def selectEmFiles(**kwargs):
# Pour A et L : mon, day1, day2, 6hLev, 6hPlev, 3h
simulation = kwargs['simulation']
frequency = kwargs['frequency']
variable = kwargs['variable']
period = kwargs['period']
realm = kwargs['realm']
#
freqs = {"mon": "", "3h": "_3h"}
f = frequency
if f in freqs:
f = freqs[f]
rep = []
# Must look for all realms, here identified by a single letter
if realm == "*":
lrealm = ["A", "L", "O", "I"]
else:
lrealm = [realm]
for realm in lrealm:
clogger.debug("Looking for realm " + realm)
# Use EM data for finding data dir
freq_for_em = f
if realm == 'I':
freq_for_em = "" # This is a special case ...
command = [
"grep", "^export EM_DIRECTORY_" + realm + freq_for_em + "=",
os.path.expanduser(os.getenv("EM_HOME")) + "/expe_" + simulation
]
try:
ex = subprocess.Popen(command,
stdout=subprocess.PIPE,
stderr=subprocess.PIPE)
except:
clogger.error("Issue getting archive_location for " + simulation +
" for realm " + realm + " with: " + repr(command))
break
if ex.wait() == 0:
dir = ex.stdout.read().split("=")[1].replace('"',
"").replace("\n", "")
clogger.debug("Looking at dir " + dir)
if os.path.exists(dir):
lfiles = os.listdir(dir)
for fil in lfiles:
# clogger.debug("Looking at file "+fil)
fileperiod = periodOfEmFile(fil, realm, f)
if fileperiod and period.intersects(fileperiod):
if fileHasVar(dir + "/" + fil, variable):
rep.append(dir + "/" + fil)
# clogger.debug("Done with Looking at file "+fil)
else:
clogger.error(
"Directory %s does not exist for simulation %s, realm %s "
"and frequency %s" % (dir, simulation, realm, f))
else:
clogger.info("No archive location found for " + simulation +
" for realm " + realm + " with: " + repr(command))
return rep
def selectEmFiles(**kwargs) :
#POur A et L : mon, day1, day2, 6hLev, 6hPlev, 3h
simulation=kwargs['simulation']
frequency=kwargs['frequency']
variable=kwargs['variable']
period=kwargs['period']
realm=kwargs['realm']
#
freqs={ "mon" : "" , "3h" : "_3h"}
f=frequency
if f in freqs : f=freqs[f]
rep=[]
# Must look for all realms, here identified by a single letter
if realm=="*" : lrealm= ["A", "L", "O", "I" ]
else: lrealm=[ realm ]
for realm in lrealm :
clogger.debug("Looking for realm "+realm)
# Use EM data for finding data dir
freq_for_em=f
if realm == 'I' : freq_for_em="" # This is a special case ...
command=["grep", "^export EM_DIRECTORY_"+realm+freq_for_em+"=",
os.path.expanduser(os.getenv("EM_HOME"))+"/expe_"+simulation ]
try :
ex = subprocess.Popen(command, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
except :
clogger.error("Issue getting archive_location for "+
simulation+" for realm "+realm+" with: "+`command`)
break
if ex.wait()==0 :
dir=ex.stdout.read().split("=")[1].replace('"',"").replace("\n","")
clogger.debug("Looking at dir "+dir)
if os.path.exists(dir) :
lfiles= os.listdir(dir)
for fil in lfiles :
#clogger.debug("Looking at file "+fil)
fileperiod=periodOfEmFile(fil,realm,f)
if fileperiod and period.intersects(fileperiod) :
if fileHasVar(dir+"/"+fil,variable) :
rep.append(dir+"/"+fil)
#clogger.debug("Done with Looking at file "+fil)
else : clogger.error("Directory %s does not exist for EM simulation %s, realm %s "
"and frequency %s"%(dir,simulation,realm,f))
else :
clogger.info("No archive location found for "+
simulation+" for realm "+realm+" with: "+`command`)
return rep
def selectExampleFiles(urls,**kwargs) :
rep=[]
if (kwargs['frequency'] == "monthly") :
for l in urls :
for realm in ["A","L"] :
#dir=l+"/"+realm+"/Origin/Monthly/"+simulation
dir=l+"/"+realm
clogger.debug("Looking at dir "+dir)
if os.path.exists(dir) :
lfiles= os.listdir(dir)
for f in lfiles :
clogger.debug("Looking at file "+f)
fileperiod=periodOfEmFile(f,realm,'mon')
if fileperiod and fileperiod.intersects(kwargs['period']) :
if fileHasVar(dir+"/"+f,kwargs['variable']) :
rep.append(dir+"/"+f)
#else: print "No var ",variable," in file", dir+"/"+f
return rep
def selectExampleFiles(urls,**kwargs) :
rep=[]
if (kwargs['frequency'] == "monthly") :
for l in urls :
for realm in ["A","L"] :
#dir=l+"/"+realm+"/Origin/Monthly/"+simulation
dir=l+"/"+realm
clogger.debug("Looking at dir "+dir)
if os.path.exists(dir) :
lfiles= os.listdir(dir)
for f in lfiles :
clogger.debug("Looking at file "+f)
fileperiod=periodOfEmFile(f,realm,'mon')
if fileperiod and fileperiod.intersects(kwargs['period']) :
if fileHasVar(dir+"/"+f,kwargs['variable']) :
rep.append(dir+"/"+f)
#else: print "No var ",variable," in file", dir+"/"+f
return rep
def selectGenericFiles(urls, return_wildcards=None,merge_periods_on=None,**kwargs):
"""
Allow to describe a ``generic`` file organization : the list of files returned
by this function is composed of files which :
- match the patterns in ``url`` once these patterns are instantiated by
the values in kwargs, and
- contain the ``variable`` provided in kwargs
- match the `period`` provided in kwargs
In the pattern strings, no keyword is mandatory. However, for remote files,
filename pattern must include ${varname}, which is instanciated by variable
name or ``filenameVar`` (given via :py:func:`~climaf.classes.calias()`); this is
for the sake of efficiency (please complain if inadequate)
Example :
>>> selectGenericFiles(project='my_projet',model='my_model', simulation='lastexp', variable='tas', period='1980', urls=['~/DATA/${project}/${model}/*${variable}*${PERIOD}*.nc)']
/home/stephane/DATA/my_project/my_model/somefilewith_tas_Y1980.nc
In the pattern strings, the keywords that can be used in addition to the argument
