|
|
|
import os, sys, re, string, glob
|
|
|
|
allmodules = ["core", "flann", "imgproc", "ml", "highgui", "video", "features2d", "calib3d", "objdetect", "legacy", "contrib", "gpu", "androidcamera", "java", "python", "stitching", "ts", "photo", "nonfree", "videostab"]
|
|
|
|
verbose = False
|
|
|
|
show_warnings = True
|
|
|
|
show_errors = True
|
|
|
|
|
|
|
|
params_blacklist = {
|
|
|
|
"fromarray" : ("object", "allowND"), # python only function
|
|
|
|
"reprojectImageTo3D" : ("ddepth"), # python only argument
|
|
|
|
"composeRT" : ("d*d*"), # wildchards in parameter names are not supported by this parser
|
|
|
|
"CvSVM::train_auto" : ("\*Grid"), # wildchards in parameter names are not supported by this parser
|
|
|
|
"error" : "args", # parameter of supporting macro
|
|
|
|
"getConvertElem" : ("from", "cn", "to", "beta", "alpha"), # arguments of returned functions
|
|
|
|
"gpu::swapChannels" : ("dstOrder") # parameter is not parsed correctly by the hdr_parser
|
|
|
|
}
|
|
|
|
|
|
|
|
params_mapping = {
|
|
|
|
"composeRT" : {
|
|
|
|
"dr3dr1" : "d*d*",
|
|
|
|
"dr3dr2" : "d*d*",
|
|
|
|
"dr3dt1" : "d*d*",
|
|
|
|
"dr3dt2" : "d*d*",
|
|
|
|
"dt3dr1" : "d*d*",
|
|
|
|
"dt3dr2" : "d*d*",
|
|
|
|
"dt3dt1" : "d*d*",
|
|
|
|
"dt3dt2" : "d*d*"
|
|
|
|
},
|
|
|
|
"CvSVM::train_auto" : {
|
|
|
|
"coeffGrid" : "\\*Grid",
|
|
|
|
"degreeGrid" : "\\*Grid",
|
|
|
|
"gammaGrid" : "\\*Grid",
|
|
|
|
"nuGrid" : "\\*Grid",
|
|
|
|
"pGrid" : "\\*Grid"
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
class DeclarationParser(object):
|
|
|
|
def __init__(self, line=None):
|
|
|
|
if line is None:
|
|
|
|
self.fdecl = ""
|
|
|
|
self.lang = ""
|
|
|
|
self.balance = 0
|
|
|
|
return
|
|
|
|
self.lang = self.getLang(line)
|
|
|
|
assert self.lang is not None
|
|
|
|
self.fdecl = line[line.find("::")+2:].strip()
|
|
|
|
self.balance = self.fdecl.count("(") - self.fdecl.count(")")
|
|
|
|
assert self.balance >= 0
|
|
|
|
|
|
|
|
def append(self, line):
|
|
|
|
self.fdecl += line
|
|
|
|
self.balance = self.fdecl.count("(") - self.fdecl.count(")")
|
|
|
|
|
|
|
|
def isready(self):
|
|
|
|
return self.balance == 0
|
|
|
|
|
|
|
|
def getLang(self, line):
|
|
|
|
if line.startswith(".. ocv:function::"):
|
|
|
|
return "C++"
|
|
|
|
if line.startswith(".. ocv:cfunction::"):
|
|
|
|
return "C"
|
|
|
|
if line.startswith(".. ocv:pyfunction::"):
|
|
|
|
return "Python2"
|
|
|
|
if line.startswith(".. ocv:pyoldfunction::"):
|
|
|
|
return "Python1"
|
|
|
|
if line.startswith(".. ocv:jfunction::"):
|
|
|
|
return "Java"
|
|
|
|
return None
|
|
|
|
|
|
|
|
def hasDeclaration(self, line):
|
|
|
|
return self.getLang(line) is not None
|
|
|
|
|
|
|
|
class ParamParser(object):
|
|
|
|
def __init__(self, line=None):
|
|
|
|
if line is None:
|
|
|
|
self.prefix = ""
|
|
|
|
self.name = ""
|
|
|
|
self.comment = ""
|
|
|
|
self.active = False
|
|
|
|
return
|
|
|
|
offset = line.find(":param")
|
|
|
|
assert offset > 0
|
|
|
|
self.prefix = line[:offset]
|
|
|
|
assert self.prefix==" "*len(self.prefix), ":param definition should be prefixed with spaces"
|
|
|
|
line = line[offset + 6:].lstrip()
|
|
|
|
name_end = line.find(":")
|
|
|
|
assert name_end > 0
|
|
|
|
self.name = line[:name_end]
|
|
|
|
self.comment = line[name_end+1:].lstrip()
|
|
|
|
self.active = True
|
|
|
|
|
|
|
|
def append(self, line):
|
|
|
|
assert self.active
|
|
|
|
if (self.hasDeclaration(line)):
|
|
|
|
self.active = False
|
|
|
|
elif line.startswith(self.prefix) or not line:
|
|
|
|
self.comment += "\n" + line.lstrip()
|
|
|
|
else:
|
|
|
|
self.active = False
|
|
|
|
|
|
|
|
def hasDeclaration(self, line):
|
|
|
|
return line.lstrip().startswith(":param")
|
|
|
|
|
|
|
|
class RstParser(object):
|
|
|
|
