from OFS.SimpleItem import SimpleItem from Products.PageTemplates.PageTemplateFile import PageTemplateFile import xml.etree.ElementTree as ET import re import logging import urllib import urlparse import base64 from HocrTxtUtils import getInt, getText, getHttpData def serialize(node): """returns a string containing an XML snippet of node""" s = ET.tostring(node, 'UTF-8') # snip off XML declaration if s.startswith('') return s[i+3:] return s class HocrTextServer(SimpleItem): """TextServer implementation for MPDL-XML eXist server""" meta_type="Hocr TextServer" manage_options=( {'label':'Config','action':'manage_changeHocrTextServerForm'}, )+SimpleItem.manage_options manage_changeHocrTextServerForm = PageTemplateFile("zpt/manage_changeHocrTextServer", globals()) def __init__(self,id,title="",serverUrl="http://localhost:8080/hocr", timeout=40, repositoryType='production'): """constructor""" self.id=id self.title=title self.timeout = timeout self.repositoryType = repositoryType self.serverUrl = serverUrl def getHttpData(self, url, data=None): """returns result from url+data HTTP request""" return getHttpData(url,data,timeout=self.timeout) def getServerData(self, pn, data=None): """returns result from text server for method+data""" url = self.serverUrl return getHttpData(url,pn,data=data,timeout=self.timeout) def getRepositoryType(self): """returns the repository type, e.g. 'production'""" return getattr(self, 'repositoryType', None) def getTextDownloadUrl(self, type='xml', docinfo=None): """returns a URL to download the current text""" docpath = docinfo.get('textURLPath', None) if not docpath: return None docpath = docpath.replace('.xml','.'+type) url = '%sgetDoc?doc=%s'%(self.serverUrl.replace('interface/',''), docpath) return url def getPlacesOnPage(self, docinfo=None, pn=None): """Returns list of GIS places of page pn""" docpath = docinfo.get('textURLPath',None) if not docpath: return None places=[] text=self.getServerData("xpath.xql", "document=%s&xpath=//place&pn=%s"%(docpath,pn)) dom = ET.fromstring(text) result = dom.findall(".//resultPage/place") for l in result: id = l.get("id") name = l.text place = {'id': id, 'name': name} places.append(place) return places def getTextInfo(self, mode='', docinfo=None): """reads document info, including page concordance, from text server""" logging.debug("getTextInfo mode=%s"%mode) if mode not in ['toc', 'figures', '']: mode = '' # check cached info if mode: # cached toc-request? if 'full_%s'%mode in docinfo: return docinfo else: # no toc-request if 'numTextPages' in docinfo: return docinfo docpath = docinfo.get('textURLPath', None) if docpath is None: logging.error("getTextInfo: no textURLPath!") return docinfo try: # we need to set a result set size pagesize = 10000 pn = 1 # fetch docinfo pagexml = self.getServerData("doc-info.xql","document=%s&info=%s&pageSize=%s&pn=%s"%(docpath,mode,pagesize,pn)) dom = ET.fromstring(pagexml) # all info in tag doc = dom.find("document") except Exception, e: logging.error("getTextInfo: Error reading doc info: %s"%e) return docinfo if doc is None: logging.error("getTextInfo: unable to find document-tag!") else: # go through all child elements for tag in doc: name = tag.tag # numTextPages if name == 'countPages': np = getInt(tag.text) if np > 0: docinfo['numTextPages'] = np # numFigureEntries elif name == 'countFigureEntries': docinfo['numFigureEntries'] = getInt(tag.text) # numTocEntries elif name == 'countTocEntries': # WTF: s1 = int(s)/30+1 docinfo['numTocEntries'] = getInt(tag.text) # numPlaces elif name == 'countPlaces': docinfo['numPlaces'] = getInt(tag.text) # pageNumbers elif name == 'pageNumbers': # contains tags with page numbers # 44 # n=scan number, no=original page no, non=normalized original page no # pageNumbers is a dict indexed by scan number pages = {} for pn in tag: page = {} n = 0 for p in pn: if p.tag == 'n': n = getInt(p.text) page['pn'] = n elif p.tag == 'no': page['no'] = p.text elif p.tag == 'non': page['non'] = p.text if n > 0: pages[n] = page docinfo['pageNumbers'] = pages #logging.debug("got pageNumbers=%s"%repr(pages)) # toc elif name == 'toc': # contains tags with table of contents/figures # 133Chapter I1.1 tocs = [] for te in tag: toc = {} for t in te: if t.tag == 'page': toc['pn'] = getInt(t.text) elif t.tag == 'level': toc['level'] = t.text