names (e.g. ${model}) are:
- ${variable} : use it if the files are split by variable and
filenames do include the variable name, as this speed up the search
- ${PERIOD} : use it for indicating the period covered by each file, if this
is applicable in the file naming; this period can appear in filenames as
YYYY, YYYYMM, YYYYMMDD, YYYYMMDDHHMM, either once only, or twice with
separator ='-' or '_'
- wildcards '?' and '*' for matching respectively one and any number of characters
"""
def store_wildcard_facet_values(f,facets_regexp, kwargs, wildcards, merge_periods_on=None,
fperiod=None,periods=None,periods_dict=None):
""""
"""
if fperiod is not None and periods is not None :
clogger.debug('Adding period %s'%fperiod)
periods.append(fperiod)
#
for kw in kwargs :
it=re.finditer(facets_regexp,f)
for oc in it :
try : facet_value=oc.group(kw)
except : continue
if type(kwargs[kw]) is str and ("*" in kwargs[kw] or "?" in kwargs[kw] ):
if facet_value is not None :
if kw not in wildcards : wildcards[kw]=set()
wildcards[kw].add(facet_value)
clogger.debug("Discover %s=%s for file=%s"%(kw,facet_value,f))
else :
clogger.debug("Logic issue for kw=%s and file=%s"%(kw,f))
#
if fperiod is not None and periods is not None :
if merge_periods_on is None : key=None
elif kw == merge_periods_on : key=facet_value
else :
#print "Skipping for kw=%s,sort=%s"%(kw,merge_periods_on)
continue
if key not in periods_dict: periods_dict[key]=set()
#print "adding period %s for key %s"%(fperiod,key)
periods_dict[key].add(fperiod)
else:
pass
#print "no Adding period for %s=%s for %s"%(kw,facet_value,f)
#print "end of store, periods_dict=",periods_dict, "wild=",wildcards
rep=[]
#
periods=None # a list of periods available
periods_dict=dict()
#
period=kwargs['period'] ;
if period == "*" :
periods=[] # List of all periods
elif type(period) is str : period=init_period(period)
#
variable=kwargs['variable']
altvar=kwargs.get('filenameVar',variable)
#
# dicts of date patterns, for globbing and for regexp
#
digit="[0-9]"
date_glob_patt={ "${PERIOD}" : "*" }
# an ordered list of dates keywords
date_keywords=date_glob_patt.keys() ; date_keywords.sort(reverse=True)
#
annee="%s{4}"%digit
mois="(01|02|03|04|05|06|07|08|09|10|11|12)"
jour="([0-3][0-9])"
heure="(00|01|02|03|04|05|06|07|08|09|10|11|12|13|14|15|16|17|18|19|20|21|22|23)"
minutes="[0-5][0-9]"
date="%s(%s(%s(%s(%s)?)?)?)?"%(annee,mois,jour,heure,minutes)
rperiod="(?P<period>(?P<start>%s)([_-](?P<end>%s))?)"%(date,date)
date_regexp_patt={ "${PERIOD}" : rperiod }
# an ordered list of dates regexp keywords
date_regexp_keywords=date_regexp_patt.keys() ; date_regexp_keywords.sort(reverse=True)
#
#
for l in urls :
# Instantiate keywords in pattern with attributes values
remote_prefix="" ;
if re.findall(".*:.*",l) :
remote_prefix=':'.join(l.split(":")[0:-1])+':'
basename=l.split(":")[-1] # This discard the remote_prefix if any
basename=basename.replace("//","/")
my_template=Template(basename)
template=my_template.safe_substitute(**kwargs)
#print "template after attributes replace : "+template
#
# Construct a pattern for globbing dates
temp2=template
for k in date_keywords : temp2=temp2.replace(k,date_glob_patt[k])
# Do globbing with plain varname
if remote_prefix :
lfiles=sorted(glob_remote_data(remote_prefix, temp2))
clogger.debug("Remote globbing %d files for varname on %s : "%\
(len(lfiles),remote_prefix+temp2))
else: # local data
lfiles=sorted(glob.glob(temp2))
clogger.debug("Before regexp filtering : Globbing %d files for varname on %s : "%(len(lfiles),temp2))
# Must filter with regexp, because * with glob is too inclusive
alt=[]
for f in lfiles :
for k in date_keywords :
if re.search(date_regexp_patt[k],f) :
alt.append(f)
continue
lfiles=alt
clogger.debug("Globbing %d files for varname on %s : "%(len(lfiles),temp2))
#
# If unsuccessful using varname, try with filenameVar
if len(lfiles)==0 and "filenameVar" in kwargs and kwargs['filenameVar'] :
# Change value of facet 'variable'
kwargs['variable']=kwargs['filenameVar']
template=my_template.safe_substitute(**kwargs)
temp2=template
for k in date_keywords : temp2=temp2.replace(k,date_glob_patt[k])
#
# Do globbing with fileVarname
if remote_prefix : #
lfiles=sorted(glob_remote_data(remote_prefix, temp2))
clogger.debug("Remote globbing %d files for filenamevar on %s: "%\
(len(lfiles),remote_prefix+temp2))
else: # local data
lfiles=sorted(glob.glob(temp2))
# Must filter with regexp, because * with glob is too inclusive
alt=[]
for f in lfiles :
for k in date_keywords :
if re.search(date_regexp_patt[k],f) :
alt.append(f)
continue
lfiles=alt
clogger.debug("Globbing %d files for filenamevar on %s: "%(len(lfiles),temp2))
#
# For discovering values for those facets which are a wildcard,
# construct a regexp with a group name for all facets (but period)
alt_basename=basename.replace("?",".").replace("*",".*")
alt_kwargs=kwargs.copy()
for kw in kwargs :
if type(kwargs[kw]) is str : # This excludes period attribute, which has a type
alt_kwargs[kw]=kwargs[kw].replace("?",".").replace("*",".*")
alt_basename=alt_basename.replace(r"${%s}"%kw,r"(?P<%s>%s)"%(kw,alt_kwargs[kw]),1)
facets_regexp=Template(alt_basename).safe_substitute(**alt_kwargs)
for k in date_regexp_keywords :
facets_regexp=facets_regexp.replace(k,date_regexp_patt[k],1)
facets_regexp=facets_regexp.replace(k,".*")
wildcards=dict()
#print "facets_regexp=",facets_regexp
#
# Construct regexp for extracting dates from filename
date_regexp=None
template_toreg=template.replace("*",".*").replace("?",r".").replace("+","\+")
