from dreqPy import dreq
import hashlib
import collections, string, os, utils_wb
import htmlTemplates as tmpl
import xml, re, uuid
import xml.dom, xml.dom.minidom
import sets
import xlsxwriter

mipTableMap = {'cfMon':'CFmon' }

knowna = [ 'LS3MIP [LWday]', 'LS3MIP [LEday]', 'CFMIP [cf1hrClimMon]', 'HighResMIP [3hr_cloud]',
           'CFMIP [cf3hr_sim_new]', 'C4MIP [L_3hr]', 'DAMIP [DAMIP_day]', 'DAMIP [DAMIP_3hr_p2]',
           'DynVar [DYVR_daily_c]', 'PMIP [PMIP-6hr]', 'HighResMIP [1hrLev]', 'PMIP [PMIP-Amon-02]',
           'PMIP [PMIP-day-02]', 'PMIP [PMIP-aero-02]']
knownl = [ 'PMIP [PMIP-Lmon-02]']
knowno = [ 'DAMIP [DAMIP_Omon_p2]', 'FAFMIP [fafOyr]', 'PMIP [PMIP-Omon-02]']

from utils_wb import uniCleanFunc
from utils_wb import workbook

if os.path.isfile( 'refDefaultP.txt' ):
  refpix = {}
  for l in open( 'refDefaultP.txt' ).readlines():
    bits = string.split( string.strip(l), '\t' )
    assert len(bits) == 4, 'Bad record found in %s' % 'refDefaultP.txt'
    refpix[bits[0]] = (bits[1],bits[2], int(bits[3]) )
else:
  refpix = None

empty=re.compile('^$')

src1 = '../workbook/trial2_20150831.xml'

dq = dreq.loadDreq(dreqXML=src1, manifest=None)
inx = dq.inx
##inx.makeVarRefs()
ix_rql_uid = {}
ix_rqvg_uid = {}
ix_ovar_uid = {}
ix_gpi_uid = {}
list_gp_ovar = collections.defaultdict( list )
xr_var_ovar = collections.defaultdict( list )
xr_var_gpi = collections.defaultdict( list )
rql_by_name = collections.defaultdict( list )

def makeVarRefs(uid, var, iref_by_uid):
    varRefs = {}
    for thisuid in var.uid.keys():
      if iref_by_uid.has_key(thisuid):
        ee1 = collections.defaultdict( list )
        for k,i in iref_by_uid[thisuid]:
          thisi = uid[i]
          sect = thisi._h.label
          if sect == 'groupItem':
            ee1[sect].append( '%s.%s' % (thisi.mip, thisi.group) )
          elif sect == 'ovar':
            ee1[sect].append( thisi.mipTable )
          elif sect == 'revisedTabItem':
            ee1[sect].append( '%s.%s' % (thisi.mip, thisi.table) )
        varRefs[thisuid] = ee1
    return varRefs

varRefs = makeVarRefs( inx.uid, inx.var, inx.iref_by_uid)

class addUnits(object):
  def __init__(self,dq,wbi='units/units.xlsx'):
    eqs = [('N m-1', 'kg s-2', 'J m-2'), ('Pa', 'N m-2', 'kg m-1 s-2'), ('cm-1','km-1','m-1')     ]
    eqss = ['N m-1','Pa','m-1','m','s','1','m s-2','m2','m3' ]

    wb = workbook( wbi )
    sh = wb.book.sheet_by_name('units')
    self.repl = {}
    self.uu = {}
    for j in range(1,sh.nrows):
      r = sh.row(j)
      if len(r) > 5 and r[5].value != u'':
        self.repl[ r[0].value ] = r[5].value
      else:
        self.uu[r[0].value] = [x.value for x in r]

    for k in self.repl:
      if self.repl[k] not in self.uu:
        print 'Bad replacement found: %s --> %s' % (k,self.repl[k])

    for i in dq.coll['var'].items:
      if i.units in self.repl:
        u = self.repl[i.units]
      else:
        u = i.units
      u = string.strip(str(u))
      if str(u) == '1.0':
        u = '1'
      if u not in self.uu:
        if u == "1.0":
          if float(u) not in self.uu:
            print 'UNITS NOT FOUND: %s (%s)' % (u,i.label)
        else:
            print 'UNITS NOT FOUND: %s (%s)' % (u,i.label)

  def uid(self,u0):
    if u0 in self.repl:
      u = self.repl[u0]
    else:
      u = u0
    if u in self.uu:
      return self.uu[u][7]
    elif u == '1.0' and float(u) in self.uu:
      return self.uu[float(u)][7]
    else:
      return None


class updates(object):
  delToks = sets.Set( ['inc','omit'] )
  def __init__(self,fndup,fnmult,idir='rev1'):
    assert os.path.isdir( idir ), 'Directory %s not found' % idir
    self.fdup = '%s/%s' % (idir,fndup)
    self.fmult = '%s/%s' % (idir,fnmult)
    for p in [self.fdup,self.fmult]:
      assert os.path.isfile( p ), 'File %s not found' % p
    self.repl = {}
    self.upd = {}
    self.twins = []
    self.ddel = {}
    self.cmvrepl = {}

  def writeCmvUpd( self, inx, fnrp='CMVreplace.csv'):
    oo = open( fnrp, 'w' )
    for k in self.cmvrepl.keys():
      if inx.iref_by_uid.has_key(k):
        kn = self.cmvrepl[k]
        for tag,ki  in inx.iref_by_uid[k]:
           vu = [ inx.uid.has_key(kk) for kk in [k,kn,ki] ]
           if all( vu ):
             oo.write( '%s\t%s\t%s\t%s\t%s\t%s\t%s\n' % (k,kn,tag,ki, inx.uid[k].label,  inx.uid[kn].label, inx.uid[ki].label) )
           else:
             print 'ERROR.088.0003: Bad index in replace info: %s .. %s .. %s' % ( str([k,kn,ki]), str(vu), tag )
      else:
        print 'ERROR.088.0004: Bad index in replace info: %s' %  k
    oo.close()

