def getrpnftndoc(filename):
"""Extract in a list of dictionaries the parsed RPN style doc string
of a Fortran file, one dic per sub or fn of prog
getrpnftndoc(filename)
"""
#Define regex
refunc = re.compile(r'^\s+([a-zA-Z][a-zA-Z0-9_*]+\s+function)\s+([a-zA-Z][a-zA-Z0-9_]+).*',re.I)
resub = re.compile(r'^\s+(subroutine|program)\s+([a-zA-Z][a-zA-Z0-9_]+).*',re.I)
retag = re.compile(r'^[*!]([a-zA-Z]+).*',re.I)
rebgn = re.compile(r'^[*!]\*\*([a-zA-Z].*)',re.I)
reend = re.compile(r'^[*!]\*.*',re.I)
ignorelist = [
re.compile(r'^#include\s+<model_macros_f\.h>',re.I),
re.compile(r'^#include\s+[\'"]impnone.cdk[\'"]',re.I),
re.compile(r'^\*[cv]dir\s+',re.I),
re.compile(r'^(\*|\s*!)?\s*$',re.I)
]
#read file
fsock = openAnything(filename)
a = fsock.read()
mylist = a.splitlines()
#initialize var for loop
indoc = False
insec = None
counter = 0
linecounter = 0
mydocdic = [{
'filename':filename,
'subname':'',
'startline':0,
'all':''
},
]
#loop through lines
for myline in mylist:
linecounter += 1
keepline = True
ignore = False
for reitem in ignorelist:
if re.search(reitem,myline,re.I):
ignore = True
if indoc and not ignore:
mydocdic[counter]['all'] += myline+"\n"
if re.search(retag,myline,re.I):
insec = (re.sub(retag,r'\1',myline)).lower()
keepline = False
elif re.search(resub,myline,re.I):
insec = 'main'
mydocdic[counter]['subname'] = \
re.sub(resub,r'\2',myline)
mydocdic[counter]['startline'] = linecounter
elif re.search(refunc,myline,re.I):
insec = 'main'
mydocdic[counter]['subname'] = \
re.sub(refunc,r'\2',myline)
mydocdic[counter]['startline'] = linecounter
#elif re.search(r'^\s+',myline): #code line
# pass
elif re.search(r'^[*cC!]\s+\W+',myline): #in section comment line
pass
elif re.search(r'^[^*cC!]\s*\W+',myline): #code line
pass
elif re.search(reend,myline): #end of doc section
indoc = False
keepline = False
if insec and keepline:
if insec in mydocdic[counter].keys():
mydocdic[counter][insec] += str(myline)+"\n"
else:
mydocdic[counter][insec] = myline+"\n"
elif re.search(rebgn,myline):
indoc = True
insec = 'title'
mydocdic[counter]['all'] += myline+"\n"
mydocdic[counter][insec] = myline+"\n"
#if len(subsecs)<1 or insec in subsecs:
# print insec,myline
return mydocdic
def tidy_f90(filename):
"""echo a tidy up version of a f90 program
tidy_f90(filename)
"""
emptypattern = [
# - merge multiple empty lines (empty comment == empty line)
(r'\n[ \t]*![ \t]*\n',"\n\n"),
(r'\n([ \t]*\n)+', "\n\n")
]
#Define regex
patternlist = [
# - remove trailing blanks
(r'[ \t]+\n', "\n"),
# - remove extra head blanks in comments (more than 3)
(r'\n[ \t]*![ \t][ \t][ \t]+',"\n! "),
# - remove extra spaces around parentesis [except after if]
(r'[ \t]*\([ \t]*', "("),
(r'[ \t]*\)', ")"),
(r'if\(', "if ("),
# - max one space after comma (none before)
(r'[ \t]+,', ","), #TODO: should not catch '\n[ \t]+\('
(r',[ \t]+', ", "),
# - at least one space before trailing &
(r'&[ \t]*\n', " &\n"), #TODO: should not add one if there
# - update doc format
(r'!author','!@author'),
(r'!revision','!@revisions'),
(r'!arguments','!@arguments'),
(r'!object','!@description'),
(r'\n!\*\*[ \t]*s/r[ \t]*','\n!/**\n!@objective'),
(r'\n!\*[ \t]*\n','\n!**/\n'),
# - misc
(r'end[ \t]+do','enddo'),
(r'end[ \t]+if','endif'),
(r'else[ \t]+if','elseif')
]
# - replace .eq. (and other) notation, one space around, lowercase
oprlist = [ (r'\.eq\.',r'=='),
(r'\.ne\.',r'/='),
(r'\.gt\.',r'>'),
(r'\.ge\.',r'>='),
(r'\.lt\.',r'<'),
(r'\.le\.',r'<=')
]
postfixlist = [ (r'\b[ \t]*\.or\.[ \t]*\b',' .or. '),
(r'\b[ \t]*\.and\.[ \t]*\b',' .and. '),
(r'[ \t]*=[ \t]*',' = '),
(r'\b\([ \t]*len[ \t]*=[ \t]*\b','(len='),
(r'\b\([ \t]*kind[ \t]*=[ \t]*\b','(kind='),
(r'= +=','=='),
(r'/ +=','/='),
(r'> +=','>='),
(r'< +=','<='),
(r'= +>','=>')
]
pattern = []
replace = []
for (opr1,opr2) in patternlist:
pattern.append(re.compile(opr1,re.I))
replace.append(opr2)
for (opr1,opr2) in oprlist:
pattern.append(re.compile(opr1,re.I))
replace.append(opr2)
# - one and only on space around = and other operators
for (opr1,opr2) in oprlist: #TODO:not in comments/string
pattern.append(re.compile(r'[ \t]*'+opr2+r'[ \t]*',re.I))
replace.append(' '+opr2+' ')
for (opr1,opr2) in postfixlist:
pattern.append(re.compile(opr1,re.I))
replace.append(opr2)
pattern.append(re.compile(opr1,re.I))
replace.append(opr2)
pattern.append(re.compile(opr1,re.I))
replace.append(opr2)
#read file
fsock = openAnything(filename)
myfilestr = fsock.read()
#s = re.sub("(?<= )(.+)(?= )", lambda m: "can use a callable for the %s text too" % m.group(1), s)
#print(re.sub(pattern,replace,myfilestr))
(myfilestr,commentlist,stringslist) = pull_comm_str(myfilestr)
i = 0
for mypatt in pattern:
myfilestr = re.sub(mypatt,replace[i],myfilestr)
i=i+1
myfilestr = indent_f90(myfilestr)
myfilestr = push_comm_str(myfilestr,commentlist,stringslist)
pattern = []
replace = []
for (opr1,opr2) in emptypattern:
pattern.append(re.compile(opr1,re.I))
replace.append(opr2)
i = 0
for mypatt in pattern:
myfilestr = re.sub(mypatt,replace[i],myfilestr)
i=i+1
#TODO:
# - replace licence
# - last end should with function/sub name
# - lower case [but not parameters]
# - external and parameters... now attributes
# - split too long lines
# - add () on subroutine name and call sub when missing
return myfilestr