def __init__(self, cpp_parser):
|
|
|
|
self.cpp_parser = cpp_parser
|
|
|
|
self.definitions = {}
|
|
|
|
self.sections_parsed = 0
|
|
|
|
self.sections_total = 0
|
|
|
|
self.sections_skipped = 0
|
|
|
|
|
|
|
|
def parse(self, module_name, module_path=None):
|
|
|
|
if module_path is None:
|
|
|
|
module_path = "../" + module_name
|
|
|
|
doclist = glob.glob(os.path.join(module_path,"doc/*.rst"))
|
|
|
|
for doc in doclist:
|
|
|
|
self.parse_rst_file(module_name, doc)
|
|
|
|
|
|
|
|
def parse_section_safe(self, module_name, section_name, file_name, lineno, lines):
|
|
|
|
try:
|
|
|
|
self.parse_section(module_name, section_name, file_name, lineno, lines)
|
|
|
|
except AssertionError, args:
|
|
|
|
if show_errors:
|
|
|
|
print >> sys.stderr, "RST parser error: assertion in \"%s\" File: %s (line %s)" % (section_name, file_name, lineno)
|
|
|
|
print >> sys.stderr, " Details: %s" % args
|
|
|
|
|
|
|
|
def parse_section(self, module_name, section_name, file_name, lineno, lines):
|
|
|
|
self.sections_total += 1
|
|
|
|
# skip sections having whitespace in name
|
|
|
|
if section_name.find(" ") >= 0 and section_name.find("::operator") < 0:
|
|
|
|
if show_errors:
|
|
|
|
print "SKIPPED: \"%s\" File: %s (line %s)" % (section_name, file_name, lineno)
|
|
|
|
self.sections_skipped += 1
|
|
|
|
return
|
|
|
|
|
|
|
|
func = {}
|
|
|
|
func["name"] = section_name
|
|
|
|
func["file"] = file_name
|
|
|
|
func["line"] = lineno
|
|
|
|
func["module"] = module_name
|
|
|
|
|
|
|
|
# parse section name
|
|
|
|
section_name = self.parse_namespace(func, section_name)
|
|
|
|
class_separator_idx = section_name.find("::")
|
|
|
|
if class_separator_idx > 0:
|
|
|
|
func["class"] = section_name[:class_separator_idx]
|
|
|
|
func["method"] = section_name[class_separator_idx+2:]
|
|
|
|
else:
|
|
|
|
func["method"] = section_name
|
|
|
|
|
|
|
|
capturing_seealso = False
|
|
|
|
skip_code_lines = False
|
|
|
|
expected_brief = True
|
|
|
|
fdecl = DeclarationParser()
|
|
|
|
pdecl = ParamParser()
|
|
|
|
|
|
|
|
for l in lines:
|
|
|
|
# read tail of function/method declaration if needed
|
|
|
|
if not fdecl.isready():
|
|
|
|
fdecl.append(ll)
|
|
|
|
if fdecl.isready():
|
|
|
|
self.add_new_fdecl(func, fdecl)
|
|
|
|
continue
|
|
|
|
|
|
|
|
# continue capture seealso
|
|
|
|
if capturing_seealso:
|
|
|
|
if not l or l.startswith(" "):
|
|
|
|
seealso = func.get("seealso",[])
|
|
|
|
seealso.extend(l.split(","))
|
|
|
|
func["seealso"] = seealso
|
|
|
|
continue
|
|
|
|
else:
|
|
|
|
capturing_seealso = False
|
|
|
|
|
|
|
|
ll = l.strip()
|
|
|
|
if ll == "..":
|
|
|
|
expected_brief = False
|
|
|
|
skip_code_lines = False
|
|
|
|
continue
|
|
|
|
|
|
|
|
# skip lines if line-skipping mode is activated
|
|
|
|
if skip_code_lines:
|
|
|
|
if not l or l.startswith(" "):
|
|
|
|
continue
|
|
|
|
else:
|
|
|
|
skip_code_lines = False
|
|
|
|
|
|
|
|
if ll.startswith(".. code-block::") or ll.startswith(".. image::"):
|
|
|
|
skip_code_lines = True
|
|
|
|
continue
|
|
|
|
|
|
|
|
# todo: parse structure members; skip them for now
|
|
|
|
if ll.startswith(".. ocv:member::"):
|
|
|
|
skip_code_lines = True
|
|
|
|
continue
|
|
|
|
|
|
|
|
#ignore references (todo: collect them)
|
|
|
|
if l.startswith(".. ["):
|
|
|
|
continue
|
|
|
|
|
|
|
|
if ll.startswith(".. "):
|
|
|
|
expected_brief = False
|
|
|
|
elif ll.endswith("::"):
|
|
|
|
# turn on line-skipping mode for code fragments
|
|
|
|
skip_code_lines = True
|
|
|
|
ll = ll[:len(ll)-2]
|
|
|
|
|
|
|
|
# continue param parsing (process params after processing .. at the beginning of the line and :: at the end)
|
|
|
|
if pdecl.active:
|
|
|
|
pdecl.append(l)
|
|
|
|
if pdecl.active:
|
|
|
|
continue
|
|
|
|
else:
|
|
|
|