  def writeVarUpd(self, inx, fnrp='uuidreplace.csv', fnrm='uuidremove.csv', fnup='uuidupdate.csv'):
    oo = open( fnrp, 'w' )
    oo2 = open( fnrm, 'w' )
    for k in self.repl.keys():
      if inx.iref_by_uid.has_key(k):
        kn = self.repl[k]
        for tag,ki  in inx.iref_by_uid[k]:
           vu = [ inx.uid.has_key(kk) for kk in [k,kn,ki] ]
           if all( vu ):
             oo.write( '%s\t%s\t%s\t%s\t%s\t%s\t%s\n' % (k,kn,tag,ki, inx.uid[k].label,  inx.uid[kn].label, inx.uid[ki].label) )
           else:
             print 'ERROR.088.0001: Bad index in replace info: %s .. %s .. %s' % ( str([k,kn,ki]), str(vu), tag )
      else:
        oo2.write( k + '\n' )
    oo.close()
    oo2.close()
    oo = open( fnup, 'w' )
    for k in self.upd.keys():
        ee = self.upd[k]
        oo.write( string.join( [k,ee['provNote'],string.join(ee['tags']),ee['label'], ee['title'] ], '\t') + '\n' )
    oo.close()

  def scandup(self):
    ii = open( self.fdup ).readlines()
    nn = (len(ii)-1)/2
    for i in range(nn):
      l1 = string.split( ii[i*2+1], '\t' )
      l2 = string.split( ii[i*2+2], '\t' )
      xx = l1[8:10]
      yy = l2[8:10]
      if xx[1] == '' and yy[1] == xx[0]:
        ths = 0
        assert not self.repl.has_key( yy[0] ), 'duplicate replacement request for %s' % yy[0]
        self.repl[ yy[0] ] = yy[1] 
      elif yy[1] == '' and xx[1] == yy[0]:
        ths = 1
        assert not self.repl.has_key( xx[0] ), 'duplicate replacement request for %s' % xx[0]
        self.repl[ xx[0] ] = xx[1] 
      elif l1[10] == 'twin' and  l2[10] == 'twin':
        ths = 2
        self.twins.append( l1[8] )
        self.twins.append( l2[8] )
      elif l1[10] in self.delToks and l2[10] in self.delToks:
        ths = 3
        self.ddel[ l1[8] ] = (l1[10],l1[11])
        self.ddel[ l2[8] ] = (l2[10],l2[11])
      elif xx[1] == '' and yy[1] == "":
        print 'WARN.087.00001: uncorrected duplication ..... %s ' %  str( l1[:5] )
      else:
        ths = -1
        print 'ERROR.xxx.0001: Match failed'
        print l1
        print l2
        assert False

  def scancmvdup(self):
    wb = workbook( 'csv2/CMORvar.xls' )
    sht = wb.book.sheet_by_name( 'Sheet1' )
    for i in range(sht.nrows):
      rr = sht.row(i)
      if len(rr) == 21 and str( rr[20].value ) != '':
        kn = rr[20].value
        ko = rr[18].value
        vn = rr[0].value
        self.cmvrepl[ ko ] = kn
        print '%s: replace %s with %s' % (vn,ko,kn)

  def scanmult(self):
    ii = open( self.fmult ).readlines()
    nn = (len(ii)-1)/3
    for i in range(nn):
      l1 = string.split( ii[i*3+1], '\t' )
      l2 = string.split( ii[i*3+2], '\t' )
      l3 = string.split( ii[i*3+3], '\t' )
      yy = [l1[9],l2[9],l3[9]]
      xx = [l1[8],l2[8],l3[8]]
      zz = (l1,l2,l3)
      for j in range(3):
        if yy[j] != '':
          assert yy[j] in xx, 'Invalid replacement option, %s' % yy[j]
          assert  not self.repl.has_key( xx[j] ), 'duplicate replacement request for %s' % xx[j]
          self.repl[ xx[j] ] = yy[j]
        elif zz[j][10] == 'twin':
          self.twins.append( zz[j][8] )
        elif zz[j][11] == 'update':
          tags = map( string.strip, string.split( zz[j][13], ',' ) )
          self.upd[ xx[j] ] = { 'provNote':zz[j][12], 'tags':tags, 'label':zz[j][0], 'title':zz[j][1] }

###
### varDup and varMult created in first parse ----- then editted to select options
### 2nd pass through then generates the replace and remove options -- taking into account cross references
### the results of the 2nd pass go back to ../workbook to generate a new set of inputs.
###
up = updates('varDup.csv', 'varMult.csv', idir='rev2')
up.scandup()
up.scancmvdup()
up.scanmult()

urep = False
urep = True
if urep:
  up.writeVarUpd( inx, fnrp='uuidreplace.csv', fnrm='uuidremove.csv', fnup='uuidupdate.csv')
  up.writeCmvUpd( inx, fnrp='CMVreplace.csv')
else:
  oo2 = open( 'uuidremove2.csv', 'w' )
  for i in dq.coll['var'].items:
    if not inx.iref_by_uid.has_key(i.uid):
      oo2.write( string.join( [i.uid,i.label,i.title,i.prov,i.description], '\t') + '\n' )
  oo2.close()

### check back references.
nbr = 0
lbr = []
for k in inx.iref_by_uid.keys():
  if not inx.uid.has_key(k):
   nbr += 1
   lbr.append(k)
print 'Missing references: ', nbr
### can now apply mappings, create updated records and write to new xml?

for i in dq.coll['requestLink'].items:
   rql_by_name[i.label].append( i.uid )
   ix_rql_uid[i.uid] = i

for i in dq.coll['requestVarGroup'].items:
   ix_rqvg_uid[i.uid] = i


if dq.coll.has_key( 'revisedTabItem' ):
  thisk = 'revisedTabItem'
else:
  thisk = 'requestVar'
oo = open( 'uuidinsert.csv', 'w' )
for i in dq.coll[thisk].items:
  if i.uid == '__new__':
    if inx.var.label.has_key( i.label ):
      if len( inx.var.label[i.label] ) == 1:
        v = inx.uid[ inx.var.label[i.label][0] ]
        oo.write( string.join( ['unique',i.label,v.label,v.uid,v.prov,i.mip], '\t' ) + '\n' )
      else:
        oo.write( string.join( ['ambiguous',i.label,i.mip,str(len(inx.var.label[i.label] ) ) ], '\t' ) + '\n' )
oo.close()
    
