yukiitohand
commented
5 years ago This problem seems to be caused by an unexpected operation: fetching extra parameters for CO2 absorption line data, although no extra parameters for CO2 is stored in the current version of the HITRAN database (if I understand correctly. Therefore, fetching such parameters is unnecessary). When I fetch CO2 absorption line data with extra parameters, say,
fetch_by_ids('CO2_2000t12500',[7,8,9,10,11,12,13,14,121,15,120,122],2000,12500,
ParameterGroups=['LineMixing','Voigt_self','Voigt_CO2','SDVoigt'])then it still creates 'extra' component in the header file, which leads to a call of getRowObjectFromString when the data is loaded to cache. I make sure that this problem doesn't happen and already fixed when I only fetch standard parameters specified by 'par_line'.
The modification I put in the last comment turned out to be insufficient, so I put another patch here, although this may not be necessary...
def getRowObjectFromString(input_string,TableName):
# restore RowObject from string, get formats and names in TableName
#print 'getRowObjectFromString:'
pos = 0
RowObject = []
for par_name in LOCAL_TABLE_CACHE[TableName]['header']['order']:
par_format = LOCAL_TABLE_CACHE[TableName]['header']['format'][par_name]
regex = '^\%([0-9]+)\.?[0-9]*([dfs])$' #
regex = FORMAT_PYTHON_REGEX
(lng,trail,lngpnt,ty) = re.search(regex,par_format).groups()
lng = int(lng)
par_value = input_string[pos:(pos+lng)]
if ty=='d': # integer value
# modified fix: only affects 'local_iso_id'
###########################################
if par_name.lower() == 'local_iso_id':
if par_value.strip().isnumeric():
par_value = int(par_value)
if par_value == 0:
par_value = 10
else:
par_value = 11+ord(par_value)-ord('A')
else:
try:
par_value = int(par_value)
except:
par_value = 0
###########################################
elif ty.lower() in set(['e','f']): # float value
par_value = float(par_value)
elif ty=='s': # string value
pass # don't strip string value
else:
print('err1')
raise Exception('Format \"%s\" is unknown' % par_format)
RowObject.append((par_name,par_value,par_format))
pos += lng
# Do the same but now for extra (comma-separated) parameters
if 'extra' in set(LOCAL_TABLE_CACHE[TableName]['header']):
csv_chunks = input_string.split(LOCAL_TABLE_CACHE[TableName]['header'].\
get('extra_separator',','))
# Disregard the first "column-fixed" container if it presents:
if LOCAL_TABLE_CACHE[TableName]['header'].get('order',[]):
pos = 1
else:
pos = 0
for par_name in LOCAL_TABLE_CACHE[TableName]['header']['extra']:
par_format = LOCAL_TABLE_CACHE[TableName]['header']['extra_format'][par_name]
regex = '^\%([0-9]+)\.?[0-9]*([dfs])$' #
regex = FORMAT_PYTHON_REGEX
(lng,trail,lngpnt,ty) = re.search(regex,par_format).groups()
lng = int(lng)
par_value = csv_chunks[pos]
if ty=='d': # integer value
try:
# modified fix: I am not sure this is necessary
###########################################
if par_name.lower() == 'local_iso_id':
if par_value.strip().isnumeric():
par_value = int(par_value)
if par_value == 0:
par_value = 10
else:
par_value = 11+ord(par_value)-ord('A')
else:
par_value = int(par_value)
###########################################
except:
par_value = 0
elif ty.lower() in set(['e','f']): # float value
try:
par_value = float(par_value)
except:
par_value = 0.0
elif ty=='s': # string value
pass # don't strip string value
else:
print('err')
raise Exception('Format \"%s\" is unknown' % par_format)
RowObject.append((par_name,par_value,par_format))
pos += 1
return RowObject
I think the current version fails to load absorption line data of CO2 isotopes labelled as A&B (11&12) to cache. I come up with a simple patch fix for this by changing and modifying some lines in the function
and
getRowObjectFromString. My suggesting patch is at two locations:I uploaded this fix at my forked repo: https://github.com/yukiitohand/hapi/ I am using Python3.