self.add_new_pdecl(func, pdecl)
|
|
|
|
# do not continue - current line can contain next parameter definition
|
|
|
|
|
|
|
|
# parse ".. seealso::" blocks
|
|
|
|
if ll.startswith(".. seealso::"):
|
|
|
|
if ll.endswith(".. seealso::"):
|
|
|
|
capturing_seealso = True
|
|
|
|
else:
|
|
|
|
seealso = func.get("seealso",[])
|
|
|
|
seealso.extend(ll[ll.find("::")+2:].split(","))
|
|
|
|
func["seealso"] = seealso
|
|
|
|
continue
|
|
|
|
|
|
|
|
# skip ".. index::"
|
|
|
|
if ll.startswith(".. index::"):
|
|
|
|
continue
|
|
|
|
|
|
|
|
# parse class & struct definitions
|
|
|
|
if ll.startswith(".. ocv:class::"):
|
|
|
|
func["class"] = ll[ll.find("::")+2:].strip()
|
|
|
|
if "method" in func:
|
|
|
|
del func["method"]
|
|
|
|
func["isclass"] = True
|
|
|
|
expected_brief = True
|
|
|
|
continue
|
|
|
|
|
|
|
|
if ll.startswith(".. ocv:struct::"):
|
|
|
|
func["class"] = ll[ll.find("::")+2:].strip()
|
|
|
|
if "method" in func:
|
|
|
|
del func["method"]
|
|
|
|
func["isstruct"] = True
|
|
|
|
expected_brief = True
|
|
|
|
continue
|
|
|
|
|
|
|
|
# parse function/method definitions
|
|
|
|
if fdecl.hasDeclaration(ll):
|
|
|
|
fdecl = DeclarationParser(ll)
|
|
|
|
if fdecl.isready():
|
|
|
|
self.add_new_fdecl(func, fdecl)
|
|
|
|
continue
|
|
|
|
|
|
|
|
# parse parameters
|
|
|
|
if pdecl.hasDeclaration(l):
|
|
|
|
pdecl = ParamParser(l)
|
|
|
|
continue
|
|
|
|
|
|
|
|
# record brief description
|
|
|
|
if expected_brief:
|
|
|
|
func["brief"] = func.get("brief", "") + "\n" + ll
|
|
|
|
if skip_code_lines:
|
|
|
|
expected_brief = False # force end brief if code block begins
|
|
|
|
continue
|
|
|
|
|
|
|
|
# record other lines as long description
|
|
|
|
func["long"] = func.get("long", "") + "\n" + ll
|
|
|
|
if skip_code_lines:
|
|
|
|
func["long"] = func.get("long", "") + "\n"
|
|
|
|
# endfor l in lines
|
|
|
|
|
|
|
|
if fdecl.balance != 0:
|
|
|
|
if show_errors:
|
|
|
|
print >> sys.stderr, "RST parser error: invalid parentheses balance in \"%s\" File: %s (line %s)" % (section_name, file_name, lineno)
|
|
|
|
return
|
|
|
|
|
|
|
|
# save last parameter if needed
|
|
|
|
if pdecl.active:
|
|
|
|
self.add_new_pdecl(func, pdecl)
|
|
|
|
|
|
|
|
# add definition to list
|
|
|
|
func = self.normalize(func)
|
|
|
|
if self.validate(func):
|
|
|
|
self.definitions[func["name"]] = func
|
|
|
|
self.sections_parsed += 1
|
|
|
|
if verbose:
|
|
|
|
self.print_info(func)
|
|
|
|
elif func:
|
|
|
|
if show_errors:
|
|
|
|
self.print_info(func, True, sys.stderr)
|
|
|
|
|
|
|
|
def parse_rst_file(self, module_name, doc):
|
|
|
|
doc = os.path.abspath(doc)
|
|
|
|
lineno = 0
|
|
|
|
whitespace_warnings = 0
|
|
|
|
max_whitespace_warnings = 10
|
|
|
|
|
|
|
|
lines = []
|
|
|
|
flineno = 0
|
|
|
|
fname = ""
|
|
|
|
prev_line = None
|
|
|
|
|
|
|
|
df = open(doc, "rt")
|
|
|
|
for l in df.readlines():
|
|
|
|
lineno += 1
|
|
|
|
# handle tabs
|
|
|
|
if l.find("\t") >= 0:
|
|
|
|
whitespace_warnings += 1
|
|
|
|
if whitespace_warnings <= max_whitespace_warnings and show_warnings:
|
|
|
|
print >> sys.stderr, "RST parser warning: tab symbol instead of space is used at file %s (line %s)" % (doc, lineno)
|
|
|
|
l = l.replace("\t", " ")
|
|
|
|
|
|
|
|
# handle first line
|
|
|
|
if prev_line == None:
|
|
|
|
prev_line = l.rstrip()
|
|
|
|
continue
|
|
|
|
|
|
|
|
ll = l.rstrip()
|
|
|
|
if len(prev_line) > 0 and len(ll) >= len(prev_line) and ll == "-" * len(ll):
|
|
|
|
# new function candidate
|
|
|
|
if len(lines) > 1:
|
|
|
|
self.parse_section_safe(module_name, fname, doc, flineno, lines[:len(lines)-1])
|
|
|
|
lines = []
|
|
|
|
flineno = lineno-1
|
|
|
|
fname = prev_line.strip()
|
|
|
|
elif flineno > 0:
|
|
|
|
lines.append(ll)
|
|
|
|
prev_line = ll
|
|
|
|
df.close()