oo = open( 'varMult.csv', 'w' )
oo2 = open( 'varDup.csv', 'w' )
oo3 = open( 'varStar.csv', 'w' )
hs = ['label','title','sn','units','description','prov','procnote','procComment','uid']
oo.write( string.join(hs, '\t' ) + '\n' )
oo2.write( string.join(hs, '\t' ) + '\n' )
oo3.write( string.join(hs, '\t' ) + '\n' )
ks = inx.var.label.keys()
ks.sort()
emptySet = sets.Set( ['','unset'] )
def entryEq(a,b):
  return a == b or (a in emptySet and b in emptySet)

deferredRecs = []
for k in ks:
  if len(inx.var.label[k]) == 2:
    v1 = inx.var.uid[inx.var.label[k][0]]
    v2 = inx.var.uid[inx.var.label[k][1]]
    cc = map( lambda x: entryEq( v1.__dict__[x], v2.__dict__[x]), ['title','sn','units','description']  )
    if all(cc):
### where duplicates are identical , collect and output at end of file.
      pv1 = string.find( v1.__dict__['prov'], 'OMIP.' ) != -1
      pv2 = string.find( v2.__dict__['prov'], 'OMIP.' ) != -1
      if pv2:
        vp = v2
        vo = v1
      else:
        if not pv1:
          print 'WARN.088.00002: no preference: %s, %s, %s' % (v1.__dict__['label'],v1.__dict__['prov'],v2.__dict__['prov'])
        vp = v1
        vo = v2
      deferredRecs.append( string.join(map( lambda x: vo.__dict__[x], hs) + [vp.uid,'identical'], '\t' ) + '\n' )
      deferredRecs.append( string.join(map( lambda x: vp.__dict__[x], hs) + ['',''], '\t' ) + '\n' )
    else:
      try:
        oo2.write( string.join(map( lambda x: str(v1.__dict__[x]), hs) + ['',''], '\t' ) + '\n' )
        oo2.write( string.join(map( lambda x: str(v2.__dict__[x]), hs) + ['',''], '\t' ) + '\n' )
      except:
        print 'SEVERE.oo2.00001',v1.__dict__
        print 'SEVERE.oo2.00002',v2.__dict__
      
  elif len(inx.var.label[k]) > 1:
    for i in inx.var.label[k]:
      oo.write( string.join(map( lambda x: inx.var.uid[i].__dict__[x], hs), '\t' ) + '\n' )

  if k[-2:] == '--':
    for i in (inx.var.label[k] + inx.var.label[k[:-2]]):
      oo3.write( string.join(map( lambda x: inx.var.uid[i].__dict__[x], hs), '\t' ) + '\n' )

## output auto-filled records for identical duplicates at end of varDup file.
for r in deferredRecs:
  oo2.write( r )
oo.close()
oo2.close()
oo3.close()



vns = inx.var.label.keys()
vns.sort()
for v in vns:
  if len( inx.var.label[v] ) > 1:
     print 'INFO.001.0001:',v, string.join( map( lambda x: inx.var.uid[x].sn, inx.var.label[v] ), ';' )

nok = 0
nerr = 0
if dq.coll.has_key( 'ovar' ):
  thisk = 'ovar'
else:
  thisk = 'CMORvar'
for i in dq.coll[thisk].items:
   vid = i.vid
   ix_ovar_uid[i.uid] = i
   xr_var_ovar[vid].append( i.uid )
   if not inx.var.uid.has_key(vid):
     print 'missing key:',i.label, i.prov, vid
     nerr += 1
   else:
     nok += 1

class rqHtml(object):

  def __init__(self,odir='./htmlSc/'):
    self.odir = odir
    if not os.path.isdir(odir):
      os.mkdir(odir)

  def mkRqlHtml(self,name):
     ## [u'comment', u'uid', u'tab', u'title', u'label', u'grid', 'defaults', u'objective', u'mip', 'globalDefault', u'gridreq']
    if len( rql_by_name[name] ) == 1:
      self.mkRqlHtml01(rql_by_name[name][0], name )
    else:
      self.mkRqlHtmlGp(name)

  def mkRqlHtmlGp(self,name):
    ee = {}
    ee['title'] = 'CMIP Request Link %s (with multiple definitions)' % name
    self.pageName = 'rql__%s.html' % name
    al =[]
    for i in range( len( rql_by_name[name] ) ):
      this = ix_rql_uid[rql_by_name[name][i]]
      al.append( tmpl.item % {'item':'<a href="rql__%s__%s.html">[%s]</a>: %s' % (name,i,i,this.title) } )
    ee['items'] = string.join(al, '\n' )
    ee['introduction'] = ''
    ee['htmlBody'] = tmpl.indexWrapper % ee
    ee['htmlHead'] = '''<title>%(title)s</title>''' % ee
    self.pageHtml = tmpl.pageWrapper % ee
    self.write()
    for i in range( len( rql_by_name[name] ) ):
      self.mkRqlHtml01(rql_by_name[name][i],i)

  def mkRqlHtml01(self,id, tag):
    this = ix_rql_uid[id]
    ee = {}
    if this.label == tag:
      ee['title'] = 'CMIP Request Link %s' % tag
      self.pageName = 'rql__%s.html' % tag
    else:
      ee['title'] = 'CMIP Request Link %s[%s]' % (this.label,tag)
      self.pageName = 'rql__%s__%s.html' % (this.label,tag)
    atts = this.__dict__.keys()
    atts.sort()
    al =[]
    for a in atts:
      if a not in ['defaults','globalDefault']:
        al.append( tmpl.item % {'item':'<b>%s</b>: %s' % (a,this.__dict__.get(a,'-- Not Set --')) } )
    ee['items'] = string.join(al, '\n' )
    ee['introduction'] = ''
    ee['htmlBody'] = tmpl.indexWrapper % ee
    ee['htmlHead'] = '''<title>%(title)s</title>''' % ee
    self.pageHtml = tmpl.pageWrapper % ee
    self.write()
     
  def mkVarHtml(self,name):
    if len( inx.var.label[name] ) == 1:
      self.mkVarHtml01(inx.var.label[name][0], name )
    else:
      self.mkVarHtmlGp(name)

