Refactoring ...

This commit is contained in:
2016-12-29 18:22:05 +01:00
parent 0526386337
commit 65f3b16e6d
9 changed files with 1015 additions and 1 deletions

View File

@ -0,0 +1,412 @@
#cython: language_level=3
from libc.stdlib cimport malloc
from ..capi.obiview cimport Alias_column_pair_p, \
obi_new_view, \
obi_open_view, \
obi_save_and_close_view, \
obi_view_get_pointer_on_column_in_view, \
obi_view_delete_column, \
obi_view_create_column_alias
from ..capi.obidmscolumn cimport OBIDMS_column_p
from obitools3.utils cimport tobytes, \
bytes2str
cdef class View :
def __init__(self,dms,int __internalCall__):
if __internalCall__!=987654:
raise RuntimeError('OBIView constructor cannot be called directly')
self._dms = dms
self._pointer = NULL
def clone(self,
object view_name,
object comments=None):
cdef bytes view_name_b = tobytes(view_name)
cdef bytes comments_b
cdef View view = View(self._dms,
987654)
if comments is not None:
comments_b = tobytes(comments)
else:
comments_b = b''
view._pointer = obi_new_view(self._pointer.dms,
view_name_b,
self._pointer,
NULL,
comments_b)
if view._pointer == NULL :
raise RuntimeError("Error : Cannot clone view %s into view %s"
% (str(self.name),
bytes2str(view_name_b))
)
return view
@staticmethod
def new(DMS dms,
object view_name,
object comments=None):
cdef bytes view_name_b = tobytes(view_name)
cdef bytes comments_b
cdef str message
cdef View view = View(dms,
987654) # @DuplicatedSignature
if comments is not None:
comments_b = tobytes(comments)
else:
comments_b = b''
view._pointer = obi_new_view(dms._pointer,
view_name_b,
NULL,
NULL,
comments_b)
if view._pointer == NULL :
message = "Error : Cannot create view %s" % bytes2str(view_name_b)
raise RuntimeError(message)
return view
@staticmethod
def open(DMS dms, # @ReservedAssignment
object view_name):
cdef bytes view_name_b = tobytes(view_name)
cdef View view = View(dms,
987654) # @DuplicatedSignature
view._pointer = obi_open_view(dms._pointer,
view_name_b)
if view._pointer == NULL :
raise RuntimeError("Error : Cannot open view %s" % bytes2str(view_name_b))
return view
def close(self):
if (self._pointer != NULL
and obi_save_and_close_view(self._pointer) < 0) :
raise Exception("Problem closing view %s" %
bytes2str(self.name))
def __dealloc__(self):
self.close()
def __repr__(self) :
cdef str s = "{name:s}\n{comments:s}\n{line_count:d} lines\n".format(name = str(self.name),
comments = str(self.comments),
line_count = self.line_count)
#for column_name in self._columns :
# s = s + repr(self._columns[column_name]) + '\n'
return s
def keys(self):
cdef int i
cdef int nb_column = self._pointer.infos.column_count
cdef Alias_column_pair_p column_p = self._pointer.infos.column_references
for i in range(nb_column) :
col_alias = bytes2str(self._pointer.infos.column_references[i].alias)
yield col_alias
def get_column(self,
object column_name):
cdef bytes column_name_b = tobytes(column_name)
cdef OBIDMS_column_p* column_pp
cdef OBIDMS_column_p column_p
cdef Column column
cdef obitype_t column_type
column_pp = obi_view_get_pointer_on_column_in_view(self._pointer,
column_name_b)
if column_pp == NULL:
raise KeyError("Cannot access to column %s in view %s" % (
bytes2str(column_name_b),
bytes2str(self.name)
))
column_p = column_pp[0]
column_type = column_p.header.returned_data_type
column = DMS.get_column_class(column_type)(self)
column._pointer = column_pp
return column
cpdef delete_column(self,
object column_name) :
cdef bytes column_name_b = tobytes(column_name)
if obi_view_delete_column(self._pointer, column_name_b) < 0 :
raise Exception("Problem deleting column %s from a view",
bytes2str(column_name_b))
cpdef rename_column(self,
object current_name,
object new_name):
cdef Column column
cdef bytes current_name_b = tobytes(current_name)
cdef bytes new_name_b = tobytes(new_name)
if (obi_view_create_column_alias(self._pointer,
tobytes(current_name_b),
tobytes(new_name_b)) < 0) :
raise Exception("Problem in renaming column %s to %s" % (
bytes2str(current_name_b),
bytes2str(new_name_b)))
cpdef View_line_selection new_selection(self,list lines=None):
return View_line_selection(self,lines)
def __iter__(self):
# Iteration on each line of all columns
# Declarations
cdef index_t line_nb
cdef View_line line
# Yield each line
for line_nb in range(self.line_count) :
line = self[line_nb]
yield line
def __getitem__(self, object item) :
if type(item) == str :
return (self._columns)[item]
elif type(item) == int :
return View_line(self, item)
def __contains__(self, str column_name):
return (column_name in self._columns)
def __len__(self):
return(self.line_count)
def __str__(self) :
cdef View_line line
cdef str to_print
to_print = ""
for line in self :
to_print = to_print + str(line) + "\n"
return to_print
@property
def dms(self):
return self._dms
# line_count property getter
@property
def line_count(self):
return self._pointer.infos.line_count
# name property getter
@property
def name(self):
return <bytes> self._pointer.infos.name
# view type property getter
@property
def type(self): # @ReservedAssignment