|
|
|
|
|
|
|
|
# don't forget about the last function section in file!!!
|
|
|
|
if len(lines) > 1:
|
|
|
|
self.parse_section_safe(module_name, fname, doc, flineno, lines)
|
|
|
|
|
|
|
|
def parse_namespace(self, func, section_name):
|
|
|
|
known_namespaces = ["cv", "gpu", "flann"]
|
|
|
|
l = section_name.strip()
|
|
|
|
for namespace in known_namespaces:
|
|
|
|
if l.startswith(namespace + "::"):
|
|
|
|
func["namespace"] = namespace
|
|
|
|
return l[len(namespace)+2:]
|
|
|
|
return section_name
|
|
|
|
|
|
|
|
def add_new_fdecl(self, func, decl):
|
|
|
|
decls = func.get("decls",[])
|
|
|
|
if (decl.lang == "C++" or decl.lang == "C"):
|
|
|
|
rst_decl = self.cpp_parser.parse_func_decl_no_wrap(decl.fdecl)
|
|
|
|
decls.append( (decl.lang, decl.fdecl, rst_decl) )
|
|
|
|
else:
|
|
|
|
decls.append( (decl.lang, decl.fdecl) )
|
|
|
|
func["decls"] = decls
|
|
|
|
|
|
|
|
def add_new_pdecl(self, func, decl):
|
|
|
|
params = func.get("params",{})
|
|
|
|
if decl.name in params:
|
|
|
|
if show_errors:
|
|
|
|
#check black_list
|
|
|
|
if decl.name not in params_blacklist.get(func["name"], []):
|
|
|
|
print >> sys.stderr, "RST parser error: redefinition of parameter \"%s\" in \"%s\" File: %s (line %s)" \
|
|
|
|
% (decl.name, func["name"], func["file"], func["line"])
|
|
|
|
else:
|
|
|
|
params[decl.name] = decl.comment
|
|
|
|
func["params"] = params
|
|
|
|
|
|
|
|
def print_info(self, func, skipped=False, out = sys.stdout):
|
|
|
|
print >> out
|
|
|
|
if skipped:
|
|
|
|
print >> out, "SKIPPED DEFINITION:"
|
|
|
|
print >> out, "name: %s" % (func.get("name","~empty~"))
|
|
|
|
print >> out, "file: %s (line %s)" % (func.get("file","~empty~"), func.get("line","~empty~"))
|
|
|
|
print >> out, "is class: %s" % func.get("isclass",False)
|
|
|
|
print >> out, "is struct: %s" % func.get("isstruct",False)
|
|
|
|
print >> out, "module: %s" % func.get("module","~unknown~")
|
|
|
|
print >> out, "namespace: %s" % func.get("namespace", "~empty~")
|
|
|
|
print >> out, "class: %s" % (func.get("class","~empty~"))
|
|
|
|
print >> out, "method: %s" % (func.get("method","~empty~"))
|
|
|
|
print >> out, "brief: %s" % (func.get("brief","~empty~"))
|
|
|
|
if "decls" in func:
|
|
|
|
print >> out, "declarations:"
|
|
|
|
for d in func["decls"]:
|
|
|
|
print >> out, " %7s: %s" % (d[0], re.sub(r"[ ]+", " ", d[1]))
|
|
|
|
if "seealso" in func:
|
|
|
|
print >> out, "seealso: ", func["seealso"]
|
|
|
|
if "params" in func:
|
|
|
|
print >> out, "parameters:"
|
|
|
|
for name, comment in func["params"].items():
|
|
|
|
print >> out, "%23s: %s" % (name, comment)
|
|
|
|
print >> out, "long: %s" % (func.get("long","~empty~"))
|
|
|
|
print >> out
|
|
|
|
|
|
|
|
def validate(self, func):
|
|
|
|
if func.get("decls",None) is None:
|
|
|
|
if not func.get("isclass",False) and not func.get("isstruct",False):
|
|
|
|
return False
|
|
|
|
if func["name"] in self.definitions:
|
|
|
|
if show_errors:
|
|
|
|
print >> sys.stderr, "RST parser error: \"%s\" from file: %s (line %s) is already documented in file: %s (line %s)" \
|
|
|
|
% (func["name"], func["file"], func["line"], self.definitions[func["name"]]["file"], self.definitions[func["name"]]["line"])
|
|
|
|
return False
|
|
|
|
return self.validateParams(func)
|
|
|
|
|
|
|
|
def validateParams(self, func):