  def mkVarHtmlGp(self,name):
    ee = {}
    ee['title'] = 'CMIP Variable %s (with multiple definitions)' % name
    self.pageName = 'var__%s.html' % name
    al =[]
    for i in range( len( inx.var.label[name] ) ):
      this = inx.var.uid[inx.var.label[name][i]]
      al.append( tmpl.item % {'item':'<a href="var__%s__%s.html">[%s]</a>: %s' % (name,i,i,this.title) } )
    ee['items'] = string.join(al, '\n' )
    ee['introduction'] = ''
    ee['htmlBody'] = tmpl.indexWrapper % ee
    ee['htmlHead'] = '''<title>%(title)s</title>''' % ee
    self.pageHtml = tmpl.pageWrapper % ee
    self.write()
    ##print 'Multi var: %s' % name
    for i in range( len( inx.var.label[name] ) ):
      self.mkVarHtml01(inx.var.label[name][i],i)

  def mkVarHtml01(self,id, tag):
    this = inx.var.uid[id]
    ee = {}
    if this.label == tag:
      ee['title'] = 'CMIP Variable %s' % tag
      self.pageName = 'var__%s.html' % tag
    else:
      ee['title'] = 'CMIP Variable %s[%s]' % (this.label,tag)
      self.pageName = 'var__%s__%s.html' % (this.label,tag)
    atts = this.__dict__.keys()
    atts.sort()
    al =[]
    for a in atts:
      if a not in ['defaults','globalDefault']:
        al.append( tmpl.item % {'item':'<b>%s</b>: %s' % (a,this.__dict__.get(a,'-- Not Set --')) } )

    if inx.iref_by_uid.has_key(this.uid):
      assert varRefs.has_key(this.uid), 'Problem with collected references'
      ee1 = varRefs[this.uid]
      ks = ee1.keys()
      ks.sort()
      for k in ks:
        al.append( tmpl.item % {'item':'<b>%s</b>: %s' % (k,string.join(ee1[k])) } )
    ee['items'] = string.join(al, '\n' )
    ee['introduction'] = ''
    ee['htmlBody'] = tmpl.indexWrapper % ee
    ee['htmlHead'] = '''<title>%(title)s</title>''' % ee
    self.pageHtml = tmpl.pageWrapper % ee
    self.write()

  def varHtml(self):
    for k in inx.var.label.keys():
      self.mkVarHtml(k)
  
  def rqlHtml(self):
    for k in rql_by_name.keys():
      self.mkRqlHtml(k)
  
  def write(self):
    oo = open( '%s/%s' % (self.odir,self.pageName), 'w' )
    oo.write( self.pageHtml )
    oo.close()

    
vh = rqHtml()
vh.varHtml()
vh.rqlHtml()

if nerr == 0:
  print 'CHECK 001: %s records checked, no missing references' % nok

shps = {'': 64, 'XYZKT': 13, '4-element vector': 2, 'XYT': 476, '2D vector field ': 2, 'KZT': 4, '2D vector field': 2, 'XYZ': 27, 'XYZT': 204, '2D': 83, 'scalar': 14, 'XY': 88, '?': 21, '2D ': 1, 'XYKT': 3, 'YZT': 16, 'ZST1': 15, 'XKT': 2, 'BasinYT': 1}
vshpchkMap = {'':'', u'all model levels above 400hPa':'alevStrat', u'all':'Xlev', 3.0:'plev3', '4.0':'plev4', \
            36.0:'plev36', u'soil levels':'sdepth', \
            1.0:'sfc?', \
           16.0:'plev16', 7.0:'plev7', 40.0:'plev40', u'all*':'Xlev', 14.0:'plev14', u'Model levels or 27Plevs':'alev|plev27', \
           u'17 (or 23 )':'plev17|plev23', u'17 (or 23)':'plev17|plev23', \
           27.0:'plev27', 17.0:'plev17', u'17 (or23)':'plev17|plev23', 8.0:'plev8', u'all model levels':'alev', 5.0:'plev5'}
ks = vshpchkMap.keys()
for k in ks:
  if type(k) == type(0.):
     vshpchkMap[str(k)] = vshpchkMap[k]

tsmap = { 'mean':[u'daily mean', u'time mean', u'time: day',
                u'Cumulative annual fraction', u'Time mean', u'weighted time mean', u'time: mean', u'mean', u'Mean'],
          '__unknown__':['','dummyAt'],
          'point':[ u'Instantaneous (end of year)', u'point', u'Synoptic', u'instantaneous', u'time: point', u'synoptic'] }
tsmap2 = {}
for k in tsmap.keys():
  for i in tsmap[k]:
    tsmap2[i] = k

if dq.coll.has_key( 'groupItem' ):
  ee = collections.defaultdict( int )
  for i in dq.coll['groupItem'].items:
    tst = tsmap2[ i.tstyle ]
    dd = ''
    if 'X' in i.shape:
      dd += 'latitude '
    if 'Y' in i.shape:
      dd += 'longitude '
    if 'Z' in i.shape:
      if i.levels == '':
        print 'ERROR.001.0001: no levels specified', i.label, i.title
      else:
        zdim = vshpchkMap[i.levels]
        dd +=  zdim 
  ## print '%s::%s::%s|%s' % (i.shape, i.levels, i.tstyle, dd)
  for i in dq.coll['groupItem'].items:
     list_gp_ovar[i.gpid].append( i.uid )

  nok = 0
  nerr = 0
  for i in dq.coll['groupItem'].items:
     vid = i.vid
     ix_gpi_uid[i.uid] = i
     xr_var_gpi[vid].append( i.uid )
     if not inx.var.uid.has_key(vid):
       nerr += 1
     else:
       nok += 1
  ##print 'groupItem to var crossref: nok = %s, nerr = %s' % (nok, nerr)


class tcmp(object):
  def __init__(self):
    pass
  def cmp(self,x,y):
    return cmp(x.title,y.title)

def atRepr(l,x,optional=False):
  if x != None:
    if optional:
      v = l.__dict__.get(x, '' )
    else:
      v = l.__dict__[x]
  else:
    v = l
  if v == '__unset__':
    return ''
  elif type(v) in [ type([]), type(())]:
    return string.join([str(i) for i in v])
  else:
    return v

class xlsx(object):
  def __init__(self,fn):
    self.wb = xlsxwriter.Workbook(fn)

  def newSheet(self,name):
    self.worksheet = self.wb.add_worksheet(name=name)
    return self.worksheet

  def close(self):
    self.wb.close()
  