#print "template before searching dates : "+template_toreg
for key in date_regexp_keywords :
#print "searchin "+key+" in "+template
start=template_toreg.find(key)
if (start>=0 ) :
date_regexp=template_toreg.replace(key,date_regexp_patt[key],1)
#print "found ",key," dateregexp ->",date_regexp
hasEnd=False
start=date_regexp.find(key)
#start=date_regexp.find(key)
if (start >=0 ) :
hasEnd=True
date_regexp=date_regexp.replace(key,date_regexp_patt[key],1)
#date_regexp=date_regexp.replace(key,date_regexp_patt[key],1)
break
#print "date_regexp before searching dates : "+date_regexp
#
for f in lfiles :
#print "processing file "+f
#
# Extract file time period
#
fperiod=None
if date_regexp :
if "P<period>" in date_regexp :
#print "date_rexgep=",date_regexp
#print "f=",f
#print "period=",re.sub(date_regexp,r'\g<period>',f)
tperiod=re.sub(date_regexp,r'\g<period>',f)
if tperiod==f :
raise classes.Climaf_Error("Cannot find a period in %s with regexp %s"%(f,date_regexp))
fperiod=init_period(tperiod)
else:
date_regexp0=date_regexp
#print "date_regexp for extracting dates : "+date_regexp0, "file="+f
start=re.sub(date_regexp0,r'\1',f)
if start==f:
raise Climaf_Data_Error("Start period not found in %s using regexp %s"%(f,regexp0)) #?
if hasEnd :
end=re.sub(date_regexp0,r'\2',f)
fperiod=init_period("%s-%s"%(start,end))
else :
fperiod=init_period(start)
#print "period for file %s is %s"%(f,fperiod)
#
# Filter file time period against required period
else :
if ( 'frequency' in kwargs and ((kwargs['frequency']=="fx") or \
kwargs['frequency']=="seasonnal" or kwargs['frequency']=="annual_cycle" )) :
# local data
if not remote_prefix and \
( (basename.find("${variable}")>=0) or variable=='*' or \
fileHasVar(f,variable) or (variable != altvar and fileHasVar(f,altvar)) ) :
clogger.debug("adding fixed field :"+f)
store_wildcard_facet_values(f,facets_regexp, kwargs, wildcards,merge_periods_on)
rep.append(f)
# remote data
elif remote_prefix :
if (basename.find("${variable}")>=0) or variable=='*' or \
(variable != altvar and (f.find(altvar)>=0) ):
clogger.debug("adding fixed field :"+remote_prefix+f)
store_wildcard_facet_values(f,facets_regexp, kwargs, wildcards,merge_periods_on)
rep.append(remote_prefix+f)
else:
raise classes.Climaf_Error(
"For remote files, filename pattern (%s) should include ${varname} "+\
"(which is instanciated by variable name or filenameVar)"%f)
else :
clogger.info("Cannot yet filter files re. time using only file content.")
store_wildcard_facet_values(f,facets_regexp, kwargs, wildcards,merge_periods_on)
rep.append(f)
#
# If file period matches requested period, check similarly for variable
#
#print "fperiod=",fperiod
#print "periods=",periods
#print "inter=",period.intersects(fperiod)
#print "date_regexp=",date_regexp
if (fperiod and ( periods is not None or period.intersects(fperiod) )) \
or not date_regexp :
#
clogger.debug('Period is OK - Considering variable filtering on %s and %s for %s'%(variable,altvar,f))
# Filter against variable
if (l.find("${variable}")>=0):
clogger.debug('appending %s based on variable in filename'%f)
store_wildcard_facet_values(f,facets_regexp, kwargs, wildcards,merge_periods_on,
fperiod,periods,periods_dict)
rep.append(remote_prefix+f)
continue
if (f not in rep):
# local data
if not remote_prefix and \
(variable=='*' or "," in variable or fileHasVar(f,variable) or \
(altvar != variable and fileHasVar(f,altvar))) :
# Should check time period in the file if not date_regexp
clogger.debug('appending %s based on multi-var or var exists in file '%f)
store_wildcard_facet_values(f,facets_regexp, kwargs, wildcards,merge_periods_on,
fperiod,periods,periods_dict)
rep.append(f)
continue
# remote data
elif remote_prefix :
if variable=='*' or "," in variable or \
(variable != altvar and (f.find(altvar)>=0) ):
# Should check time period in the file if not date_regexp
clogger.debug('appending %s based on multi-var or altvar '%(remote_prefix+f))
store_wildcard_facet_values(f,facets_regexp, kwargs, wildcards, merge_periods_on,
fperiod,periods,periods_dict)
rep.append(remote_prefix+f)
continue
else:
mess="For remote files, filename pattern (%s) should include"%(remote_prefix+f)
mess+=" ${varname} (which is instanciated by variable name or filenameVar)"
raise classes.Climaf_Error(mess)
else:
if not fperiod :
clogger.debug('not appending %s because period is None '%f)
elif not period.intersects(fperiod) :
clogger.debug('not appending %s because period doesn t intersect %s'%(f,period))
else:
clogger.debug('not appending %s for some other reason %s'%(f))
# Break on first url with any matching data
if len(rep)>0 :
clogger.debug('url %s does match for '%l + `kwargs`)
break
# For wildcard facets, discover facet values + checks
for facet in wildcards:
s=wildcards[facet]
if return_wildcards is not None :
if facet=="period" :
#print "s=",s," periods_dict=",periods_dict
for val in periods_dict :
periods_dict[val]=sort_periods_list(list(periods_dict[val]))
clogger.info("Attribute period='*' has values %s"%(periods_dict))
return_wildcards["period"]=periods_dict
else:
if len(s) == 1 :
s=s.pop()
clogger.info("Attribute %s='%s' has matching value '%s'"%(facet,kwargs[facet],s))
return_wildcards[facet]=s
else:
rep=list(s); rep.sort()
return_wildcards[facet]=rep
message="Attribute %s='%s' has multiple values : %s"%(facet,kwargs[facet],list(s))
if return_wildcards : clogger.info(message)
else: clogger.error(message)
s=return_wildcards[facet]
else:
clogger.debug("return_wildcards is None")
return rep
def selectGenericFiles(urls, **kwargs):
"""
Allow to describe a ``generic`` file organization : the list of files returned
by this function is composed of files which :
- match the patterns in ``url`` once these patterns are instantiated by
the values in kwargs, and
- contain the ``variable`` provided in kwargs
- match the `period`` provided in kwargs
In the pattern strings, no keyword is mandatory. However, for remote files,
filename pattern must include ${varname}, which is instanciated by variable
name or ``filenameVar`` (given via :py:func:`~climaf.classes.calias()`); this is
for the sake of efficiency (please complain if inadequate)
Example :
>>> selectGenericFiles(project='my_projet',model='my_model', simulation='lastexp', variable='tas', period='1980', urls=['~/DATA/${project}/${model}/*${variable}*YYYY*.nc)']
/home/stephane/DATA/my_project/my_model/somefilewith_tas_Y1980.nc
In the pattern strings, the keywords that can be used in addition to the argument
names (e.g. ${model}) are:
- ${variable} : use it if the files are split by variable and
filenames do include the variable name, as this speed up the search
- YYYY, YYYYMM, YYYYMMDD : use it for indicating the start date of
the period covered by each file, if this is applicable in the
file naming; use a second time for end date, if applicable
(otherwise the assumption is that the whole year -resp. month or
day- is included in the file
- wildcards '?' and '*' for matching respectively one and any number of characters
"""
rep = []
period = kwargs['period']
if type(period) is str: period = init_period(period)
variable = kwargs['variable']
altvar = kwargs.get('filenameVar', variable)
# a dict and an ordered list of date globbing patterns
dt = dict(YYYY="????",
YYYYMM="??????",
YYYYMMDD="????????",
YYYYMMDDHH="??????????")
lkeys = dt.keys()
lkeys.sort(reverse=True)
# a dict and an ordered list for matching dates
dr = dict(YYYY="([0-9]{4})",
YYYYMM="([0-9]{6})",
YYYYMMDD="([0-9]{8})",
YYYYMMDDHH="([0-9]{10})")
rkeys = dr.keys()
rkeys.sort(reverse=True)
#
for l in urls:
# Instantiate keywords in pattern with attributes values
if re.findall(".*:.*", l): # remote data
remote_prefix = ':'.join(l.split(":")[0:-1]) + ':'
template = Template(l.split(":")[-1]).safe_substitute(**kwargs)
else: # local data
remote_prefix = ""
template = Template(l).safe_substitute(**kwargs)
#print "template after attributes replace : "+template
#
# Construct a pattern for globbing dates
temp2 = template
for k in lkeys:
temp2 = temp2.replace(k, dt[k])
if remote_prefix:
lfiles = sorted(glob_remote_data(remote_prefix, temp2))
clogger.debug("Remote globbing %d files for varname on %s : " %
(len(lfiles), remote_prefix + temp2))
else: # local data
lfiles = sorted(glob.glob(temp2))
clogger.debug("Globbing %d files for varname on %s : " %
(len(lfiles), temp2))
#
# If unsuccessful using varname, try with filenameVar
if len(lfiles
) == 0 and "filenameVar" in kwargs and kwargs['filenameVar']:
# Change value of facet 'variable'
kwargs['variable'] = kwargs['filenameVar']
if remote_prefix: # remote data
template = Template(l.split(":")[-1]).safe_substitute(**kwargs)
else: # local data
template = Template(l).safe_substitute(**kwargs)
temp2 = template
for k in lkeys:
temp2 = temp2.replace(k, dt[k])
#
if remote_prefix: #
lfiles = sorted(glob_remote_data(remote_prefix, temp2))
clogger.debug("Globbing %d files for filenamevar on %s: " %
(len(lfiles), remote_prefix + temp2))
else: # local data
lfiles = sorted(glob.glob(temp2))
clogger.debug("Globbing %d files for filenamevar on %s: " %
(len(lfiles), temp2))
#
# Construct regexp for extracting dates from filename
regexp = None
#print "template before searching dates : "+template
for key in rkeys:
#print "searchin "+key+" in "+=Template(l)
start = template.find(key)
if (start >= 0):
#print "found "+key
regexp = template.replace(key, dr[key], 1)
hasEnd = False
start = regexp.find(key)
if (start >= 0):
hasEnd = True
regexp = regexp.replace(key, dr[key], 1)
break
#print "regexp before searching dates : "+regexp
#
for f in lfiles:
#print "processing file "+f
#
# Analyze file time period
fperiod = None
if regexp:
regexp0 = regexp.replace("*", ".*").replace("?", r".")
#print "regexp for extracting dates : "+regexp
start = re.sub(regexp0, r'\1', f)
if start == f:
raise Climaf_Data_Error("Start period not found") #?
if hasEnd:
end = re.sub(regexp0, r'\2', f)
fperiod = init_period("%s-%s" % (start, end))
else:
fperiod = init_period(start)
#print "period for file %s is %s"%(f,fperiod)
#
# Filter file time period against required period
else:
if ( 'frequency' in kwargs and ((kwargs['frequency']=="fx") or \
kwargs['frequency']=="seasonnal" or kwargs['frequency']=="annual_cycle" )) :
# local data
if remote_prefix and \
( (l.find("${variable}")>=0) or variable=='*' or \
fileHasVar(f,variable) or (variable != altvar and fileHasVar(f,altvar)) ) :
clogger.debug("adding fixed field :" + f)
rep.append(f)
# remote data
elif remote_prefix is not "":
if (l.split(":")[-1].find("${variable}")>=0) or variable=='*' or \
(variable != altvar and (f.find(altvar)>=0) ):
clogger.debug("adding fixed field :" +
remote_prefix + f)
rep.append(remote_prefix + f)
else:
raise Climaf_Data_Error(
"For remote files, filename pattern (%s) should include ${varname} (which is instanciated by variable name or filenameVar)"
% f)
else:
clogger.info(
"Cannot yet filter files re. time using only file content."