|
|
|
|
documentedParams = func.get("params",{}).keys()
|
|
|
|
params = []
|
|
|
|
|
|
|
|
for decl in func.get("decls", []):
|
|
|
|
if len(decl) > 2:
|
|
|
|
args = decl[2][3] # decl[2] -> [ funcname, return_ctype, [modifiers], [args] ]
|
|
|
|
for arg in args:
|
|
|
|
# arg -> [ ctype, name, def val, [mod], argno ]
|
|
|
|
if arg[0] != "...":
|
|
|
|
params.append(arg[1])
|
|
|
|
params = list(set(params))#unique
|
|
|
|
|
|
|
|
# 1. all params are documented
|
|
|
|
for p in params:
|
|
|
|
if p not in documentedParams and show_warnings:
|
|
|
|
print >> sys.stderr, "RST parser warning: parameter \"%s\" of \"%s\" is undocumented. File: %s (line %s)" % (p, func["name"], func["file"], func["line"])
|
|
|
|
|
|
|
|
# 2. only real params are documented
|
|
|
|
for p in documentedParams:
|
|
|
|
if p not in params and show_warnings:
|
|
|
|
if p not in params_blacklist.get(func["name"], []):
|
|
|
|
print >> sys.stderr, "RST parser warning: unexisting parameter \"%s\" of \"%s\" is documented. File: %s (line %s)" % (p, func["name"], func["file"], func["line"])
|
|
|
|
return True
|
|
|
|
|
|
|
|
def normalize(self, func):
|
|
|
|
if not func:
|
|
|
|
return func
|
|
|
|
fnname = func["name"]
|
|
|
|
fnname = self.normalizeText(fnname)
|
|
|
|
fnname = re.sub(r'_\?D$', "_nD", fnname) # tailing _?D can be mapped to _nD
|
|
|
|
fnname = re.sub(r'\?D$', "ND", fnname) # tailing ?D can be mapped to ND
|
|
|
|
fnname = re.sub(r'\(s\)$', "s", fnname) # tailing (s) can be mapped to s
|
|
|
|
func["name"] = fnname
|
|
|
|
if "method" in func:
|
|
|
|
func["method"] = self.normalizeText(func["method"])
|
|
|
|
if "class" in func:
|
|
|
|
func["class"] = self.normalizeText(func["class"])
|
|
|
|
if "brief" in func:
|
|
|
|
func["brief"] = self.normalizeText(func.get("brief",None))
|
|
|
|
if not func["brief"]:
|
|
|
|
del func["brief"]
|
|
|
|
if "long" in func:
|
|
|
|
func["long"] = self.normalizeText(func.get("long",None))
|
|
|
|
if not func["long"]:
|
|
|
|
del func["long"]
|
|
|
|
if "decls" in func:
|
|
|
|
func["decls"].sort()
|
|
|
|
if "params" in func:
|
|
|
|
params = {}
|
|
|
|
for name, comment in func["params"].items():
|
|
|
|
cmt = self.normalizeText(comment)
|
|
|
|
if cmt:
|
|
|
|
params[name] = cmt
|
|
|
|
# expand some wellknown params
|
|
|
|
pmap = params_mapping.get(fnname)
|
|
|
|
if pmap:
|
|
|
|
for name, alias in pmap.items():
|
|
|
|
params[name] = params[alias]
|
|
|
|
func["params"] = params
|
|
|
|
if "seealso" in func:
|
|
|
|
seealso = []
|
|
|
|
for see in func["seealso"]:
|
|
|
|
item = self.normalizeText(see.rstrip(".")).strip("\"")
|
|
|
|
if item and (item.find(" ") < 0 or item.find("::operator") > 0):
|
|
|
|
seealso.append(item)
|
|
|
|
func["seealso"] = list(set(seealso))
|
|
|
|
if not func["seealso"]:
|
|
|
|
del func["seealso"]
|
|
|
|
|
|
|
|
# special case for old C functions - section name should omit "cv" prefix
|
|
|
|
if not func.get("isclass",False) and not func.get("isstruct",False):
|
|
|
|
self.fixOldCFunctionName(func)
|
|
|
|
return func
|
|
|
|
|
|
|
|
def fixOldCFunctionName(self, func):
|
|
|
|
if not "decls" in func:
|
|
|
|
return
|
|
|
|
fname = None
|
|
|
|
for decl in func["decls"]:
|
|
|
|
if decl[0] != "C" and decl[0] != "Python1":
|
|
|
|
return
|
|
|
|
if decl[0] == "C":
|
|