def dumpxlsx( fn, key, atl):
  wb = xlsx( fn )
  sht = wb.newSheet( key )
  j = 0
  for i in range(len(atl)):
     sht.write( j,i, atl[i] )
  ll = dq.coll[key].items[:]
  ll.sort( tcmp().cmp )
  for l in ll:
    uid = atRepr(l,'uid')
    j+=1
    i=0
    for x in atl:
       sht.write( j,i, atRepr(l,x) )
       i+=1
    if key == 'var' and refpix != None:
      if uid in refpix:
        p = refpix[uid][2]
      else:
        p = 201
      sht.write( j,i, p )
  wb.close()
      
def dumpcsv( fn, key, atl, optionalSet='' ):
  oo = open(fn, 'w' )
  ll = dq.coll[key].items[:]
  ll.sort( tcmp().cmp )
  oo.write( string.join( atl, '\t' ) + '\n' )
  for l in ll:
    try:
      oo.write( string.join( map( lambda x: str(atRepr(l,x,x in optionalSet)), atl), '\t' ) + '\n' )
    except:
      print 'SEVERE.090.0001: print %s' % str(atl)
      print l
      print key
      raise
  oo.close()

def atlSort( ll ):
  oo = []
  l1 = ['label','title']
  l2 = ['uid','defaults','globalDefault']
  for i in l1:
    if i in ll:
      oo.append(i)
  ll.sort()
  for i in ll:
    if i not in l1 + l2:
      oo.append(i)
  if 'uid' in ll:
    oo.append( 'uid' )
  return oo

for k in dq.coll.keys():
  if len( dq.coll[k].items ) > 0:
    expl = dq.coll[k].items[0]
    atl = atlSort( expl.__dict__.keys() )
    atl1 = [a for a in atl if a != 'parent' and a[0] != '_']
    ###print k, atl1
    optionalSet = set( [a for a in atl1 if not expl.__class__.__dict__[a].required] )
    dumpcsv( 'csv2/%s.csv' % k, k, atl1, optionalSet=optionalSet )
    if k == 'var':
       dumpxlsx( 'csv2/var.xlsx', k, atl1 )
  
class annotate(object):
  def __init__(self,src,dreq):
    assert os.path.isfile( src), '%s not found' % src 
    self.doc = xml.dom.minidom.parse( src  )
    self.dreq = dreq
    self.removedUids = {}

  def iniVar(self,dq):
    a = addUnits(dq)
    this = self.doc.getElementsByTagName('var')[0]
    dil = this.getElementsByTagName('item')
    self.vid = {}
    for item in dil:
      uid = item.getAttribute( 'uid' )  
      title = item.getAttribute( 'title' )  
      label = item.getAttribute( 'label' )  
      self.vid[uid] = (label,title)
      units = item.getAttribute( 'units' )  
      if units in a.repl:
        print 'INFO.units.0001: replacing %s --> %s' % (units,a.repl[ units ])
        units = a.repl[ units ]
      u2 = a.uid( units )
      if u2 == None:
        units = string.strip( str(units) )
        u2 = a.uid( units )
      if u2 != None:
        item.setAttribute( 'unid', u2 )
      else:
        print 'Units not recognised: %s (%s) -- %s' % (units,label, type(units))
     
  def tableMap(self,dq):
    that = self.doc.getElementsByTagName('CMORvar')[0]
    dil = that.getElementsByTagName('item')
    ee1 = {}
    for item in dil:
      mt = item.getAttribute( 'mipTable' )
      if mt in mipTableMap:
        mt1 = mipTableMap[mt]
        item.setAttribute( 'mipTable', mt1 )
        item.setAttribute( 'mtid', 'MIPtable::%s' % mt1 )
      
  def strTtl(self,dq):
    this = self.doc.getElementsByTagName('cellMethods')[0]
    dil = this.getElementsByTagName('item')
    eecm = {}
    for item in dil:
      eecm[item.getAttribute( 'uid' )] = item.getAttribute( 'label' )

    that = self.doc.getElementsByTagName('spatialShape')[0]
    dil = that.getElementsByTagName('item')
    ee1 = {}
    for item in dil:
      uid = item.getAttribute( 'uid' )
      lab = item.getAttribute( 'label' )
      ttl = item.getAttribute( 'title' )
      ee1[uid] = (lab,ttl)

    this = self.doc.getElementsByTagName('structure')[0]
    dil = this.getElementsByTagName('item')
    
    filterStr = True
    estr = []
    strForce = ['str-a076','str-x269','str-x100']
    for item in dil:
      uid = item.getAttribute( 'uid' )  
      lab = item.getAttribute( 'label' )  
      if len( dq.inx.iref_by_uid[uid] ) == 0 and filterStr:
         if lab not in strForce:
           print 'UNUSED STRUCTURE: %s, %s' % (lab,uid)
           this.removeChild( item )
           self.removedUids[uid] = 'structure: %s' % lab
      else:
        tmid = item.getAttribute( 'tmid' )  
        spid = item.getAttribute( 'spid' )  
        cmid = item.getAttribute( 'cmid' )  
        cml = eecm.get( cmid, '' )
        if cml != '':
          cml = ' [%s]' % cml
        o = item.getAttribute( 'odims' )  
        c = item.getAttribute( 'coords' )  
        if spid not in ee1:
          print 'SEVERE:spid.0001: spid not found: %s' % spid 
          sl = '__unknowm__'
          st = '__unknowm__'
        else:
          sl,st = ee1[spid]
        if tmid not in dq.inx.uid:
          print 'BAD time record uid: ',tmid
          print lab, uid
          raise
        title = '%s, %s [%s]' % (dq.inx.uid[tmid].title, st, sl) 
        if string.strip( c ) != '' or string.strip( o ) != '':
          title += ' {%s:%s}' % (o,c)
        title += cml
        if title != item.getAttribute( 'title' ):
           print 'STRUCT: %s:: %s' % (title, item.getAttribute( 'title' ) )
           if item.getAttribute( 'title' )[:4] == 'str-':
             item.setAttribute( 'title', title )
             estr.append( item.getAttribute( 'label' ) )
    if len(estr) > 0:
      oo = open( 'scandreq_estr.txt', 'w' )
      for i in estr:
        oo.write( i + '\n' )
      oo.close()
    