elif t.tag == 'content': toc['content'] = t.text elif t.tag == 'level-string': toc['level-string'] = t.text elif t.tag == 'real-level': toc['real-level'] = t.text tocs.append(toc) # save as full_toc/full_figures docinfo['full_%s'%mode] = tocs return docinfo def processPageInfo(self, dom, docinfo, pageinfo): """processes page info divs from dom and stores in docinfo and pageinfo""" # assume first second level div is pageMeta alldivs = dom.find("div") if alldivs is None or alldivs.get('class', '') != 'pageMeta': logging.error("processPageInfo: pageMeta div not found!") return for div in alldivs: dc = div.get('class') # pageNumberOrig if dc == 'pageNumberOrig': pageinfo['pageNumberOrig'] = div.text # pageNumberOrigNorm elif dc == 'pageNumberOrigNorm': pageinfo['pageNumberOrigNorm'] = div.text # pageHeaderTitle elif dc == 'pageHeaderTitle': pageinfo['pageHeaderTitle'] = div.text #logging.debug("processPageInfo: pageinfo=%s"%repr(pageinfo)) return def getTextPage(self, mode="text", pn=1, docinfo=None, pageinfo=None): """returns single page from fulltext""" logging.debug("getTextPage Hocr mode=%s, pn=%s"%(mode,pn)) # check for cached text -- but ideally this shouldn't be called twice if pageinfo.has_key('textPage'): logging.debug("getTextPage: using cached text") return pageinfo['textPage'] docpath = docinfo.get('textURLPath', None) docpath=docpath.replace("pages","hocr") logging.debug("getTextPage docpath= %s"%docpath) if not docpath: return None # stuff for constructing full urls selfurl = docinfo['viewerUrl'] textParams = {'document': docpath, 'pn': pn} if 'characterNormalization' in pageinfo: textParams['characterNormalization'] = pageinfo['characterNormalization'] if not mode: # default is dict mode = 'text' logging.debug("getTextPage mode=%s, pn=%s"%(mode,pn)) modes = mode.split(',') # check for multiple layers if len(modes) > 1: logging.debug("getTextPage: more than one mode=%s"%mode) # search mode if 'search' in modes: # add highlighting highlightQuery = pageinfo.get('highlightQuery', None) if highlightQuery: textParams['highlightQuery'] = highlightQuery textParams['highlightElement'] = pageinfo.get('highlightElement', '') textParams['highlightElementPos'] = pageinfo.get('highlightElementPos', '') # ignore mode in the following modes.remove('search') # pundit mode punditMode = False if 'pundit' in modes: punditMode = True # ignore mode in the following modes.remove('pundit') # other modes don't combine if 'dict' in modes: # dict is called textPollux in the backend textmode = 'textPollux' elif 'xml' in modes: # xml mode textmode = 'xml' textParams['characterNormalization'] = 'orig' elif 'gis' in modes: textmode = 'gis' else: # text is default mode textmode = 'text' textParams['mode'] = textmode logging.debug("getTextPage (textparams: %s"%textParams) try: # fetch the page pagexml = self.getServerData(pn,urllib.urlencode(textParams)) return pagexml except Exception, e: logging.error("getTextPage: Error reading page: %s"%e) return None return None def addPunditAttributes(self, pagediv, pageinfo, docinfo): """add about attributes for pundit annotation tool""" textid = docinfo.get('DRI', "fn=%s"%docinfo.get('documentPath', '???')) pn = pageinfo.get('pn', '1') # TODO: use pn as well? # check all div-tags divs = pagediv.findall(".//div") for d in divs: id = d.get('id') if id: d.set('about', "http://echo.mpiwg-berlin.mpg.de/%s/pn=%s/#%s"%(textid,pn,id)) cls = d.get('class','') cls += ' pundit-content' d.set('class', cls.strip()) return pagediv def getSearchResults(self, mode, query=None, pageinfo=None, docinfo=None): """loads list of search results and stores XML in docinfo""" logging.debug("getSearchResults mode=%s query=%s"%(mode, query)) if mode == "none": return docinfo cachedQuery = docinfo.get('cachedQuery', None) if cachedQuery is not None: # cached search result if cachedQuery == '%s_%s'%(mode,query): # same query return docinfo else: # different query del docinfo['resultSize'] del docinfo['resultXML'] # cache query docinfo['cachedQuery'] = '%s_%s'%(mode,query) # fetch full results docpath = docinfo['textURLPath'] params = {'document': docpath, 'mode': 'text', 'queryType': mode, 'query': query, 'queryResultPageSize': 1000, 'queryResultPN': 1, 'characterNormalization': pageinfo.get('characterNormalization', 'reg')} pagexml = self.getServerData("doc-query.xql",urllib.urlencode(params)) #pagexml = self.getServerData("doc-query.xql","document=%s&mode=%s&queryType=%s&query=%s&queryResultPageSize=%s&queryResultPN=%s&s=%s&viewMode=%s&characterNormalization=%s&highlightElementPos=%s&highlightElement=%s&highlightQuery=%s"%(docpath, 'text', queryType, urllib.quote(query), pagesize, pn, s, viewMode,characterNormalization, highlightElementPos, highlightElement, urllib.quote(highlightQuery))) dom = ET.fromstring(pagexml) # page content is in