|
|
fname = decl[2][0]
|
|
|
|
if fname is None:
|
|
|
|
return
|
|
|
|
|
|
|
|
fname = fname.replace(".", "::")
|
|
|
|
if fname.startswith("cv::cv"):
|
|
|
|
if fname[6:] == func.get("name", ""):
|
|
|
|
func["name"] = fname[4:]
|
|
|
|
func["method"] = fname[4:]
|
|
|
|
elif show_warnings:
|
|
|
|
print >> sys.stderr, "\"%s\" - section name is \"%s\" instead of \"%s\". File: %s (line %s)" % (fname, func["name"], fname[6:], func["file"], func["line"])
|
|
|
|
#self.print_info(func)
|
|
|
|
|
|
|
|
def normalizeText(self, s):
|
|
|
|
if s is None:
|
|
|
|
return s
|
|
|
|
s = re.sub(r"\.\. math::[ ]*\n+(.*?)(\n[ ]*\n|$)", mathReplace2, s)
|
|
|
|
s = re.sub(r":math:`([^`]+?)`", mathReplace, s)
|
|
|
|
s = re.sub(r" *:sup:", "^", s)
|
|
|
|
|
|
|
|
s = s.replace(":ocv:class:", "")
|
|
|
|
s = s.replace(":ocv:struct:", "")
|
|
|
|
s = s.replace(":ocv:func:", "")
|
|
|
|
s = s.replace(":ocv:cfunc:","")
|
|
|
|
s = s.replace(":c:type:", "")
|
|
|
|
s = s.replace(":c:func:", "")
|
|
|
|
s = s.replace(":ref:", "")
|
|
|
|
s = s.replace(":math:", "")
|
|
|
|
s = s.replace(":func:", "")
|
|
|
|
|
|
|
|
s = s.replace("]_", "]")
|
|
|
|
s = s.replace(".. note::", "Note:")
|
|
|
|
s = s.replace(".. table::", "")
|
|
|
|
s = s.replace(".. ocv:function::", "")
|
|
|
|
s = s.replace(".. ocv:cfunction::", "")
|
|
|
|
|
|
|
|
# remove ".. identifier:" lines
|
|
|
|
s = re.sub(r"(^|\n)\.\. [a-zA-Z_0-9]+(::[a-zA-Z_0-9]+)?:(\n|$)", "\n ", s)
|
|
|
|
# unwrap urls
|
|
|
|
s = re.sub(r"`([^`<]+ )<(https?://[^>]+)>`_", "\\1(\\2)", s)
|
|
|
|
# remove tailing ::
|
|
|
|
s = re.sub(r"::(\n|$)", "\\1", s)
|
|
|
|
|
|
|
|
# normalize line endings
|
|
|
|
s = re.sub(r"\r\n", "\n", s)
|
|
|
|
# remove extra line breaks before/after _ or ,
|
|
|
|
s = re.sub(r"\n[ ]*([_,])\n", r"\1 ", s)
|
|
|
|
# remove extra line breaks after `
|
|
|
|
#s = re.sub(r"`\n", "` ", s)
|
|
|
|
# remove extra space after ( and before .,)
|
|
|
|
s = re.sub(r"\([\n ]+", "(", s)
|
|
|
|
s = re.sub(r"[\n ]+(\.|,|\))", "\\1", s)
|
|
|
|
# remove extra line breaks after ".. note::"
|
|
|
|
s = re.sub(r"\.\. note::\n+", ".. note:: ", s)
|
|
|
|
# remove extra line breaks before *
|
|
|
|
s = re.sub(r"\n+\*", "\n*", s)
|
|
|
|
# remove extra line breaks after *
|
|
|
|
s = re.sub(r"\n\*\n+", "\n* ", s)
|
|
|
|
# remove extra line breaks before #.
|
|
|
|
s = re.sub(r"\n+#\.", "\n#.", s)
|
|
|
|
# remove extra line breaks after #.
|
|
|
|
s = re.sub(r"\n#\.\n+", "\n#. ", s)
|
|
|
|
# remove extra line breaks before `
|
|
|
|
#s = re.sub(r"\n[ ]*`", " `", s)
|
|
|
|
# remove trailing whitespaces
|
|
|
|
s = re.sub(r"[ ]+$", "", s)
|
|
|
|
# remove .. for references
|
|
|
|
#s = re.sub(r"\.\. \[", "[", s)
|
|
|
|
# unescape
|
|
|
|
s = re.sub(r"\\(.)", "\\1", s)
|
|
|
|
|
|
|
|
# remove whitespace before .
|
|
|
|
s = re.sub(r"[ ]+\.", ".", s)
|
|
|
|
# remove tailing whitespace
|
|
|
|
s = re.sub(r" +(\n|$)", "\\1", s)
|
|
|
|
# remove leading whitespace
|
|
|
|
s = re.sub(r"(^|\n) +", "\\1", s)
|
|
|
|
# compress line breaks
|
|
|
|
s = re.sub(r"\n\n+", "\n\n", s)
|
|
|
|
# remove other newlines
|
|
|
|
s = re.sub(r"([^.\n\\=])\n([^*#\n]|\*[^ ])", "\\1 \\2", s)
|
|
|
|
# compress whitespace
|
|
|
|
s = re.sub(r" +", " ", s)
|
|
|
|
|
|
|
|
# restore math
|
|
|
|
s = re.sub(r" *<BR> *","\n", s)