  def rvgCheck(self,dq):
    """Remove request variable groups which have no requestLink"""
    this = self.doc.getElementsByTagName('requestVarGroup')[0]
    dil = this.getElementsByTagName('item')
    nn = 0
    for item in dil:
        uid = item.getAttribute( 'uid' )  
        if ('requestLink' not in dq.inx.iref_by_sect[uid].a) and ('tableSection' not in dq.inx.iref_by_sect[uid].a):
          if item.getAttribute( 'label' ) in ['aermonthly']:
             print 'WARN.0010: overriding variable group pruning: ',item.getAttribute( 'label' )
          else:
            ##self.removedUids[ uid ] = 'requestVarGroup:  %s' % item.getAttribute( 'label' )
            ##this.removeChild(item)
            print 'INFO.rvg.0001: New rvg?? ',uid, item.getAttribute( 'label' )
            nn+=1
    ##print 'WARN.Unused variable groups removed: %s' % nn
    this = self.doc.getElementsByTagName('requestVar')[0]
    dil = this.getElementsByTagName('item')
    nn = 0
    s1 = {i.uid for i in dq.coll['requestVarGroup'].items if i.uid not in self.removedUids}
    self.usedCmv = set()
    for item in dil:
        uid = item.getAttribute( 'uid' )  
        vid = item.getAttribute( 'vid' )
        vgid = item.getAttribute( 'vgid' )
        if vgid not in s1:
          this.removeChild(item)
          self.removedUids[ uid ] = 'requestVar:  %s' % item.getAttribute( 'label' )
          nn+=1
        else:
          self.usedCmv.add( vid )
    print 'Unused request variables removed: %s' % nn
        
  def rqvCheck(self,dq):
    wb = workbook( 'ingest/rqv_cmv_remap.xls' )
    rqvredirect = {('aermonthly','siconc'):'SImon' }
    sh = wb.book.sheet_by_name('maps')
    self.repl = {}
    self.uu = {}
    for j in range(sh.nrows):
      r = [str(x.value) for x in sh.row(j)]
      rqvredirect[ ( r[0], r[1] ) ] = r[2]

    this = self.doc.getElementsByTagName('requestVar')[0]
    dil = this.getElementsByTagName('item')
    badrqv = set()
    self.removedCmv = set()
    for item in dil:
        uid = item.getAttribute( 'uid' )  
        vid = item.getAttribute( 'vid' )  
        if vid not in dq.inx.uid:
          badrqv.add( uid )
        elif dq.inx.uid[vid]._h.label == 'remarks':
          badrqv.add( uid )
        else:
          cmv = (dq.inx.uid[vid].mipTable,dq.inx.uid[vid].label)
          if cmv in rqvredirect:
            targ = (rqvredirect[cmv],dq.inx.uid[vid].label)
            if targ not in self.cmvLookUp:
              print 'SEVERE.rgvrdi.00001: attempt to redirect to non-existant var: %s --> %s' % (str(cmv),str(targ))
            else:
               nid = self.cmvLookUp[targ]
               item.setAttribute( 'vid', nid )
               self.removedCmv.add( vid )
               print 'INFO.rgvrdi.00002: redirect rql to new CMOR var: %s --> %s' % (str(cmv),str(targ))
    
  def cmvCheck2(self,dq):
    this = self.doc.getElementsByTagName('CMORvar')[0]
    dil = this.getElementsByTagName('item')
    self.usedVar = set()

    self.keepVar = {'ugrido',}
    for item in dil:
        uid = item.getAttribute( 'uid' )  
        if ((uid not in self.usedCmv) or (uid in self.removedCmv)) and (item.getAttribute( 'label' ) not in self.keepVar):
          print 'INFO.cmv.04004: removing unused CMORvar: %s, %s, %s, %s: ' % (uid,item.getAttribute( 'label' ),item.getAttribute( 'table' ),item.getAttribute( 'mipTable' ))
          this.removeChild(item)
          self.removedUids[ uid ] = 'CMORvar:  %s' % item.getAttribute( 'label' )
        else:
          self.usedVar.add( item.getAttribute( 'vid' ) )

    this = self.doc.getElementsByTagName('var')[0]
    dil = this.getElementsByTagName('item')
    self.varLookUp = {}
    for item in dil:
        uid = item.getAttribute( 'uid' )  
        self.varLookUp[ item.getAttribute( 'label' ) ] = uid
        if uid not in self.usedVar and item.getAttribute( 'label' ) not in self.keepVar:
          print 'INFO.var.04004: removing unused var: %s, %s, %s, %s: ' % (uid,item.getAttribute( 'label' ),item.getAttribute( 'table' ),item.getAttribute( 'units' ))
          this.removeChild(item)
          self.removedUids[ uid ] = 'var:  %s' % item.getAttribute( 'label' )

    this = self.doc.getElementsByTagName('CMORvar')[0]
    wb = workbook( 'ingest/extraCmv.xls' )
    sh = wb.book.sheet_by_name('extra')
    extra = []
    self.uu = {}
    for j in range(1,sh.nrows):
      extra.append( [str(x.value) for x in sh.row(j)] + ['',''] )
    extra_old = [('E6hrZ','ps','str-a076','ps','atmos','Surface Pressure','Surface Pressure .. needed for vertical coordinates','6hr', 'CMIP extra','scanDreq.py','1','float',''),
             ('AERmon','ps','str-013','ps','atmos','Surface Pressure','Surface Pressure .. needed for vertical coordinates','mon', 'CMIP extra','scanDreq.py','1','float',''),
             ('E3hrPt','ps','str-d11','ps','atmos','Surface Pressure','Surface Pressure .. needed for vertical coordinates','3hr', 'CMIP extra','scanDreq.py','1','float',''),]
    eh = ['label','modeling_realm','title','description','frequency','provNote','prov','defaultPriority','type','positive']
    hids = set()
    idfp = eh.index('defaultPriority') + 3
    for e in extra:
     assert e[11] in ['float'], 'bad type in cmv extra ... see ingest/extraCmv.xls: %s' % e[11] 
     if (e[0],e[1]) not in self.cmvLookUp:
       e[idfp] = str( int( float( e[idfp] ) ) )
       stid = self.strLookUp[e[2]]
       vid = self.varLookUp[e[1]]
       new = self.doc.createElement( 'item' )
       new.setAttribute( 'mipTable', e[0] )
       new.setAttribute( 'mtid', 'MIPtable::%s' % e[0] )
       new.setAttribute( 'stid', stid )
       new.setAttribute( 'vid', vid )
       new.setAttribute( 'rowIndex', '0' )
       for a in ['deflate','deflate_level','shuffle']:
          new.setAttribute( a, '' )
       hid = hashlib.new( 'sha1', 'extra-fields::' + ':'.join(e[:8]) ).hexdigest()
       assert hid not in hids
       new.setAttribute( 'uid', hid )
       print 'INFO.extra.00011: ',e

       for k in range(len(eh) ):
         new.setAttribute( eh[k], e[3+k] )
       this.appendChild( new )
     else:
       print 'SKIPPING extraCMV: %s.%s' % (e[0],e[1])
         
