from jyquickhelper import add_notebook_menu
add_notebook_menu()
On reprend la fonction donnée dans l'énoncé.
def distance_edition(mot1, mot2):
dist = { (-1,-1): 0 }
for i,c in enumerate(mot1) :
dist[i,-1] = dist[i-1,-1] + 1
dist[-1,i] = dist[-1,i-1] + 1
for j,d in enumerate(mot2) :
opt = [ ]
if (i-1,j) in dist :
x = dist[i-1,j] + 1
opt.append(x)
if (i,j-1) in dist :
x = dist[i,j-1] + 1
opt.append(x)
if (i-1,j-1) in dist :
x = dist[i-1,j-1] + (1 if c != d else 0)
opt.append(x)
dist[i,j] = min(opt)
return dist[len(mot1)-1,len(mot2)-1]
%timeit distance_edition("idstzance","distances")
188 µs ± 28.6 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)
Les préliminaires :
%load_ext cython
Puis :
%%cython --annotate
cimport cython
def cidistance_edition(str mot1, str mot2):
cdef int dist [500][500]
cdef int cost, c
cdef int l1 = len(mot1)
cdef int l2 = len(mot2)
dist[0][0] = 0
for i in range(l1):
dist[i+1][0] = dist[i][0] + 1
dist[0][i+1] = dist[0][i] + 1
for j in range(l2):
cost = dist[i][j+1] + 1
c = dist[i+1][j] + 1
if c < cost : cost = c
c = dist[i][j]
if mot1[i] != mot2[j] : c += 1
if c < cost : cost = c
dist[i+1][j+1] = cost
cost = dist[l1][l2]
return cost
Generated by Cython 0.29.21
Yellow lines hint at Python interaction.
Click on a line that starts with a "+
" to see the C code that Cython generated for it.
01: cimport cython
02:
+03: def cidistance_edition(str mot1, str mot2):
/* Python wrapper */ static PyObject *__pyx_pw_46_cython_magic_f072b3f10e4cb6a87b39cd12da494e91_1cidistance_edition(PyObject *__pyx_self, PyObject *__pyx_args, PyObject *__pyx_kwds); /*proto*/ static PyMethodDef __pyx_mdef_46_cython_magic_f072b3f10e4cb6a87b39cd12da494e91_1cidistance_edition = {"cidistance_edition", (PyCFunction)(void*)(PyCFunctionWithKeywords)__pyx_pw_46_cython_magic_f072b3f10e4cb6a87b39cd12da494e91_1cidistance_edition, METH_VARARGS|METH_KEYWORDS, 0}; static PyObject *__pyx_pw_46_cython_magic_f072b3f10e4cb6a87b39cd12da494e91_1cidistance_edition(PyObject *__pyx_self, PyObject *__pyx_args, PyObject *__pyx_kwds) { PyObject *__pyx_v_mot1 = 0; PyObject *__pyx_v_mot2 = 0; PyObject *__pyx_r = 0; __Pyx_RefNannyDeclarations __Pyx_RefNannySetupContext("cidistance_edition (wrapper)", 0); { static PyObject **__pyx_pyargnames[] = {&__pyx_n_s_mot1,&__pyx_n_s_mot2,0}; PyObject* values[2] = {0,0}; if (unlikely(__pyx_kwds)) { Py_ssize_t kw_args; const Py_ssize_t pos_args = PyTuple_GET_SIZE(__pyx_args); switch (pos_args) { case 2: values[1] = PyTuple_GET_ITEM(__pyx_args, 1); CYTHON_FALLTHROUGH; case 1: values[0] = PyTuple_GET_ITEM(__pyx_args, 