return bytes2str(self._pointer.infos.view_type)
# comments property getter
@property
def comments(self):
return <bytes> self._pointer.infos.comments
# TODO setter that concatenates new comments?
cdef class Line_selection(list):
def __init__(self, View view, lines=None) :
if view._pointer == NULL:
raise Exception("Error: trying to create a line selection with an invalidated view")
self._view = view
self._view_name = view.name
if lines is not None:
self.extend(lines)
def extend(self, iterable):
cdef index_t i
cdef index_t max_i = self._view.line_count
for i in iterable:
if i > max_i:
raise RuntimeError("Error: trying to select line %d beyond the line count %d of view %s" %
(i,
max_i,
self._view_name)
)
list.append(self,i)
def append(self, index_t idx) :
if idx >= self._view.line_count :
raise IndexError("Error: trying to select line %d beyond the line count %d of view %s" %
(idx,
self._view.line_count,
bytes2str(self.name))
)
list.append(self,idx)
cdef index_t* __build_binary_list__(self):
cdef index_t* line_selection_p = NULL
cdef int i
cdef size_t l_selection = len(self)
line_selection_p = <index_t*> malloc((l_selection + 1) * sizeof(index_t)) # +1 for the -1 flagging the end of the array
for i in range(l_selection) :
line_selection_p[i] = self[i]
line_selection_p[l_selection] = -1 # flagging the end of the array
return line_selection_p
cpdef View materialize(self,
object view_name,
object comments=""):
cdef View view = View(987654)
cdef bytes view_name_b=tobytes(view_name)
view._pointer = obi_new_view(self._view._pointer.dms,
view_name_b,
self._view._pointer,
self.__build_binary_list__(),
tobytes(comments))
if view._pointer == NULL :
raise RuntimeError("Error : Cannot clone view %s into view %s"
% (str(self.name),
view_name)
)
return view
cdef class Line :
def __init__(self, View view, index_t line_nb) :
self._index = line_nb
self._view = view
def __getitem__(self, str column_name) :
return ((self._view)._columns)[column_name][self._index]
# def __setitem__(self, str column_name, object value):
# # TODO detect multiple elements (dict type)? put somewhere else? but more risky (in get)
# # TODO OBI_QUAL ?
# cdef type value_type
# cdef str value_obitype
# cdef bytes value_b
#
# if column_name not in self._view :
# if value == None :
# raise Exception("Trying to create a column from a None value (can't guess type)")
# value_type = type(value)
# if value_type == int :
# value_obitype = 'OBI_INT'
# elif value_type == float :
# value_obitype = 'OBI_FLOAT'
# elif value_type == bool :
# value_obitype = 'OBI_BOOL'
# elif value_type == str or value_type == bytes :
# if value_type == str :
# value_b = str2bytes(value)
# else :
# value_b = value
# if only_ATGC(value_b) : # TODO detect IUPAC
# value_obitype = 'OBI_SEQ'
# elif len(value) == 1 :
# value_obitype = 'OBI_CHAR'
# elif (len(value) > 1) :
# value_obitype = 'OBI_STR'
# else :
# raise Exception("Could not guess the type of a value to create a new column")
# self._view.add_column(column_name, type=value_obitype)
#
# (((self._view)._columns)[column_name]).set_line(self._index, value)
#
# def __iter__(self):
# for column_name in ((self._view)._columns) :
# yield column_name
#
# def __contains__(self, str column_name):
# return (column_name in self._view._columns)
def __repr__(self):
cdef dict line
cdef str column_name
line = {}
# for column_name in self._view._columns :
# line[column_name] = self[column_name]
return str(line)
# cpdef dict get_view_infos(self, str view_name) :
#
# cdef Obiview_infos_p view_infos_p
# cdef dict view_infos_d
# cdef Alias_column_pair_p column_refs
# cdef int i, j
# cdef str column_name
#
# view_infos_p = obi_view_map_file(self._pointer,
# tobytes(view_name))
# view_infos_d = {}
# view_infos_d["name"] = bytes2str(view_infos_p.name)
# view_infos_d["comments"] = bytes2str(view_infos_p.comments)
# view_infos_d["view_type"] = bytes2str(view_infos_p.view_type)
# view_infos_d["column_count"] = <int> view_infos_p.column_count
# view_infos_d["line_count"] = <int> view_infos_p.line_count
# view_infos_d["created_from"] = bytes2str(view_infos_p.created_from)
# view_infos_d["creation_date"] = bytes2str(obi_format_date(view_infos_p.creation_date))
# if (view_infos_p.all_lines) :
# view_infos_d["line_selection"] = None
# else :
# view_infos_d["line_selection"] = {}
# view_infos_d["line_selection"]["column_name"] = bytes2str((view_infos_p.line_selection).column_name)
# view_infos_d["line_selection"]["version"] = <int> (view_infos_p.line_selection).version
# view_infos_d["column_references"] = {}
# column_references = view_infos_p.column_references
# for j in range(view_infos_d["column_count"]) :
# column_name = bytes2str((column_references[j]).alias)
# view_infos_d["column_references"][column_name] = {}
# view_infos_d["column_references"][column_name]["original_name"] = bytes2str((column_references[j]).column_refs.column_name)
# view_infos_d["column_references"][column_name]["version"] = (column_references[j]).column_refs.version
#
# obi_view_unmap_file(self._pointer, view_infos_p)
#
# return view_infos_d