexp + @param indexName: name of the index to be searched in + @param searchTerm: term to be searched + """ + if not searchTerm: + return [] + ret=[] + words=self.findWordRegExp(indexName,searchTerm) # suche nach allen Treffern + logging.info("wd:%s"%words) + for word in words: + + ret+=self.searchInLineIndexDocs(indexName,word) + + + x= unique(ret) + logging.info("words_done") + return x + + def showInLineIndex(self): + """get the index for debug purposes""" + print "show" + for key in self.lineIndexes.keys(): + logging.info("index:%s"%key) + for x in self.lineIndexes[key].iterkeys(): + logging.info("word:%s"%repr(x)) + #for y in self.lineIndex[x].iterkeys(): + # print "doc",repr(y),repr(self.lineIndex[x][y]) + + return self.lineIndexes + + def searchInLineIndexDocs(self,indexName,word,uniq=True,regExp=False): + """search occurences in an index + @param indexName: name of the index to be searched in + @param word: word to be searched + @param unique: (optional) unify the list of results + @param regExp: (optional) use regular expressions + """ + + if regExp: + return self.searchRegExpInLineIndexDocs(indexName,word) + + try: + + lst=list(self.lineIndexes[indexName].get(word).keys()) + except: + logging.error("error: searchInLineIndexDocs (%s %s)"%(sys.exc_info()[0:2])) + lst=[] + if uniq: + return unique(lst) + else: + return lst + + def getLinesFromIndex(self,indexName,word,doc,regExp=False): + """return all lines from a document where word is found + @param indexName: Name of the index + @param word: word to be searched + @param doc: name of the document (usuallay the p-number) + @param regExp: (optional) use regExp + """ + + if not regExp: + return self.lineIndexes[indexName].get(word)[doc] + else: # wenn regexp, suche welches word + for w in self.findWordRegExp(indexName,word): + if self.lineIndexes[indexName].get(w): # ein word in im dex gefunden + try: + dc=self.lineIndex[indexName].get(word)[doc] + return dc # und ein document dann gib es zurueck + except: + pass #andernfalls weiter + + def cleanInLineIndex(self,indexName): + """empty an InlineIndex + @param indexName: name of the index + """ + for x in list(self.lineIndexes[indexName].keys()): + del(self.lineIndexes[indexName][x]) + print [x for x in self.lineIndexes[indexName].keys()] + + return "ok" + + def storeInLineIndex(self,indexName,key,value): + """store in index, key is normally a word or grapheme + and value is a tuple (documentname, line) where the word can be found + @param indexName: name of the index + @param key: key in index + @param value: value in index, value is a tuple (document name, line) + """ + logging.error("indexing: %s %s"%(indexName,key)) + if (not hasattr(self,'lineIndexes')): + + self.lineIndexes={} + + if self.lineIndexes.get(indexName,None) is None: + #index exisitiert noch nicht dann anlegen + + self.lineIndexes[indexName]=OOBTree() + lis=self.lineIndexes + li=lis[indexName] + + if li.has_key(key): + +# if li[key].has_key(value[0]) and (not (value[1] in li[key][value[0]])): + if li[key].has_key(value[0]): + tmp=li[key][value[0]] + tmp.append(value[1]) # add it if now in the array + li[key][value[0]]=tmp[0:] + else: + li[key][value[0]]=[value[1]] # new array for lines + + else: + + li[key]=OOBTree()# new btree for lines + li[key][value[0]]=[value[1]] + + + self.lineIndexes=lis + + transaction.get().commit() + + + def showFile(self,fileId,wholePage=False): + """show a file + @param fileId: P-Number of the document to be displayed + """ + f=self.CDLICatalog({'title':fileId}) + if not f: + return "" + + if wholePage: + logging.info("whole") + return f[0].getObject().getLastVersion().view() + else: + return f[0].getObject().getLastVersionFormattedData() + + + def showWordInFile(self,fileId,word,lineList=None,regExp=True,indexName=""): + """get lines with word fromFileId""" + + file=self.showFile(fileId) + logging.info("regEXP %s"%regExp) + ret=[] + if regExp: # wenn regexp dann generiere alle worte aus der list die der regexp entsprechen + wordlist=self.findWordRegExp(indexName,word) + else: + wordlist=[word] + + for line in file.split("\n"): + found=False + for word in wordlist: + try: # just a hack because of possible unicode errors in line + if line.find(word)>-1: + if lineList: #liste of moeglichen Zeilennummern + num=line.split(".")[0] #Zeilenummer ist alles vor dem . in der Zeile + + if num in lineList: + + ret.append(line) + else: # nimm alles ohne line check + ret.append(line) + + break; + except: + pass + return ret + + def tagWordInFile(self,fileId,word,lineList=None,regExp=True,indexName=""): + """get lines with word fromFileId""" + + file=self.showFile(fileId) + tagStr=u'%s' + ret=[] + + if regExp: # wenn regexp dann generiere alle worte aus der list die der regexp entsprechen + wordlist=self.findWordRegExp(indexName,word) + else: + wordlist=[word] + + for line in file.split("\n"): + line = unicodify(line) + found=False + for word in wordlist: + if line.find(word)>-1: #word ist gefunden dann makiere und breche die Schleife ab + if lineList: #liste of moeglichen Zeilennummern + num=line.split(".")[0] #Zeilenummer ist alles vor dem . in der Zeile + + if num in lineList: + + ret.append(line.replace(word,tagStr%word)) + + else: # nimm alles ohne line check + ret.append(line.replace(word,tagStr%word)) + found=True + break + if not found: #word wurde nicht gefunden keine makierung + ret.append(line) + + return u'
\n'.join(ret) + + def URLquote(self,str): + """quote url""" + return urllib.quote(str) + + def URLunquote(self,str): + """unquote url""" + return urllib.unquote(str) + + def URLquote_plus(self,str): + """quote url""" + return urllib.quote_plus(str) + + def URLunquote_plus(self,str): + """unquote url""" + return urllib.unquote_plus(str) + def forceunlock(self): "break all locks" ret=[] for f in self.ZopeFind(self,obj_metatypes="CDLI file",search_sub=1): un=f[1].forceunlock() - print un + if un and un !="": ret.append((f[0],un)) - print ret - return ret - + + return ret + + def getChangesByAuthor(self,author,n=100): """getChangesByAuthor""" zcat=self.CDLIObjectsCatalog @@ -1580,7 +2352,7 @@ class CDLIRoot(Folder): tmp[id]=(x.getObject().aq_parent,nr) - return self.cdli_main.findObjectsFromListWithVersion(list=tmp.values()) + return self.cdli_main.findObjectsFromListWithVersion(list=tmp.values(),author=author) def getLastChanges(self,n=100): """get the last n changes""" @@ -1612,27 +2384,73 @@ class CDLIRoot(Folder): def checkThreads(self): """check threads""" - return threading.enumerate() - + ret="" + for thread in threading.enumerate(): + ret+="

%s (%s): %s

"%(repr(thread),thread.getName(),thread.isAlive()) + + return ret + + + def uploadATFRPC(self,data,username): + """upload an atffile via xml-rpc""" + uploader=uploadATFThread() + + #generate an random id for the upload object + from random import randint + if (not self.REQUEST.SESSION.get('idTmp',None)): + + idTmp=str(randint(0,1000000000)) + self.REQUEST.SESSION['idTmp']=idTmp + else: + idTmp=self.REQUEST.SESSION.get('idTmp',None) + + + uploader.set(data,0,username,idTmp) + + stObj=uploader.run() + + processor=uploadATFfinallyThread() + + basketname=stObj.returnValue['basketNameFromFile'] + + processor.set("uploadchanged",basketname=basketname,SESSION=stObj.returnValue,username=username,serverport=self.REQUEST['SERVER_PORT']) + + processor.run() + + + return generateXMLReturn(stObj.returnValue) + def uploadATF(self,repeat=None,upload=None,basketId=0,RESPONSE=None): - """standard ausgabe""" + """upload an atf file / basket file""" #self._v_uploadATF.returnValue=None + + #generate an random id for the upload thread + from random import randint + if (not self.REQUEST.SESSION.get('idTmp',None)): + idTmp=str(randint(0,1000000000)) + self.REQUEST.SESSION['idTmp']=idTmp + else: + idTmp=self.REQUEST.SESSION.get('idTmp',None) + + threadName=repeat if not threadName or threadName=="": + #new thread not called from the waiting page tmpVar=False thread=uploadATFThread() threadName=thread.getName()[0:] - if not hasattr(self,'_v_uploadATF'): + if (not hasattr(self,'_v_uploadATF')): self._v_uploadATF={} self._v_uploadATF[threadName]=thread #self._xmltrans.start() #thread=Thread(target=self._v_uploadATF) - - self._v_uploadATF[threadName].set(upload,basketId,self.REQUEST['AUTHENTICATED_USER'],serverport=self.REQUEST['SERVER_PORT']) + logging.info("set thread. extern") + self._v_uploadATF[threadName].set(upload,basketId,self.REQUEST['AUTHENTICATED_USER'],idTmp,serverport=self.REQUEST['SERVER_PORT']) #thread.start() + logging.info("start thread. extern") self._v_uploadATF[threadName].start() @@ -1647,14 +2465,14 @@ class CDLIRoot(Folder): else: #recover thread, if lost - if not hasattr(self,'_v_uploadATF'): + if (not hasattr(self,'_v_uploadATF')): self._v_uploadATF={} if not self._v_uploadATF.get(threadName,None): for thread in threading.enumerate(): if threadName == thread.getName(): self._v_uploadATF[threadName]=thread - if not self._v_uploadATF[threadName].returnValue: + if self._v_uploadATF.get(threadName,None) and (not self._v_uploadATF[threadName].returnValue): wait_template=self.aq_parent.ZopeFind(self.aq_parent,obj_ids=['wait_template']) @@ -1666,38 +2484,18 @@ class CDLIRoot(Folder): return pt(txt='/uploadATF',threadName=threadName) else: -# tmp={} -# for key in self._v_uploadATF[threadName].returnValue.keys(): -# t=self._v_uploadATF[threadName].returnValue[key] -# if type(t) is ListType: -# tmp[key]=self._v_uploadATF[threadName].returnValue[key][0:] -# else: -# tmp[key]=self._v_uploadATF[threadName].returnValue[key] -# print repr(tmp[key]),repr(key) -# -# # - #tmp=self.cdli_main.tmpStore2[threadName] - tmp=self._v_uploadATF[threadName].returnValue - - #self._v_uploadATF[threadName].continueVar=False - - self.REQUEST.SESSION['changed']=[x[0].getId() for x in tmp['changed']] - self.REQUEST.SESSION['errors']=[x.getId() for x in tmp['errors']] - self.REQUEST.SESSION['newPs']=tmp['newPs'] - self.REQUEST.SESSION['tmpdir']=tmp['dir'] - #del(self.cdli_main.tmpStore2[threadName]) - print "here the templates" - + tmp=getattr(self.temp_folder,idTmp).returnValue + pt=PageTemplateFile(os.path.join(package_home(globals()),'zpt','uploadCheck.zpt')).__of__(self) - print "init p" - return pt(changed=tmp['changed'],errors=tmp['errors'],dir=tmp['dir'],newPs=tmp['newPs'],basketLen=tmp['basketLen'],numberOfFiles=tmp['numberOfFiles'], + + return pt(changed=tmp['changed'],lockerrors=tmp['lockerrors'],errors=tmp['errors'],dir=tmp['dir'],newPs=tmp['newPs'],basketLen=tmp['basketLen'],numberOfFiles=tmp['numberOfFiles'], basketNameFromId=tmp['basketNameFromId'],basketNameFromFile=tmp['basketNameFromFile'],basketId=tmp['basketId']) def redoUpload(self,threadName): """redo the upload""" tmp=self.cdli_main.tmpStore2[threadName] pt=PageTemplateFile(os.path.join(package_home(globals()),'zpt','uploadCheck.zpt')).