0); CYTHON_FALLTHROUGH; case 0: break; default: goto __pyx_L5_argtuple_error; } kw_args = PyDict_Size(__pyx_kwds); switch (pos_args) { case 0: if (likely((values[0] = __Pyx_PyDict_GetItemStr(__pyx_kwds, __pyx_n_s_mot1)) != 0)) kw_args--; else goto __pyx_L5_argtuple_error; CYTHON_FALLTHROUGH; case 1: if (likely((values[1] = __Pyx_PyDict_GetItemStr(__pyx_kwds, __pyx_n_s_mot2)) != 0)) kw_args--; else { __Pyx_RaiseArgtupleInvalid("cidistance_edition", 1, 2, 2, 1); __PYX_ERR(0, 3, __pyx_L3_error) } } if (unlikely(kw_args > 0)) { if (unlikely(__Pyx_ParseOptionalKeywords(__pyx_kwds, __pyx_pyargnames, 0, values, pos_args, "cidistance_edition") < 0)) __PYX_ERR(0, 3, __pyx_L3_error) } } else if (PyTuple_GET_SIZE(__pyx_args) != 2) { goto __pyx_L5_argtuple_error; } else { values[0] = PyTuple_GET_ITEM(__pyx_args, 0); values[1] = PyTuple_GET_ITEM(__pyx_args, 1); } __pyx_v_mot1 = ((PyObject*)values[0]); __pyx_v_mot2 = ((PyObject*)values[1]); } goto __pyx_L4_argument_unpacking_done; __pyx_L5_argtuple_error:; __Pyx_RaiseArgtupleInvalid("cidistance_edition", 1, 2, 2, PyTuple_GET_SIZE(__pyx_args)); __PYX_ERR(0, 3, __pyx_L3_error) __pyx_L3_error:; __Pyx_AddTraceback("_cython_magic_f072b3f10e4cb6a87b39cd12da494e91.cidistance_edition", __pyx_clineno, __pyx_lineno, __pyx_filename); __Pyx_RefNannyFinishContext(); return NULL; __pyx_L4_argument_unpacking_done:; if (unlikely(!__Pyx_ArgTypeTest(((PyObject *)__pyx_v_mot1), (&PyUnicode_Type), 1, "mot1", 1))) __PYX_ERR(0, 3, __pyx_L1_error) if (unlikely(!__Pyx_ArgTypeTest(((PyObject *)__pyx_v_mot2), (&PyUnicode_Type), 1, "mot2", 1))) __PYX_ERR(0, 3, __pyx_L1_error) __pyx_r = __pyx_pf_46_cython_magic_f072b3f10e4cb6a87b39cd12da494e91_cidistance_edition(__pyx_self, __pyx_v_mot1, __pyx_v_mot2); int __pyx_lineno = 0; const char *__pyx_filename = NULL; int __pyx_clineno = 0; /* function exit code */ goto __pyx_L0; __pyx_L1_error:; __pyx_r = NULL; __pyx_L0:; __Pyx_RefNannyFinishContext(); return __pyx_r; } static PyObject *__pyx_pf_46_cython_magic_f072b3f10e4cb6a87b39cd12da494e91_cidistance_edition(CYTHON_UNUSED PyObject *__pyx_self, PyObject *__pyx_v_mot1, PyObject *__pyx_v_mot2) { int __pyx_v_dist[0x1F4][0x1F4]; int __pyx_v_cost; int __pyx_v_c; int __pyx_v_l1; int __pyx_v_l2; PyObject *__pyx_v_i = NULL; PyObject *__pyx_v_j = NULL; PyObject *__pyx_r = NULL; __Pyx_RefNannyDeclarations __Pyx_RefNannySetupContext("cidistance_edition", 0); /* … */ /* function exit code */ __pyx_L1_error:; __Pyx_XDECREF(__pyx_t_2); __Pyx_XDECREF(__pyx_t_3); __Pyx_XDECREF(__pyx_t_7); __Pyx_XDECREF(__pyx_t_11); __Pyx_XDECREF(__pyx_t_12); __Pyx_AddTraceback("_cython_magic_f072b3f10e4cb6a87b39cd12da494e91.cidistance_edition", __pyx_clineno, __pyx_lineno, __pyx_filename); __pyx_r = NULL; __pyx_L0:; __Pyx_XDECREF(__pyx_v_i); __Pyx_XDECREF(__pyx_v_j); __Pyx_XGIVEREF(__pyx_r); __Pyx_RefNannyFinishContext(); return __pyx_r; } /* … */ __pyx_tuple_ = PyTuple_Pack(9, __pyx_n_s_mot1, __pyx_n_s_mot2, __pyx_n_s_dist, __pyx_n_s_cost, __pyx_n_s_c, __pyx_n_s_l1, __pyx_n_s_l2, __pyx_n_s_i, __pyx_n_s_j); if (unlikely(!__pyx_tuple_)) __PYX_ERR(0, 3, __pyx_L1_error) __Pyx_GOTREF(__pyx_tuple_); __Pyx_GIVEREF(__pyx_tuple_); /* … */ __pyx_t_1 = PyCFunction_NewEx(&__pyx_mdef_46_cython_magic_f072b3f10e4cb6a87b39cd12da494e91_1cidistance_edition, NULL, __pyx_n_s_cython_magic_f072b3f10e4cb6a87b); if (unlikely(!__pyx_t_1)) __PYX_ERR(0, 3, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_1); if (PyDict_SetItem(__pyx_d, __pyx_n_s_cidistance_edition, __pyx_t_1) < 0) __PYX_ERR(0, 3, __pyx_L1_error) __Pyx_DECREF(__pyx_t_1); __pyx_t_1 = 0;
04: cdef int dist [500][500]
05: cdef int cost, c
+06: cdef int l1 = len(mot1)
if (unlikely(__pyx_v_mot1 == Py_None)) { PyErr_SetString(PyExc_TypeError, "object of type 'NoneType' has no len()"); __PYX_ERR(0, 6, __pyx_L1_error) } __pyx_t_1 = __Pyx_PyUnicode_GET_LENGTH(__pyx_v_mot1); if (unlikely(__pyx_t_1 == ((Py_ssize_t)-1))) __PYX_ERR(0, 6, __pyx_L1_error) __pyx_v_l1 = __pyx_t_1;
+07: cdef int l2 = len(mot2)
if (unlikely(__pyx_v_mot2 == Py_None)) { PyErr_SetString(PyExc_TypeError, "object of type 'NoneType' has no len()"); __PYX_ERR(0, 7, __pyx_L1_error) } __pyx_t_1 = __Pyx_PyUnicode_GET_LENGTH(__pyx_v_mot2); if (unlikely(__pyx_t_1 == ((Py_ssize_t)-1))) __PYX_ERR(0, 7, __pyx_L1_error) __pyx_v_l2 = __pyx_t_1;
08:
+09: dist[0][0] = 0
((__pyx_v_dist[0])[0]) = 0;
+10: for i in range(l1):
__pyx_t_2 = __Pyx_PyInt_From_int(__pyx_v_l1); if (unlikely(!__pyx_t_2)) __PYX_ERR(0, 10, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_2); __pyx_t_3 = __Pyx_PyObject_CallOneArg(__pyx_builtin_range, __pyx_t_2); if (unlikely(!__pyx_t_3)) __PYX_ERR(0, 10, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_3); __Pyx_DECREF(__pyx_t_2); __pyx_t_2 = 0; if (likely(PyList_CheckExact(__pyx_t_3)) || PyTuple_CheckExact(__pyx_t_3)) { __pyx_t_2 = __pyx_t_3; __Pyx_INCREF(__pyx_t_2); __pyx_t_1 = 0; __pyx_t_4 = NULL; } else { __pyx_t_1 = -1; __pyx_t_2 = PyObject_GetIter(__pyx_t_3); if (unlikely(!__pyx_t_2)) __PYX_ERR(0, 10, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_2); __pyx_t_4 = Py_TYPE(__pyx_t_2)->tp_iternext; if (unlikely(!