pagediv = None # ElementTree 1.2 in Python 2.6 can't do div[@class='queryResultPage'] alldivs = dom.findall("div") for div in alldivs: dc = div.get('class') # page content div if dc == 'queryResultPage': pagediv = div elif dc == 'queryResultHits': docinfo['resultSize'] = getInt(div.text) if pagediv is not None: # store XML in docinfo docinfo['resultXML'] = ET.tostring(pagediv, 'UTF-8') return docinfo def getResultsPage(self, mode="text", query=None, pn=None, start=None, size=None, pageinfo=None, docinfo=None): """returns single page from the table of contents""" logging.debug("getResultsPage mode=%s, pn=%s"%(mode,pn)) # get (cached) result self.getSearchResults(mode=mode, query=query, pageinfo=pageinfo, docinfo=docinfo) resultxml = docinfo.get('resultXML', None) if not resultxml: logging.error("getResultPage: unable to find resultXML") return "Error: no result!" if size is None: size = pageinfo.get('resultPageSize', 10) if start is None: start = (pn - 1) * size fullresult = ET.fromstring(resultxml) if fullresult is not None: # paginate first = start-1 len = size del fullresult[:first] del fullresult[len:] tocdivs = fullresult # check all a-tags links = tocdivs.findall(".//a") for l in links: href = l.get('href') if href: # assume all links go to pages linkUrl = urlparse.urlparse(href) linkParams = urlparse.parse_qs(linkUrl.query) # take some parameters params = {'pn': linkParams['pn'], 'highlightQuery': linkParams.get('highlightQuery',''), 'highlightElement': linkParams.get('highlightElement',''), 'highlightElementPos': linkParams.get('highlightElementPos','') } url = self.getLink(params=params) l.set('href', url) return serialize(tocdivs) return "ERROR: no results!" def getToc(self, mode='text', docinfo=None): """returns list of table of contents from docinfo""" logging.debug("getToc mode=%s"%mode) if mode == 'text': queryType = 'toc' else: queryType = mode if not 'full_%s'%queryType in docinfo: # get new toc docinfo = self.getTextInfo(queryType, docinfo) return docinfo.get('full_%s'%queryType, []) def getTocPage(self, mode='text', pn=None, start=None, size=None, pageinfo=None, docinfo=None): """returns single page from the table of contents""" logging.debug("getTocPage mode=%s, pn=%s start=%s size=%s"%(mode,repr(pn),repr(start),repr(size))) fulltoc = self.getToc(mode=mode, docinfo=docinfo) if len(fulltoc) < 1: logging.error("getTocPage: unable to find toc!") return "Error: no table of contents!" if size is None: size = pageinfo.get('tocPageSize', 30) if start is None: start = (pn - 1) * size # paginate first = (start - 1) last = first + size tocs = fulltoc[first:last] tp = '
' for toc in tocs: pageurl = self.getLink('pn', toc['pn']) tp += '
' tp += '
[%s %s]
'%(toc['level-string'], toc['content']) tp += '
Page: %s
'%(pageurl, toc['pn']) tp += '
\n' tp += '
\n' return tp def manage_changeHocrTextServer(self,title="",serverUrl="http://mpdl-text.mpiwg-berlin.mpg.de/mpdl/interface/",timeout=40,repositoryType=None,RESPONSE=None): """change settings""" self.title=title self.timeout = timeout self.serverUrl = serverUrl if repositoryType: self.repositoryType = repositoryType if RESPONSE is not None: RESPONSE.redirect('manage_main') # management methods def manage_addHocrTextServerForm(self): """Form for adding""" pt = PageTemplateFile("zpt/manage_addHocrTextServer", globals()).__of__(self) return pt() def manage_addHocrTextServer(self,id,title="",serverUrl="http://mpdl-text.mpiwg-berlin.mpg.de/mpdl/interface/",timeout=40,RESPONSE=None): #def manage_addHocrTextServer(self,id,title="",serverUrl="http://mpdl-text.mpiwg-berlin.mpg.de:30030/mpdl/interface/",timeout=40,RESPONSE=None): """add zogiimage""" newObj = HocrTextServer(id,title,serverUrl,timeout) self.Destination()._setObject(id, newObj) if RESPONSE is not None: RESPONSE.redirect('manage_main')