__of__(self) - return pt(changed=tmp['changed'],errors=tmp['errors'],dir=tmp['dir'],newPs=tmp['newPs'],basketLen=tmp['basketLen'],numberOfFiles=tmp['numberOfFiles'], + return pt(changed=tmp['changed'],lockerrors=tmp['lockerrors'],errors=tmp['errors'],dir=tmp['dir'],newPs=tmp['newPs'],basketLen=tmp['basketLen'],numberOfFiles=tmp['numberOfFiles'], basketNameFromId=tmp['basketNameFromId'],basketNameFromFile=tmp['basketNameFromFile'],basketId=tmp['basketId']) def uploadATFfinally(self,procedure='',comment="",basketname='',unlock=None,repeat=None,RESPONSE=None): @@ -1709,10 +2507,16 @@ class CDLIRoot(Folder): if not threadName or threadName=="": thread=uploadATFfinallyThread() threadName=thread.getName()[0:] + + if (not hasattr(self,'_v_uploadATF')): + self._v_uploadATF={} + + self._v_uploadATF[threadName]=thread - - self._v_uploadATF[threadName].set(procedure,comment=comment,basketname=basketname,unlock=unlock,SESSION=self.REQUEST.SESSION,username=self.REQUEST['AUTHENTICATED_USER'],serverport=self.REQUEST['SERVER_PORT']) + idTmp=self.REQUEST.SESSION['idTmp'] + stObj=getattr(self.temp_folder,idTmp) + self._v_uploadATF[threadName].set(procedure,comment=comment,basketname=basketname,unlock=unlock,SESSION=stObj.returnValue,username=self.REQUEST['AUTHENTICATED_USER'],serverport=self.REQUEST['SERVER_PORT']) self._v_uploadATF[threadName].start() @@ -1745,14 +2549,14 @@ class CDLIRoot(Folder): pt=PageTemplateFile(os.path.join(package_home(globals()),'zpt','uploadATFWait.zpt')).__of__(self) return pt(txt='/uploadATFfinally',threadName=threadName) else: - + self.REQUEST.SESSION['idTmp']=None if RESPONSE is not None: RESPONSE.redirect(self.absolute_url()) - def importFiles(self,comment="",author="" ,folderName="/Users/dwinter/Documents/workspace/cdli/atf", files=None,ext=None): + def importFiles(self,comment="",author="" ,folderName="/Users/dwinter/atf", files=None,ext=None): """import files""" root=self.cdli_main - + count=0 if not files: files=os.listdir(folderName) @@ -1762,11 +2566,11 @@ class CDLIRoot(Folder): obj=self.ZopeFind(root,obj_ids=[folder]) if ext: - ext.result+="

Adding: %s

"%f + ext.result="

adding: %s

"%f+ext.result if not obj: manage_addCDLIFileFolder(root,folder,folder) fobj=getattr(root,folder) - #get_transaction().commit() + #transaction.get().commit() else: fobj=obj[0][1] @@ -1779,18 +2583,24 @@ class CDLIRoot(Folder): else: fobj2=obj2[0][1] - file2=file(os.path.join(folderName,f)) + file2=os.path.join(folderName,f) id=f manage_addCDLIFile(fobj2,f,'','') id=f ob=fobj2._getOb(f) ob.title=id - manage_addCDLIFileObject(ob,id,comment,author,file2,content_type='') + manage_addCDLIFileObject(ob,id,comment,author,file2,content_type='',from_tmp=True) self.CDLICatalog.catalog_object(ob) #self.CDLICatalog.manage_catalogFoundItems(obj_ids=[id],search_sub=1) #self.CDLICatalog.manage_catalogObject(self.REQUEST, self.REQUEST.RESPONSE, 'CDLICatalog', urlparse.urlparse(ob.absolute_url())[1]) - + count+=1 + + if count > 1000: + print "committing" + transaction.get().commit() + count=0 + transaction.get().commit() return "ok" @@ -1810,7 +2620,10 @@ def manage_addCDLIRoot(self, id, title=' ob=CDLIRoot() ob.id=str(id) ob.title=title - self._setObject(id, ob) + try: + self._setObject(id, ob) + except: + pass ob=self._getOb(id) checkPermission=getSecurityManager().checkPermission @@ -1826,4 +2639,3 @@ def manage_addCDLIRoot(self, id, title=' if REQUEST is not None: return self.manage_main(self, REQUEST, update_menu=1) -