__pyx_t_4)) __PYX_ERR(0, 10, __pyx_L1_error) } __Pyx_DECREF(__pyx_t_3); __pyx_t_3 = 0; for (;;) { if (likely(!__pyx_t_4)) { if (likely(PyList_CheckExact(__pyx_t_2))) { if (__pyx_t_1 >= PyList_GET_SIZE(__pyx_t_2)) break; #if CYTHON_ASSUME_SAFE_MACROS && !CYTHON_AVOID_BORROWED_REFS __pyx_t_3 = PyList_GET_ITEM(__pyx_t_2, __pyx_t_1); __Pyx_INCREF(__pyx_t_3); __pyx_t_1++; if (unlikely(0 < 0)) __PYX_ERR(0, 10, __pyx_L1_error) #else __pyx_t_3 = PySequence_ITEM(__pyx_t_2, __pyx_t_1); __pyx_t_1++; if (unlikely(!__pyx_t_3)) __PYX_ERR(0, 10, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_3); #endif } else { if (__pyx_t_1 >= PyTuple_GET_SIZE(__pyx_t_2)) break; #if CYTHON_ASSUME_SAFE_MACROS && !CYTHON_AVOID_BORROWED_REFS __pyx_t_3 = PyTuple_GET_ITEM(__pyx_t_2, __pyx_t_1); __Pyx_INCREF(__pyx_t_3); __pyx_t_1++; if (unlikely(0 < 0)) __PYX_ERR(0, 10, __pyx_L1_error) #else __pyx_t_3 = PySequence_ITEM(__pyx_t_2, __pyx_t_1); __pyx_t_1++; if (unlikely(!__pyx_t_3)) __PYX_ERR(0, 10, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_3); #endif } } else { __pyx_t_3 = __pyx_t_4(__pyx_t_2); if (unlikely(!__pyx_t_3)) { PyObject* exc_type = PyErr_Occurred(); if (exc_type) { if (likely(__Pyx_PyErr_GivenExceptionMatches(exc_type, PyExc_StopIteration))) PyErr_Clear(); else __PYX_ERR(0, 10, __pyx_L1_error) } break; } __Pyx_GOTREF(__pyx_t_3); } __Pyx_XDECREF_SET(__pyx_v_i, __pyx_t_3); __pyx_t_3 = 0; /* … */ } __Pyx_DECREF(__pyx_t_2); __pyx_t_2 = 0;
+11: dist[i+1][0] = dist[i][0] + 1
__pyx_t_5 = __Pyx_PyIndex_AsSsize_t(__pyx_v_i); if (unlikely((__pyx_t_5 == (Py_ssize_t)-1) && PyErr_Occurred())) __PYX_ERR(0, 11, __pyx_L1_error) __pyx_t_3 = __Pyx_PyInt_AddObjC(__pyx_v_i, __pyx_int_1, 1, 0, 0); if (unlikely(!__pyx_t_3)) __PYX_ERR(0, 11, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_3); __pyx_t_6 = __Pyx_PyIndex_AsSsize_t(__pyx_t_3); if (unlikely((__pyx_t_6 == (Py_ssize_t)-1) && PyErr_Occurred())) __PYX_ERR(0, 11, __pyx_L1_error) __Pyx_DECREF(__pyx_t_3); __pyx_t_3 = 0; ((__pyx_v_dist[__pyx_t_6])[0]) = (((__pyx_v_dist[__pyx_t_5])[0]) + 1);
+12: dist[0][i+1] = dist[0][i] + 1
__pyx_t_5 = __Pyx_PyIndex_AsSsize_t(__pyx_v_i); if (unlikely((__pyx_t_5 == (Py_ssize_t)-1) && PyErr_Occurred())) __PYX_ERR(0, 12, __pyx_L1_error) __pyx_t_3 = __Pyx_PyInt_AddObjC(__pyx_v_i, __pyx_int_1, 1, 0, 0); if (unlikely(!__pyx_t_3)) __PYX_ERR(0, 12, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_3); __pyx_t_6 = __Pyx_PyIndex_AsSsize_t(__pyx_t_3); if (unlikely((__pyx_t_6 == (Py_ssize_t)-1) && PyErr_Occurred())) __PYX_ERR(0, 12, __pyx_L1_error) __Pyx_DECREF(__pyx_t_3); __pyx_t_3 = 0; ((__pyx_v_dist[0])[__pyx_t_6]) = (((__pyx_v_dist[0])[__pyx_t_5]) + 1);
+13: for j in range(l2):