|
|
|
|
|
|
|
|
# remove extra space before .
|
|
|
|
s = re.sub(r"[\n ]+\.", ".", s)
|
|
|
|
|
|
|
|
s = s.replace("**", "")
|
|
|
|
s = s.replace("``", "\"")
|
|
|
|
s = s.replace("`", "\"")
|
|
|
|
s = s.replace("\"\"", "\"")
|
|
|
|
|
|
|
|
s = s.strip()
|
|
|
|
return s
|
|
|
|
|
|
|
|
def printSummary(self):
|
|
|
|
print
|
|
|
|
print "RST Parser Summary:"
|
|
|
|
print " Total sections: %s" % self.sections_total
|
|
|
|
print " Skipped sections: %s" % self.sections_skipped
|
|
|
|
print " Parsed sections: %s" % self.sections_parsed
|
|
|
|
print " Invalid sections: %s" % (self.sections_total - self.sections_parsed - self.sections_skipped)
|
|
|
|
|
|
|
|
# statistic by language
|
|
|
|
stat = {}
|
|
|
|
classes = 0
|
|
|
|
structs = 0
|
|
|
|
for name, d in self.definitions.items():
|
|
|
|
if d.get("isclass", False):
|
|
|
|
classes += 1
|
|
|
|
elif d.get("isstruct", False):
|
|
|
|
structs += 1
|
|
|
|
else:
|
|
|
|
for decl in d.get("decls",[]):
|
|
|
|
stat[decl[0]] = stat.get(decl[0],0) + 1
|
|
|
|
|
|
|
|
print
|
|
|
|
print " classes documented: %s" % classes
|
|
|
|
print " structs documented: %s" % structs
|
|
|
|
for lang in sorted(stat.items()):
|
|
|
|
print " %7s functions documented: %s" % lang
|
|
|
|
print
|
|
|
|
|
|
|
|
def mathReplace2(match):
|
|
|
|
m = mathReplace(match)
|
|
|
|
#print "%s ===> %s" % (match.group(0), m)
|
|
|
|
return "\n\n"+m+"<BR><BR>"
|
|
|
|
|
|
|
|
def hdotsforReplace(match):
|
|
|
|
return '... '*int(match.group(1))
|
|
|
|
|
|
|
|
def matrixReplace(match):
|
|
|
|
m = match.group(2)
|
|
|
|
m = re.sub(r" *& *", " ", m)
|
|
|
|
return m
|
|
|
|
|
|
|
|
def mathReplace(match):
|
|
|
|
m = match.group(1)
|
|
|
|
|
|
|
|
m = m.replace("\n", "<BR>")
|
|
|
|
m = re.sub(r"\\text(tt|rm)?{(.*?)}", "\\2", m)
|
|
|
|
m = re.sub(r"\\mbox{(.*?)}", "\\1", m)
|
|
|
|
m = re.sub(r"\\mathrm{(.*?)}", "\\1", m)
|
|
|
|
m = re.sub(r"\\vecthree{(.*?)}{(.*?)}{(.*?)}", "[\\1 \\2 \\3]", m)
|
|
|
|
m = re.sub(r"\\bar{(.*?)}", "\\1`", m)
|
|
|
|
m = re.sub(r"\\sqrt\[(\d)*\]{(.*?)}", "sqrt\\1(\\2)", m)
|
|
|
|
m = re.sub(r"\\sqrt{(.*?)}", "sqrt(\\1)", m)
|
|
|
|
m = re.sub(r"\\frac{(.*?)}{(.*?)}", "(\\1)/(\\2)", m)
|
|
|
|
m = re.sub(r"\\fork{(.*?)}{(.*?)}{(.*?)}{(.*?)}", "\\1 \\2; \\3 \\4", m)
|
|
|
|
m = re.sub(r"\\forkthree{(.*?)}{(.*?)}{(.*?)}{(.*?)}{(.*?)}{(.*?)}", "\\1 \\2; \\3 \\4; \\5 \\6", m)
|
|
|
|
m = re.sub(r"\\stackrel{(.*?)}{(.*?)}", "\\1 \\2", m)
|
|
|
|
m = re.sub(r"\\sum _{(.*?)}", "sum{by: \\1}", m)
|
|
|
|
|
|
|
|
m = re.sub(r" +", " ", m)
|
|
|
|
m = re.sub(r"\\begin{(?P<gtype>array|bmatrix)}(?:{[\|lcr\. ]+})? *(.*?)\\end{(?P=gtype)}", matrixReplace, m)
|
|
|
|
m = re.sub(r"\\hdotsfor{(\d+)}", hdotsforReplace, m)
|
|
|
|
m = re.sub(r"\\vecthreethree{(.*?)}{(.*?)}{(.*?)}{(.*?)}{(.*?)}{(.*?)}{(.*?)}{(.*?)}{(.*?)}", "<BR>|\\1 \\2 \\3|<BR>|\\4 \\5 \\6|<BR>|\\7 \\8 \\9|<BR>", m)