)
rep.append(f)
if (fperiod and period.intersects(fperiod)) or not regexp:
clogger.debug(
'Period is OK - Considering variable filtering on %s and %s for %s'
% (variable, altvar, f))
# Filter against variable
if (l.find("${variable}") >= 0):
clogger.debug(
'appending %s based on variable in filename' % f)
rep.append(remote_prefix + f)
continue
if (f not in rep):
# local data
if not remote_prefix and \
(variable=='*' or "," in variable or fileHasVar(f,variable) or \
(altvar != variable and fileHasVar(f,altvar))) :
# Should check time period in the file if not regexp
clogger.debug(
'appending %s based on multi-var or var exists in file '
% f)
rep.append(f)
continue
# remote data
elif remote_prefix:
if variable=='*' or "," in variable or \
(variable != altvar and (f.find(altvar)>=0) ):
# Should check time period in the file if not regexp
clogger.debug(
'appending %s based on multi-var or altvar ' %
(remote_prefix + f))
rep.append(remote_prefix + f)
continue
else:
mess = "For remote files, filename pattern (%s) should include" % (
remote_prefix + f)
mess += " ${varname} (which is instanciated by variable name or filenameVar)"
raise Climaf_Data_Error(mess)
else:
if not fperiod:
clogger.debug('not appending %s because period is None ' %
f)
else:
if not period.intersects(fperiod):
clogger.debug(
'not appending %s because period doesn t intersect %s'
% (f, period))
# Break on first url with any matching data
if len(rep) > 0:
clogger.debug('url %s does match for ' % l + ` kwargs `)
break
return rep
def selectGenericFiles(urls, **kwargs):
"""
Allow to describe a ``generic`` file organization : the list of files returned
by this function is composed of files which :
- match the patterns in ``url`` once these patterns are instantiated by
the values in kwargs, and
- contain the ``variable`` provided in kwargs
- match the `period`` provided in kwargs
In the pattern strings, no keyword is mandatory
Example :
>>> selectGenericFiles(project='my_projet',model='my_model', simulation='lastexp', variable='tas', period='1980', urls=['~/DATA/${project}/${model}/*${variable}*YYYY*.nc)']
/home/stephane/DATA/my_project/my_model/somefilewith_tas_Y1980.nc
In the pattern strings, the keywords that can be used in addition to the argument
names (e.g. ${model}) are:
- ${variable} : use it if the files are split by variable and
filenames do include the variable name, as this speed up the search
- YYYY, YYYYMM, YYYYMMDD : use it for indicating the start date of
the period covered by each file, if this is applicable in the
file naming; use a second time for end date, if applicable
(otherwise the assumption is that the whole year -resp. month or
day- is included in the file
- wildcards '?' and '*' for matching respectively one and any number of characters
"""
rep=[]
period=kwargs['period']
if type(period) is str : period=init_period(period)
variable=kwargs['variable']
mustHaveVariable=False
if "filenameVar" in kwargs and kwargs['filenameVar'] :
kwargs['variable']=kwargs['filenameVar']
mustHaveVariable=True
for l in urls :
template=Template(l)
# There is no use to look for files which path is not specific
# to the required variable when we know it should
if l.find("${variable}") < 0 and mustHaveVariable :
continue
#
# Instantiate keywords in pattern with attributes values
template=template.safe_substitute(**kwargs)
#print "template after attributes replace : "+template
#
# Construct a pattern for globbing dates
temp2=template
dt=dict(YYYY="????",YYYYMM="??????",YYYYMMDD="????????")
for k in dt : temp2=temp2.replace(k,dt[k])
clogger.debug("Globbing on : "+temp2)
lfiles=glob.glob(temp2)
#
# Analyze all filenames
for f in lfiles :
# print "looking at file"+f
# Construct regexp for extracting dates from filename
dt=dict(YYYY="([0-9]{4})",YYYYMM="([0-9]{6})",
YYYYMMDD="([0-9]{10})")
regexp=None
# print "template before searching dates : "+template
lkeys=dt.keys() ; lkeys.sort(reverse=True)
for key in lkeys :
# print "searchin "+key+" in "+template
start=template.find(key)
if (start>=0 ) :
# print "found "+key
regexp=template.replace(key,dt[key],1)
hasEnd=False
start=regexp.find(key)
if (start >=0 ) :
hasEnd=True
regexp=regexp.replace(key,dt[key],1)
break
#
# Analyze file time period
fperiod=None
if regexp :
regexp=regexp.replace("*",".*").replace("?",r".")