__pyx_t_3 = __Pyx_PyInt_From_int(__pyx_v_l2); if (unlikely(!__pyx_t_3)) __PYX_ERR(0, 13, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_3); __pyx_t_7 = __Pyx_PyObject_CallOneArg(__pyx_builtin_range, __pyx_t_3); if (unlikely(!__pyx_t_7)) __PYX_ERR(0, 13, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_7); __Pyx_DECREF(__pyx_t_3); __pyx_t_3 = 0; if (likely(PyList_CheckExact(__pyx_t_7)) || PyTuple_CheckExact(__pyx_t_7)) { __pyx_t_3 = __pyx_t_7; __Pyx_INCREF(__pyx_t_3); __pyx_t_5 = 0; __pyx_t_8 = NULL; } else { __pyx_t_5 = -1; __pyx_t_3 = PyObject_GetIter(__pyx_t_7); if (unlikely(!__pyx_t_3)) __PYX_ERR(0, 13, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_3); __pyx_t_8 = Py_TYPE(__pyx_t_3)->tp_iternext; if (unlikely(!__pyx_t_8)) __PYX_ERR(0, 13, __pyx_L1_error) } __Pyx_DECREF(__pyx_t_7); __pyx_t_7 = 0; for (;;) { if (likely(!__pyx_t_8)) { if (likely(PyList_CheckExact(__pyx_t_3))) { if (__pyx_t_5 >= PyList_GET_SIZE(__pyx_t_3)) break; #if CYTHON_ASSUME_SAFE_MACROS && !CYTHON_AVOID_BORROWED_REFS __pyx_t_7 = PyList_GET_ITEM(__pyx_t_3, __pyx_t_5); __Pyx_INCREF(__pyx_t_7); __pyx_t_5++; if (unlikely(0 < 0)) __PYX_ERR(0, 13, __pyx_L1_error) #else __pyx_t_7 = PySequence_ITEM(__pyx_t_3, __pyx_t_5); __pyx_t_5++; if (unlikely(!__pyx_t_7)) __PYX_ERR(0, 13, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_7); #endif } else { if (__pyx_t_5 >= PyTuple_GET_SIZE(__pyx_t_3)) break; #if CYTHON_ASSUME_SAFE_MACROS && !CYTHON_AVOID_BORROWED_REFS __pyx_t_7 = PyTuple_GET_ITEM(__pyx_t_3, __pyx_t_5); __Pyx_INCREF(__pyx_t_7); __pyx_t_5++; if (unlikely(0 < 0)) __PYX_ERR(0, 13, __pyx_L1_error) #else __pyx_t_7 = PySequence_ITEM(__pyx_t_3, __pyx_t_5); __pyx_t_5++; if (unlikely(!__pyx_t_7)) __PYX_ERR(0, 13, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_7); #endif } } else { __pyx_t_7 = __pyx_t_8(__pyx_t_3); if (unlikely(!__pyx_t_7)) { PyObject* exc_type = PyErr_Occurred(); if (exc_type) { if (likely(__Pyx_PyErr_GivenExceptionMatches(exc_type, PyExc_StopIteration))) PyErr_Clear(); else __PYX_ERR(0, 13, __pyx_L1_error) } break; } __Pyx_GOTREF(__pyx_t_7); } __Pyx_XDECREF_SET(__pyx_v_j, __pyx_t_7); __pyx_t_7 = 0; /* … */ } __Pyx_DECREF(__pyx_t_3); __pyx_t_3 = 0;
+14: cost = dist[i][j+1] + 1