      

  def cmvCheck(self,dq):
    this = self.doc.getElementsByTagName('CMORvar')[0]
    dil = this.getElementsByTagName('item')
    kk = 0
    kka = 0
    nrm0 = 0
    nrm1 = 0
    self.cmvLookUp = {}
    for item in dil:
      title = item.getAttribute( 'title' )  
      if title[:6] == '__from':
        kka += 1
        vid = item.getAttribute( 'vid' )  
        if vid in self.vid:
          title2 = self.vid[vid][1]
          item.setAttribute( 'title', title2 )
          kk += 1
      realm = item.getAttribute( 'modeling_realm' )
      if realm in ['','?']:
        stid = item.getAttribute( 'stid' )
        st = dq.inx.uid[stid]
        odims = st.odims
        this = None
        if odims == 'iceband':
          this = 'seaIce'
        elif odims not in ['','?']:
          this = 'atmos'
        else:
          sp = dq.inx.uid[st.spid]
          if sp.label in ['TR-na','XY-O', 'YB-O', 'YB-R']:
            this = 'ocean'
          elif sp.label != 'XY-na':
            this = 'atmos'
          else:
            prov = item.getAttribute( 'prov' )
            if prov in knowna:
              this = 'atmos'
            elif prov in knowno:
              this = 'ocean'
            elif prov in knownl:
              this = 'land'
            else:
              lab  = item.getAttribute( 'label' )
              print 'ERROR.cmv.00006: no realm: %s, %s, %s, %s ..' % (lab,odims, sp.label, prov)
        if this == None:
          nrm1 += 1
        else:
          nrm0 += 1
          item.setAttribute( 'modeling_realm', this )
      self.cmvLookUp[(item.getAttribute( 'mipTable' ),item.getAttribute( 'label' ) )] = item.getAttribute( 'uid' )
    print ('CMOR Var realms set: %s' % nrm0 )
    if nrm1 > 0:
      print ('SEVERE.cmv.00005: realm unset: %s' % nrm1)
    print ('CMOR Var titles reset: %s [%s]' % (kk,kka))
    
  def mipProv(self,dq):
    s1 = re.compile( '\[([a-zA-Z0-9]*)\]' )
    cc = collections.defaultdict(list)
    dd = collections.defaultdict(int)
    mips = set()
    for i in dq.coll['mip'].items:
      mips.add( i.uid )
    for i in dq.coll['var'].items:
      cc[i.prov].append( i.label )

    ee = {}
    for i in sorted( cc.keys() ):
      if i[:9] == 'CMIP6 end':
        m = s1.findall( i )
        assert len( m ) == 1, 'FAILED TO PARSE: %s' % i
        this = m[0]
      else:
        i5 = i.find( 'CMIP5' ) != -1
        io = i.find( 'OMIP' ) != -1
        icx = i.find( 'CORDEX' ) != -1
        ip = i.find( 'PMIP' ) != -1
        icc = i.find( 'CCMI' ) != -1
        isp = i.find( 'SPECS' ) != -1
        icf = i.find( 'CFMIP' ) != -1
        iac = i.find( 'AerChemMIP' ) != -1
        if i5 and io:
          print 'WARNING .. unclear provenance: ',i,cc[i]
          this = 'CMIP5/OMIP'
        elif i5:
          this = 'CMIP5'
        elif io:
          this = 'OMIP'
        elif icx:
          this = 'CORDEX'
        elif ip:
          this = 'PMIP'
        elif icc:
          this = 'CCMI'
        elif isp:
          this = 'SPECS'
        elif icf:
          this = 'CFMIP'
        elif iac:
          this = 'AerChemMIP'
        else:
          print 'WARNING .. unclear provenance [2]: ',i,cc[i]
          this = 'unknown'
    
      ee[i] = this
      dd[this] += len( cc[i] )
    self.dd = dd
    self.ee = ee
    this = self.doc.getElementsByTagName('var')[0]
    dil = this.getElementsByTagName('item')
    print 'FIXING var provmip attribute, %s items' % len(dil)
    kk = 0
    for item in dil:
      kk += 1
      p = item.getAttribute( 'prov' )  
      p0 = item.getAttribute( 'provmip' )  
      if p0 not in mips:
        if p in mips:
          item.setAttribute( 'provmip', p )
        else:
          assert ee.has_key(p), 'Unmatched key: %s' % p
          assert ee[p] in mips, 'Unknown provenance: %s, %s' % (p,ee[p])
          item.setAttribute( 'provmip', ee[p] )
      
  def fixCellMethods(self,dq):
    this = self.doc.getElementsByTagName('structure')[0]
    dil = this.getElementsByTagName('item')
    cmrep = collections.defaultdict( set )
    cmc = collections.defaultdict( int )
    self.strLookUp = {}
    nrep = 0
    for item in dil:
      self.strLookUp[ item.getAttribute( 'label' ) ] = item.getAttribute( 'uid' )
      cm = item.getAttribute( 'cell_methods' )  
      if string.find( cm, "area: where" ) != -1:
        cm1 = string.replace( cm, "area: where", "area: mean where" )
        item.setAttribute( 'cell_methods', cm1 )  
        cmrep[cm].add(cm1)
        cmc[cm1] += 1
        nrep += 1
      elif string.find( cm, "time:mean" ) != -1:
        cm1 = string.replace( cm, "time:mean", "time: mean" )
        item.setAttribute( 'cell_methods', cm1 )  
        cmrep[cm].add(cm1)
        cmc[cm1] += 1
        nrep += 1
      elif string.find( cm, "weighted b " ) != -1:
        cm1 = string.replace( cm, "weighted b ", "weighted by " )
        item.setAttribute( 'cell_methods', cm1 )  
        cmrep[cm].add(cm1)
        cmc[cm1] += 1
        nrep += 1
    print ('FIXED CELL METHODS .. %s' % nrep )
    for k in cmc:
      print ('%s: %s' % (k,cmc[k]) )