|
|
|
|
|
|
|
|
m = re.sub(r"\\left[ ]*\\lfloor[ ]*", "[", m)
|
|
|
|
m = re.sub(r"[ ]*\\right[ ]*\\rfloor", "]", m)
|
|
|
|
m = re.sub(r"\\left[ ]*\([ ]*", "(", m)
|
|
|
|
m = re.sub(r"[ ]*\\right[ ]*\)", ")", m)
|
|
|
|
m = re.sub(r"([^\\])\$", "\\1", m)
|
|
|
|
|
|
|
|
m = m.replace("\\times", "x")
|
|
|
|
m = m.replace("\\pm", "+-")
|
|
|
|
m = m.replace("\\cdot", "*")
|
|
|
|
m = m.replace("\\sim", "~")
|
|
|
|
m = m.replace("\\leftarrow", "<-")
|
|
|
|
m = m.replace("\\rightarrow", "->")
|
|
|
|
m = m.replace("\\leftrightarrow", "<->")
|
|
|
|
m = re.sub(r" *\\neg *", " !", m)
|
|
|
|
m = re.sub(r" *\\neq? *", " != ", m)
|
|
|
|
m = re.sub(r" *\\geq? *", " >= ", m)
|
|
|
|
m = re.sub(r" *\\leq? *", " <= ", m)
|
|
|
|
m = re.sub(r" *\\vee *", " V ", m)
|
|
|
|
m = re.sub(r" *\\oplus *", " (+) ", m)
|
|
|
|
m = re.sub(r" *\\mod *", " mod ", m)
|
|
|
|
m = re.sub(r"( *)\\partial *", "\\1d", m)
|
|
|
|
|
|
|
|
m = re.sub(r"( *)\\quad *", "\\1 ", m)
|
|
|
|
m = m.replace("\\,", " ")
|
|
|
|
m = m.replace("\\:", " ")
|
|
|
|
m = m.replace("\\;", " ")
|
|
|
|
m = m.replace("\\!", "")
|
|
|
|
|
|
|
|
m = m.replace("\\\\", "<BR>")
|
|
|
|
m = m.replace("\\wedge", "/\\\\")
|
|
|
|
m = re.sub(r"\\(.)", "\\1", m)
|
|
|
|
|
|
|
|
m = re.sub(r"\([ ]+", "(", m)
|
|
|
|
m = re.sub(r"[ ]+(\.|,|\))(<BR>| |$)", "\\1\\2", m)
|
|
|
|
m = re.sub(r" +\|[ ]+([a-zA-Z0-9_(])", " |\\1", m)
|
|
|
|
m = re.sub(r"([a-zA-Z0-9_)}])[ ]+(\(|\|)", "\\1\\2", m)
|
|
|
|
|
|
|
|
m = re.sub(r"{\((-?[a-zA-Z0-9_]+)\)}", "\\1", m)
|
|
|
|
m = re.sub(r"{(-?[a-zA-Z0-9_]+)}", "(\\1)", m)
|
|
|
|
m = re.sub(r"\(([0-9]+)\)", "\\1", m)
|
|
|
|
m = m.replace("{", "(")
|
|
|
|
m = m.replace("}", ")")
|
|
|
|
|
|
|
|
#print "%s ===> %s" % (match.group(0), m)
|
|
|
|
return m
|
|
|
|
|
|
|
|
if __name__ == "__main__":
|
|
|
|
if len(sys.argv) < 2:
|
|
|
|
print "Usage:\n", os.path.basename(sys.argv[0]), " <module path>"
|
|
|
|
exit(0)
|
|
|
|
|
|
|
|
if len(sys.argv) >= 3:
|
|
|
|
if sys.argv[2].lower() == "verbose":
|
|
|
|
verbose = True
|
|
|
|
|
|
|
|
rst_parser_dir = os.path.dirname(os.path.abspath(sys.argv[0]))
|
|
|
|
hdr_parser_path = os.path.join(rst_parser_dir, "../python/src2")
|
|
|
|
|
|
|
|
sys.path.append(hdr_parser_path)
|
|
|
|
import hdr_parser
|
|
|
|
|
|
|
|
module = sys.argv[1]
|
|
|
|
|
|
|
|
if module != "all" and not os.path.isdir(os.path.join(rst_parser_dir, "../" + module)):
|
|
|
|
print "Module \"" + module + "\" could not be found."
|
|
|
|
exit(1)
|
|
|
|
|
|
|
|
parser = RstParser(hdr_parser.CppHeaderParser())
|
|
|
|
|
|
|
|
if module == "all":
|
|
|
|
for m in allmodules:
|
|
|
|
parser.parse(m, os.path.join(rst_parser_dir, "../" + m))
|
|
|
|
else:
|
|
|
|
parser.parse(module, os.path.join(rst_parser_dir, "../" + module))
|
|
|
|
|
|
|
|
# summary
|
|
|
|
parser.printSummary()
|