# print "regexp for extracting dates : "+regexp
start=re.sub(regexp,r'\1',f)
if hasEnd :
end=re.sub(regexp,r'\2',f)
fperiod=init_period("%s-%s"%(start,end))
else :
fperiod=init_period(start)
#
# Filter file time period against required period
else :
if ( 'frequency' in kwargs and kwargs['frequency']=="fx") :
if (l.find("${variable}")>=0) or fileHasVar(f,variable) :
clogger.debug("adding fixed field :"+f)
rep.append(f)
else :
clogger.warning("Cannot yet filter files re. time using only file content. TBD")
rep.append(f)
if (fperiod and period.intersects(fperiod)) or not regexp :
# Filter against variable
if (l.find("${variable}")>=0) or fileHasVar(f,variable) :
# Should check time period in the file if not regexp
# print "appending "+f
rep.append(f)
return rep
def zonal_mean_slice(model, variable, basin, season, ref=None, add_product_in_title=True, method='regrid_model_on_ref',
custom_plot_params={}, safe_mode=True, do_cfile=True, y='lin', ymin=None, plot_on_latitude=False, horizontal_regridding=True,
apply_period_manager=True):
# -----------------------------------------------------------------------------------------------------------------------------
# -- 1/ Moyenne zonale du modele:
# -> soit on a la variable en moyenne zonale deja calculee (zoblabla)
# -> soit on la calcule a partir des masks de bassin definis dans model:
# * path_mesh_mask donne le path vers les fichiers
# * mesh_masks est un dictionnaire qui pointe le fichier de mask pour chaque bassin (GLO, ATL, PAC, IND et ALLBAS)
# -- Test pour voir si on a deja les moyennes zonales par bassin dans les variables dispos
context = 'full_field'
# -- Si method=='regrid_model_on_obs', on regrille le modele sur la grille de la reference, et on utilisera
# les masks de bassin de la reference qui sont dans son repertoire ($bassin_mask.nc)
# Cette methode favorise la structure latitudinale de la section (=> gradient eq/pole)
if method=='regrid_model_on_ref':
# -- Apply the frequency and time manager (IGCM_OUT)
wmodel=model.copy() ; wmodel.update(dict(variable=variable))
if apply_period_manager:
frequency_manager_for_diag(wmodel, diag='clim')
get_period_manager(wmodel)
# -- Get the model data
model_dat = ds(**wmodel)
# -- Compute the climatology on the model grid and mask the zeros
clim_model = clim_average( mask(model_dat, miss=0.0), season)
#
# -- Get the reference
if ref:
ref.update(dict(variable=variable))
# -- Get the reference data
ref_dat = ds(**ref)
#
# -- Get the context => model_model or bias
context = ('bias' if 'product' in ref_dat.kvp else 'model_model')
#
# -- Compute the climatology
clim_ref = clim_average(ref_dat, season)
#
# -- Regrid the model on the obs
if safe_mode:
try:
rgrd_clim_model = lonlatvert_interpolation( regrid(clim_model, clim_ref, option='remapdis'), clim_ref, horizontal_regridding=False )
except:
print '--> Error in lonlatvert_interpolation( regrid(clim_model, clim_ref, option="remapdis"), clim_ref, horizontal_regridding=False )'
print '--> Set safe_mode=False to see the error'
rgrd_clim_model = clim_model
else:
rgrd_clim_model = lonlatvert_interpolation( regrid(clim_model, clim_ref, option='remapdis'), clim_ref, horizontal_regridding=False )
print '----'
print '----'
print '----'
print 'rgrd_clim_model = ',cfile(rgrd_clim_model)
print 'clim_model = ', cfile(clim_model)
print 'clim_ref = ', cfile(clim_ref)
print '----'
print '----'
print '----'
#
# -- Get the reference mask
if 'path_mesh_mask' in ref:
mask_file = ref['path_mesh_mask'] + ref['mesh_masks'][basin]
else:
mask_file = os.path.dirname(str.split(ref_dat.baseFiles(),' ')[0])+'/'+basin.lower()+'_mask.nc'
print '----'
print '----'
print '----'
print '---> mask_file = ', mask_file
print '----'
print '----'
print '----'
mask_dat = fds( mask_file, variable='mask', period='fx')
basin_mask = mask( mask_dat, miss=0.0)
#
# -- Apply the mask to the model and the ref
masked_model = multiply(rgrd_clim_model, basin_mask)
masked_ref = multiply(clim_ref, basin_mask)
#
if 'product' not in ref:
masked_model = regridn(masked_model, cdogrid='r360x180', option='remapdis')
masked_ref = regridn(masked_ref, cdogrid='r360x180', option='remapdis')
# -- Compute the zonal means
ZM_MODEL = zonmean(masked_model)
ZM_REF = zonmean(masked_ref)
#
#print '==='
#print '==='
#print '==='
#print '=== ZM_MODEL = ',cfile(ZM_MODEL)
#print '=== ZM_REF = ',cfile(ZM_REF)
#print '==='
#print '==='
#print '==='
# -- Interpolate vertically and compute the difference
if safe_mode:
try:
ZM_bias = diff_zonmean(ZM_MODEL, ZM_REF)
except:
print '--> Error in diff_zonmean(ZM_MODEL, ZM_REF)'
print '--> Set safe_mode=False to track the error'
ZM_bias = minus(ZM_MODEL, ZM_REF)
else:
ZM_bias = diff_zonmean(ZM_MODEL, ZM_REF)
# -- Compute the zonal mean for the basin using the obs masks
else:
print 'No reference (obs) provided in zonal_mean_slice for method regrid_model_on_obs'
# -- Get the reference mask
if 'path_mesh_mask' in model:
mask_file = model['path_mesh_mask'] + model['mesh_masks'][basin]
else:
mask_file = os.path.dirname(str.split(model_dat.baseFiles(),' ')[0])+'/'+basin.lower()+'_mask.nc'
print 'mask_file = ', mask_file
mask_dat = fds( mask_file, variable='mask', period='fx')
basin_mask = mask( mask_dat, miss=0.0)
#
# -- Apply the mask to the model and the ref
masked_model = multiply(clim_model, basin_mask)
#
if 'product' not in model:
masked_model = regridn(masked_model, cdogrid='r360x180', option='remapdis')
# -- Compute the zonal means
ZM_MODEL = zonmean(masked_model)
#
if method=='regrid_ref_on_model':
#
# -> Cette methode
#if variable=='thetao': tmpzonmvar = 'zotem'+region.lower()
#if variable=='so': tmpzonmvar = 'zosal'+region.lower()
# -- Apply the frequency and time manager (IGCM_OUT)
#wmodel=model.copy() #; wmodel.update(dict(variable=tmpzonmvar))
#wmodel.update(dict(variable=variable))
#frequency_manager_for_diag(wmodel, diag='clim')
#get_period_manager(wmodel)
#model_dat = ds(**wmodel) # -> on regarde si ds() trouve un fichier qui correspondn a la variable
#if tmp.baseFiles():
# # --> Fix to add nav_lat to the file
# if not fileHasDim(cfile(tmp),'nav_lat'):
# zonmean_model = add_nav_lat(tmp, nav_lat_file=nav_lat_zovarbasin_file(grid=whichORCAGrid(cfile(tmp))),
# coordinates=build_coordinates_zovarbasin(cfile(tmp)))
# else:
# zonmean_model = tmp
# modvar_climato_zonmean_basin = mask(clim_average(zonmean_model, season), miss=0.0)
#else:
#
# -- Apply the frequency and time manager (IGCM_OUT)
wmodel=model.copy() ; wmodel.update(dict(variable=variable))
if apply_period_manager:
frequency_manager_for_diag(wmodel, diag='SE')
get_period_manager(wmodel)
model_dat = ds(**wmodel)
model_clim = ccdo(clim_average(model_dat,season), operator='setctomiss,0')
if fileHasVar(cfile(model_clim), 'lev'):
model_clim_ok = rename_depth(model_clim)
else:
model_clim_ok = model_clim
#
# if method=='regrid_model_on_1deg_grid':
# # --> In this case, we regrid the model on the obs
#maskfile = model['path_mesh_mask'] + model['mesh_masks'][basin]
#wmask = ccdo(fds(maskfile, variable='tmask', period='fx'), operator='setctomiss,0')
#modvar_climato_masked = multiply(model_clim_ok, wmask)
# modvar_rgrd = regridn(modvar_climato_masked, cdogrid='r360x180',option='remapdis')
# modvar_climato_zonmean_basin = zonmean(modvar_rgrd)
# #
#else:
set_fixed_fields('ccdfzonalmean_bas', basin, model)
# calculer la moyenne zonale pour le bassin choisi
model_clim_zonmean_basin = ccdfzonalmean_bas(model_clim_ok, point_type='T', basin=str(basin).lower())