##
## defective code .. can not easily do structure mapping here .. need to think about that ...
##
  def sectionCopy(self,dq):
    this = self.doc.getElementsByTagName('CMORvar')[0]
    thisRqv = self.doc.getElementsByTagName('requestVar')[0]
    xx = [i for i in dq.coll['requestVarGroup'].items if i.label == 'OMIP-Omon']
    assert len(xx) == 1, 'OMIP-Omon request variable group not found'
    omipOmonUid = xx[0].uid
    dil = this.getElementsByTagName('item')
    for item in dil:
      mipt = item.getAttribute( 'mipTable' )  
      prov = item.getAttribute( 'prov' )  
      provn = item.getAttribute( 'provNote' )  
      if mipt == 'Oyr' and prov[:12] == "CMIP6 [OMIP]" and provn == 'bgc':
        rowix = int( item.getAttribute( 'rowIndex' ) )
        if rowix < 65:
          var = item.getAttribute( 'label' )
          new = item.cloneNode(True)
          new.setAttribute( 'defaultPriority', '2' )
          new.setAttribute( 'mipTable', 'Omon' )
          new.setAttribute( 'prov', 'Copy from Oyr' )
          new.setAttribute( 'provNote', 'sdq.001' )
          vid =  str( uuid.uuid1() ) 
          new.setAttribute( 'uid', vid )
          this.appendChild( new )
##
## create request var
##
          new2 = self.doc.createElement( 'item' )
          uid =  str( uuid.uuid1() ) 
          new2.setAttribute( 'uid', uid )
          new2.setAttribute( 'priority', '2' )
          new2.setAttribute( 'vid', vid )
          new2.setAttribute( 'vgid', omipOmonUid )
          new2.setAttribute( 'mip', 'OMIP' )
          new2.setAttribute( 'table', 'OMIP-Omon' )

          if omipOmonUid not in dq.inx.uid:
            print 'ERROR.005.0001: vgid %s not found' % omipOmonUid
          
          thisRqv.appendChild(new2)

  def missingRefs(self,mrefs,dq,clear=True):
    this = self.doc.getElementsByTagName('remarks')[0]
    if clear:
      dil = this.getElementsByTagName('item')
      for d in dil:
        this.removeChild(d)
    for k in mrefs.keys():
      if len(  mrefs[k] ) == 1:
        tid = mrefs[k][0][2]
        tattr = mrefs[k][0][1]
        tn = None
      else:
        tid = None
        ee = collections.defaultdict(int)
        tn = str( len( mrefs[k] ) )
        for t in mrefs[k]:
          s = self.dreq.inx.uid[t[2]]._h.label
          ee['%s.%s' % (s,t[1])] += 1
        if len( ee.keys() ) == 1:
          tattr = ee.keys()[0]
        else:
          tattr = '__multiple__'
      if tid == None or (tid not in self.removedUids):
        item = self.doc.createElement( 'item' )
        assert type(k) == type( '' ), 'Attempt to set uid with bad type: %s' % str(k)
        item.setAttribute( 'uid', k )  
        item.setAttribute( 'tattr', tattr )  
        if tn != None:
          item.setAttribute( 'techNote', tn )  
        if tid != None:
          item.setAttribute( 'tid', tid )  
          if tid not in dq.inx.uid:
            print 'ERROR.005.0002: tid %s not found' % tid
        item.setAttribute( 'class', 'missingLink' )  
        item.setAttribute( 'description', 'Missing links detected and marked for fixing' )  
        item.setAttribute( 'prov', 'scanDreq.py:annotate' )  
        this.appendChild( item )
  
    parent = self.doc.getElementsByTagName('annex')[0]
    for this in parent.childNodes:
      if this.nodeType == this.ELEMENT_NODE:
        dil = this.getElementsByTagName('item')
        print 'INFO.nodescan.00001: ',this.localName,len(dil)
        for item in dil:
          for k in item.attributes.keys():
            v = item.getAttribute( k )
            if type( v ) not in [type( '' ),type( u'' )]:
              print 'SEVERE.0001: tuple in attribute value',this.localName,k,v
    txt = self.doc.toprettyxml(indent='\t', newl='\n', encoding=None)
    oo = open( 'out/annotated_20150731_i1.xml', 'w' )
    lines = string.split( txt, '\n' )
    for line in lines:
      l = utils_wb.uniCleanFunc( string.strip(line) )
      if empty.match(l):
        continue
      else:
        oo.write(l + '\n')
    oo.close()

def anno():
  oo = open( 'var1.csv', 'w' )
  ks = ['label','title','sn','units','description','prov','procnote','procComment','uid']
  ks2 = [ 'ovar','groupItem','revisedTabItem']
  oo.write( string.join(ks + ks2, '\t' ) + '\n' )
  for i in dq.coll['var'].items:
     if i.label[-2:] != '--':
       ee1 = varRefs.get( i.uid, {} )
       r2 = map( lambda x: string.join( atRepr( ee1.get(x, [] ), None ) ), ks2 )
       oo.write( string.join(map( lambda x: atRepr(i,x), ks) + r2, '\t' ) + '\n' )
  oo.close()

  print 'ANNOTATING: ', dq.c.docl[0][0]
  an = annotate( dq.c.docl[0][0], dq )
###############################################################################
### mode 3: using exports/a1/ ... 
### should add structures to exports/a1/ ... i.e. structures + spatial and temporal shapes + CMOR dimensions.
###
  ###an.sectionCopy(dq)
  an.iniVar( dq )
  an.fixCellMethods(dq)
  an.mipProv(dq)
  an.cmvCheck(dq)
  an.rvgCheck(dq)
  an.strTtl(dq)
  an.rqvCheck(dq)
  an.cmvCheck2(dq)
  for k in an.removedUids:
    print 'WARN.REMOVED: %s: %s' % (k,an.removedUids[k])
  an.missingRefs( dq.inx.missingIds, dq )

if __name__ == '__main__':
  anno()