__pyx_t_6 = __Pyx_PyIndex_AsSsize_t(__pyx_v_i); if (unlikely((__pyx_t_6 == (Py_ssize_t)-1) && PyErr_Occurred())) __PYX_ERR(0, 14, __pyx_L1_error) __pyx_t_7 = __Pyx_PyInt_AddObjC(__pyx_v_j, __pyx_int_1, 1, 0, 0); if (unlikely(!__pyx_t_7)) __PYX_ERR(0, 14, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_7); __pyx_t_9 = __Pyx_PyIndex_AsSsize_t(__pyx_t_7); if (unlikely((__pyx_t_9 == (Py_ssize_t)-1) && PyErr_Occurred())) __PYX_ERR(0, 14, __pyx_L1_error) __Pyx_DECREF(__pyx_t_7); __pyx_t_7 = 0; __pyx_v_cost = (((__pyx_v_dist[__pyx_t_6])[__pyx_t_9]) + 1);
+15: c = dist[i+1][j] + 1
__pyx_t_7 = __Pyx_PyInt_AddObjC(__pyx_v_i, __pyx_int_1, 1, 0, 0); if (unlikely(!__pyx_t_7)) __PYX_ERR(0, 15, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_7); __pyx_t_9 = __Pyx_PyIndex_AsSsize_t(__pyx_t_7); if (unlikely((__pyx_t_9 == (Py_ssize_t)-1) && PyErr_Occurred())) __PYX_ERR(0, 15, __pyx_L1_error) __Pyx_DECREF(__pyx_t_7); __pyx_t_7 = 0; __pyx_t_6 = __Pyx_PyIndex_AsSsize_t(__pyx_v_j); if (unlikely((__pyx_t_6 == (Py_ssize_t)-1) && PyErr_Occurred())) __PYX_ERR(0, 15, __pyx_L1_error) __pyx_v_c = (((__pyx_v_dist[__pyx_t_9])[__pyx_t_6]) + 1);
+16: if c < cost : cost = c
__pyx_t_10 = ((__pyx_v_c < __pyx_v_cost) != 0); if (__pyx_t_10) { __pyx_v_cost = __pyx_v_c; }
+17: c = dist[i][j]
__pyx_t_6 = __Pyx_PyIndex_AsSsize_t(__pyx_v_i); if (unlikely((__pyx_t_6 == (Py_ssize_t)-1) && PyErr_Occurred())) __PYX_ERR(0, 17, __pyx_L1_error) __pyx_t_9 = __Pyx_PyIndex_AsSsize_t(__pyx_v_j); if (unlikely((__pyx_t_9 == (Py_ssize_t)-1) && PyErr_Occurred())) __PYX_ERR(0, 17, __pyx_L1_error) __pyx_v_c = ((__pyx_v_dist[__pyx_t_6])[__pyx_t_9]);
+18: if mot1[i] != mot2[j] : c += 1
__pyx_t_7 = __Pyx_PyObject_GetItem(__pyx_v_mot1, __pyx_v_i); if (unlikely(!__pyx_t_7)) __PYX_ERR(0, 18, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_7); __pyx_t_11 = __Pyx_PyObject_GetItem(__pyx_v_mot2, __pyx_v_j); if (unlikely(!__pyx_t_11)) __PYX_ERR(0, 18, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_11); __pyx_t_12 = PyObject_RichCompare(__pyx_t_7, __pyx_t_11, Py_NE); __Pyx_XGOTREF(__pyx_t_12); if (unlikely(!__pyx_t_12)) __PYX_ERR(0, 18, __pyx_L1_error) __Pyx_DECREF(__pyx_t_7); __pyx_t_7 = 0; __Pyx_DECREF(__pyx_t_11); __pyx_t_11 = 0; __pyx_t_10 = __Pyx_PyObject_IsTrue(__pyx_t_12); if (unlikely(__pyx_t_10 < 0)) __PYX_ERR(0, 18, __pyx_L1_error) __Pyx_DECREF(__pyx_t_12); __pyx_t_12 = 0; if (__pyx_t_10) { __pyx_v_c = (__pyx_v_c + 1); }
+19: if c < cost : cost = c
__pyx_t_10 = ((__pyx_v_c < __pyx_v_cost) != 0); if (__pyx_t_10) { __pyx_v_cost = __pyx_v_c; }
+20: dist[i+1][j+1] = cost