# -- Ajouter les latitudes ici??
#else:
ZM_MODEL = model_clim_zonmean_basin
print '--'
print '--'
print '--'
print 'cfile(ZM_MODEL) = ',cfile(ZM_MODEL)
print '--'
print '--'
print '--'
#
# -----------------------------------------------------------------------------------------------------------------------------
# -- 2/ Moyenne zonale de la ref:
# -- -> les refs sont fournies avec les masks de bassins; si la ref est un modele,
# on peut recuperer path_mesh_mask et mesh_masks (et donc les fichiers de masks de bassins)
if ref:
# calculer la climatologie pour la saison choisie
if 'variable' not in ref: ref.update(dict(variable=variable))
ref_dat = ds(**ref)
ref_clim = clim_average(ref_dat, season)
# -- Check whether the ref is a model or an obs to set the appropriate context
context = ('bias' if 'product' in ref_dat.kvp else 'model_model')
# 1. Si le context est 'model_model', on verifie si la variable ne serait pas disponible en moyenne zonale
# - si oui, on travaille directement avec celle-ci
# - si non, on recupere les masks de bassins
# 2. Si le context est 'bias', on recupere les masks de bassins qui doivent etre dans le repertoire des obs
# A partir des masks, on calcule les moyennes zonales par bassin
#zovarbas_ref = ref.copy() ; zovarbas.update(dict(variable=tmpzonmvar))
#tmpref = ds(**zovarbas_ref) # -> on regarde si ds() trouve un fichier qui correspondn a la variable
# -- Si on a les variables pre-calculees en moyennes zonales pour le model et les obs, on utilise ces moyennes zonales
# --> Ok si on utilise WOA13-v2 comme reference
#if tmpref.baseFiles() and tmp.baseFiles():
# ref_clim_zonmean_basin_interp = regridn(mask(clim_average(tmpref, season), miss=0.0), cdogrid='r1x180', option='remapdis')
# model_clim_zonmean_basin_interp = regridn(model_clim_zonmean_basin, cdogrid='r1x180', option='remapdis')
# ZM_OBS = zonmean_interpolation(ref_clim_zonmean_basin_interp, model_clim_zonmean_basin_interp)
# ZM_MODEL = model_clim_zonmean_basin_interp
#else:
# #
## ref_clim = mask(clim_average(ref_dat, season), miss=0.0)
if fileHasVar(cfile(ref_clim), 'lev'):
ref_clim_ok = rename_depth(ref_clim)
else:
ref_clim_ok = ref_clim
print "cfile(ref_clim_ok) = ",cfile(ref_clim_ok)
#
# -- Si 'ref' est un autre simulation et a des mesh_masks, on les utilisent
if context=='model_model' and 'mesh_masks' in ref:
set_fixed_fields('ccdfzonalmean_bas', basin, ref)
ref_clim_interp = ref_clim_ok
else:
# -> Sinon, on regrille 'obs' sur le modele, et on utilise les masks de bassins
# -> du model pour calculer les moyennes zonales
#ref_clim_interp = regrid(ref_clim_ok, model_clim_ok, option='remapdis')
ref_clim_interp = ccdo(ref_clim_ok, operator='remapdis,'+cfile(model_clim_ok))
test = lonlatvert_interpolation(ref_clim_interp, model_clim_ok, horizontal_regridding=False)
ref_clim_zonmean_basin = ccdfzonalmean_bas(test, point_type='T', basin=str(basin).lower())
# calculer la moyenne zonale pour le bassin choisi
ZM_OBS = zonmean_interpolation(ref_clim_zonmean_basin, model_clim_zonmean_basin, horizontal_regridding=False)
ZM_bias = minus(ZM_MODEL, ZM_OBS)
#
# -- Now compute the difference (bias)
if method=='regrid_on_1deg':
print 'Not yet available : ',method
# Plot
#
# -- Get the period for display in the plot: we build a tmp_period string
# -- Check whether the period is described by clim_period, years or period (default)
# -- and make a string with it
tmp_period = build_period_str(wmodel)
#
# -- Title of the plot -> If 'customname' is in the dictionary of dat, it will be used
# -- as the title. If not, it checks whether dat is a reference or a model simulation
# -- and builds the title
title = build_plot_title(wmodel, None)# add_product_in_title='') #add_product_in_title)
#
# -- Get the default plot parameters with the function 'plot_params'
# -- We also update with a custom dictionary of params (custom_plot_params) if the user sets one
p = plot_params(variable+'_zonmean', context, custom_plot_params=custom_plot_params)
p.update(dict(y=y,
contours=1,
tiMainFontHeightF=0.023,tiMainFont="helvetica-bold",
gsnStringFontHeightF=0.019,
options="cnMissingValFillColor=gray|trYReverse=True|"+\
"vpHeightF=0.4|vpWidthF=0.8|"+\
"pmLabelBarWidthF=0.075|pmLabelBarOrthogonalPosF=0.01|lbLabelFontHeightF=0.012|"
))
if ymin: p['options']=p['options']+'|trYMinF='+str(ymin)
#
# -- Set the left, center and right strings of the plot
p.update(dict(gsnRightString = tmp_period,
gsnCenterString = variable+' '+method,
gsnLeftString = basin))
#
if ref:
ZM = ZM_bias
else:
ZM = ZM_MODEL
#
plot_zonmean = plot(ZM, title=title, **p)
# -- If the user doesn't want to do the cfile within plot_climato, set do_cfile=False
# -- Otherwise we check if the plot has been done successfully.
# -- If not, the user can set safe_mode=False and clog('debug') to debug.
return safe_mode_cfile_plot(plot_zonmean, do_cfile, safe_mode)
def selectGenericFiles(urls, **kwargs):
"""
Allow to describe a ``generic`` file organization : the list of files returned
by this function is composed of files which :
- match the patterns in ``url`` once these patterns are instantiated by
the values in kwargs, and
- contain the ``variable`` provided in kwargs
- match the `period`` provided in kwargs
In the pattern strings, no keyword is mandatory
Example :
>>> selectGenericFiles(project='my_projet',model='my_model', simulation='lastexp', variable='tas', period='1980', urls=['~/DATA/${project}/${model}/*${variable}*YYYY*.nc)']
/home/stephane/DATA/my_project/my_model/somefilewith_tas_Y1980.nc
In the pattern strings, the keywords that can be used in addition to the argument
names (e.g. ${model}) are:
- ${variable} : use it if the files are split by variable and
filenames do include the variable name, as this speed up the search
- YYYY, YYYYMM, YYYYMMDD : use it for indicating the start date of
the period covered by each file, if this is applicable in the
file naming; use a second time for end date, if applicable
(otherwise the assumption is that the whole year -resp. month or
day- is included in the file
- wildcards '?' and '*' for matching respectively one and any number of characters
"""
rep=[]
period=kwargs['period']
if type(period) is str : period=init_period(period)
variable=kwargs['variable']
altvar=kwargs.get('filenameVar',variable)
# a dict and an ordered list of date globbing patterns
dt=dict(YYYY="????",YYYYMM="??????",YYYYMMDD="????????")
lkeys=dt.keys() ; lkeys.sort(reverse=True)
# a dict and an ordered list for matching dates
dr=dict(YYYY="([0-9]{4})",YYYYMM="([0-9]{6})", YYYYMMDD="([0-9]{8})")
rkeys=dr.keys() ; rkeys.sort(reverse=True)
#
for l in urls :
# Instantiate keywords in pattern with attributes values
template=Template(l).safe_substitute(**kwargs)
#print "template after attributes replace : "+template
#
# Construct a pattern for globbing dates
temp2=template ;
for k in lkeys : temp2=temp2.replace(k,dt[k])
lfiles=glob.glob(temp2)
clogger.debug("Globbing %d files for varname on %s : "%(len(lfiles),temp2))
#
# If unsuccessful using varname, try with filenameVar
if len(lfiles)==0 and "filenameVar" in kwargs and kwargs['filenameVar'] :
kwargs['variable']=kwargs['filenameVar']
template=Template(l).safe_substitute(**kwargs)
temp2=template
for k in lkeys : temp2=temp2.replace(k,dt[k])
#
lfiles=glob.glob(temp2)
clogger.debug("Globbing %d files for filenamevar on %s: "%(len(lfiles),temp2))
# Construct regexp for extracting dates from filename
regexp=None
#print "template before searching dates : "+template
for key in rkeys :
#print "searchin "+key+" in "+=Template(l)
start=template.find(key)
if (start>=0 ) :
#print "found "+key
regexp=template.replace(key,dr[key],1)
hasEnd=False
start=regexp.find(key)
if (start >=0 ) :
hasEnd=True
regexp=regexp.replace(key,dr[key],1)
break
#print "regexp before searching dates : "+regexp
#
for f in lfiles :
#print "processing file "+f
#
# Analyze file time period
fperiod=None
if regexp :
regexp0=regexp.replace("*",".*").replace("?",r".")
#print "regexp for extracting dates : "+regexp
start=re.sub(regexp0,r'\1',f)
if start==f:
raise Climaf_Data_Error("Start period not found") #? LV
if hasEnd :
end=re.sub(regexp0,r'\2',f)
fperiod=init_period("%s-%s"%(start,end))
else :
fperiod=init_period(start)
#print "period for file %s is %s"%(f,fperiod)
#
# Filter file time period against required period
else :
if ( 'frequency' in kwargs and ((kwargs['frequency']=="fx") or \
kwargs['frequency']=="seasonnal" or kwargs['frequency']=="annual_cycle" )) :
if (l.find("${variable}")>=0) or fileHasVar(f,variable) or fileHasVar(f,altvar) :
clogger.debug("adding fixed field :"+f)
rep.append(f)
else :
clogger.warning("Cannot yet filter files re. time using only file content. TBD")
rep.append(f)
if (fperiod and period.intersects(fperiod)) or not regexp :
clogger.debug('Period is OK - Considering variable filtering on %s and %s for %s'%(variable,altvar,f))
# Filter against variable
if (l.find("${variable}")>=0):
clogger.debug('appending %s based on variable in filename'%f)
rep.append(f)
continue
if f not in rep and ( fileHasVar(f,variable) or fileHasVar(f,altvar) or ("," in variable)):
# Should check time period in the file if not regexp
clogger.debug('appending %s based on multi-var or var exists in file '%f)
rep.append(f)
else:
if not fperiod :
clogger.debug('not appending %s because period is None '%f)
else:
if not period.intersects(fperiod) :
clogger.debug('not appending %s because period doesn t intersect %s'%(f,period))
return rep