__pyx_t_12 = __Pyx_PyInt_AddObjC(__pyx_v_i, __pyx_int_1, 1, 0, 0); if (unlikely(!__pyx_t_12)) __PYX_ERR(0, 20, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_12); __pyx_t_9 = __Pyx_PyIndex_AsSsize_t(__pyx_t_12); if (unlikely((__pyx_t_9 == (Py_ssize_t)-1) && PyErr_Occurred())) __PYX_ERR(0, 20, __pyx_L1_error) __Pyx_DECREF(__pyx_t_12); __pyx_t_12 = 0; __pyx_t_12 = __Pyx_PyInt_AddObjC(__pyx_v_j, __pyx_int_1, 1, 0, 0); if (unlikely(!__pyx_t_12)) __PYX_ERR(0, 20, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_12); __pyx_t_6 = __Pyx_PyIndex_AsSsize_t(__pyx_t_12); if (unlikely((__pyx_t_6 == (Py_ssize_t)-1) && PyErr_Occurred())) __PYX_ERR(0, 20, __pyx_L1_error) __Pyx_DECREF(__pyx_t_12); __pyx_t_12 = 0; ((__pyx_v_dist[__pyx_t_9])[__pyx_t_6]) = __pyx_v_cost;
+21: cost = dist[l1][l2]
__pyx_v_cost = ((__pyx_v_dist[__pyx_v_l1])[__pyx_v_l2]);
+22: return cost
__Pyx_XDECREF(__pyx_r); __pyx_t_2 = __Pyx_PyInt_From_int(__pyx_v_cost); if (unlikely(!__pyx_t_2)) __PYX_ERR(0, 22, __pyx_L1_error) __Pyx_GOTREF(__pyx_t_2); __pyx_r = __pyx_t_2; __pyx_t_2 = 0; goto __pyx_L0;
mot1, mot2 = "idstzance","distances"
%timeit cidistance_edition(mot1, mot2)
16.9 µs ± 3.47 µs per loop (mean ± std. dev. of 7 runs, 10000 loops each)
import sys
from pyquickhelper.loghelper import run_cmd
code = """
def cdistance_edition(str mot1, str mot2):
cdef int dist [500][500]
cdef int cost, c
cdef int l1 = len(mot1)
cdef int l2 = len(mot2)
dist[0][0] = 0
for i in range(l1):
dist[i+1][0] = dist[i][0] + 1
dist[0][i+1] = dist[0][i] + 1
for j in range(l2):
cost = dist[i][j+1] + 1
c = dist[i+1][j] + 1
if c < cost : cost = c
c = dist[i][j]
if mot1[i] != mot2[j] : c += 1
if c < cost : cost = c
dist[i+1][j+1] = cost
cost = dist[l1][l2]
return cost
"""
name = "cedit_distance"
with open(name + ".pyx","w") as f : f.write(code)
setup_code = """
from distutils.core import setup
from Cython.Build import cythonize
setup(
ext_modules = cythonize("__NAME__.pyx",
compiler_directives={'language_level' : "3"})
)
""".replace("__NAME__",name)
with open("setup.py","w") as f:
f.write(setup_code)
cmd = "{0} setup.py build_ext --inplace".format(sys.executable)
out,err = run_cmd(cmd)
if err is not None and err != '':
raise Exception(err)
import pyximport
pyximport.install()
import cedit_distance
from cedit_distance import cdistance_edition
mot1, mot2 = "idstzance","distances"
%timeit cdistance_edition(mot1, mot2)
11.4 µs ± 1.93 µs per loop (mean ± std. dev. of 7 runs, 100000 loops each)
La version Cython est 10 fois plus rapide. Et cela ne semble pas dépendre de la dimension du problème.
mot1 = mot1 * 10
mot2 = mot2 * 10
%timeit distance_edition(mot1,mot2)
%timeit cdistance_edition(mot1, mot2)
11.5 ms ± 561 µs per loop (mean ± std. dev. of 7 runs, 100 loops each) 724 µs ± 30 µs per loop (mean ± std. dev. of 7 runs, 1000 loops each)