Commit dafea281f153005547b4bb728b1231c786cf6ed3 - tlsh

Add Tlsh_fromTlshStr() to python API and extend py_ext/test.py to test for this. Add description of expanded python API to README.md file. Scott4man 8 years ago

4 changed file(s) with 64 addition(s) and 10 deletion(s). Raw diff Collapse all Expand all

-0

README.md less more

89	89	import tlsh
90	90	tlsh.hash(data)
91	91	```
	92
92	93
93	94	Note that the data must contain at least 256 bytes to generate a hash value and that
94	95	it must have a certain amount of randomness.

105	106	with only a single instance of the pattern, then the difference will be increased
106	107	if the file lenght is included. But by using the `diffxlen` function, the file
107	108	length will be removed from consideration.
	109
	110	Note that the python API has been extended to miror the C++ API. See
	111	py_ext/tlshmodule.cpp and the py_ext/test.py script to see the full API set.
108	112
109	113	# Design Choices
110	114

-0

Testing/exp/python_test_EXP less more

7	7	h2.diff(h1) 427
8	8	h1.diff(hex2) 427
9	9	h2.diff(hex1) 427
	10	tlsh.Tlsh.fromTlshStr E951784702042376169012B1BA5A76EAF36092FC3311A595B4856235278F9F973763EF
	11	h3.diff(h2) 0

-0

py_ext/test.py less more

33	33	print('h2.diff(h1)', h2.diff(h1))
34	34	print('h1.diff(hex2)', h1.diff(hex2))
35	35	print('h2.diff(hex1)', h2.diff(hex1))
	36
	37	h3 = tlsh.Tlsh()
	38	h3.fromTlshStr(hex2)
	39	print('tlsh.Tlsh.fromTlshStr', hex2)
	40	print('h3.diff(h2)', h3.diff(h2))
	41
	42

+51

-10

py_ext/tlshmodule.cpp less more

11	11	#else
12	12	# define BYTES_VALUE_CHAR "s"
13	13	#endif
14
15		#define MIN_TLSH_LEN 512
16	14
17	15	static char tlsh_doc[] =
18	16	"TLSH C version - similarity matching and searching";

50	48	}
51	49
52	50	Tlsh tlsh1, tlsh2;
53		tlsh1.fromTlshStr(hash1);
54		tlsh2.fromTlshStr(hash2);
	51	if (tlsh1.fromTlshStr(hash1) != 0) {
	52	return PyErr_Format(PyExc_ValueError, "argument %s is not a TLSH hex string", hash1);
	53	}
	54	if (tlsh2.fromTlshStr(hash2) != 0) {
	55	return PyErr_Format(PyExc_ValueError, "argument %s is not a TLSH hex string", hash2);
	56	}
55	57
56	58	int score = tlsh1.totalDiff(&tlsh2);
57	59

66	68	}
67	69
68	70	Tlsh tlsh1, tlsh2;
69		tlsh1.fromTlshStr(hash1);
70		tlsh2.fromTlshStr(hash2);
	71	if (tlsh1.fromTlshStr(hash1) != 0) {
	72	return PyErr_Format(PyExc_ValueError, "argument %s is not a TLSH hex string", hash1);
	73	}
	74	if (tlsh2.fromTlshStr(hash2) != 0) {
	75	return PyErr_Format(PyExc_ValueError, "argument %s is not a TLSH hex string", hash2);
	76	}
71	77
72	78	int score = tlsh1.totalDiff(&tlsh2, false);
73	79

90	96	Tlsh tlsh;
91	97	} tlsh_TlshObject;
92	98
	99	static PyObject * Tlsh_fromTlshStr(tlsh_TlshObject , PyObject );
93	100	static PyObject * Tlsh_update(tlsh_TlshObject , PyObject );
94	101	static PyObject * Tlsh_final(tlsh_TlshObject *);
95	102	static PyObject * Tlsh_hexdigest(tlsh_TlshObject *);
96	103	static PyObject * Tlsh_diff(tlsh_TlshObject , PyObject );
97	104
98	105	static PyMethodDef Tlsh_methods[] = {
	106	{"fromTlshStr", (PyCFunction) Tlsh_fromTlshStr, METH_VARARGS,
	107	"Create a TLSH instance from a hex string."
	108	},
99	109	{"update", (PyCFunction) Tlsh_update, METH_VARARGS,
100	110	"Update the TLSH with the given string."
101	111	},

156	166	};
157	167
158	168	static PyObject *
	169	Tlsh_fromTlshStr(tlsh_TlshObject self, PyObject args)
	170	{
	171	char *str;
	172	Py_ssize_t len;
	173
	174	PyObject *arg;
	175
	176	if (PyTuple_Size(args) != 1)
	177	return PyErr_Format(PyExc_TypeError, "function takes exactly 1 argument (%lu given)", PyTuple_Size(args));
	178
	179	arg = PyTuple_GetItem(args, 0);
	180	if (PyBytes_AsStringAndSize(arg, &str, &len) == -1) {
	181	PyErr_SetString(PyExc_ValueError, "argument is not a TLSH hex string");
	182	return NULL;
	183	}
	184
	185	if (len != TLSH_STRING_LEN) {
	186	PyErr_SetString(PyExc_ValueError, "argument length incorrect: not a TLSH hex string");
	187	return NULL;
	188	}
	189
	190	if (self->tlsh.fromTlshStr(str) != 0) {
	191	PyErr_SetString(PyExc_ValueError, "argument value incorrect: not a TLSH hex string");
	192	return NULL;
	193	}
	194	self->finalized = true;
	195
	196	Py_RETURN_NONE;
	197	}
	198
	199	static PyObject *
159	200	Tlsh_update(tlsh_TlshObject self, PyObject args)
160	201	{
161	202	const char *str;

168	209	PyErr_SetString(PyExc_ValueError, "final() has already been called");
169	210	return NULL;
170	211	}
171		if (self->required_data < MIN_TLSH_LEN) {
172		self->required_data += len > MIN_TLSH_LEN ? MIN_TLSH_LEN : len;
	212	if (self->required_data < MIN_DATA_LENGTH) {
	213	self->required_data += len > MIN_DATA_LENGTH ? MIN_DATA_LENGTH : len;
173	214	}
174	215
175	216	self->tlsh.update((unsigned char *) str, (unsigned int) len);

184	225	PyErr_SetString(PyExc_ValueError, "final() has already been called");
185	226	return NULL;
186	227	}
187		if (self->required_data < MIN_TLSH_LEN) {
188		return PyErr_Format(PyExc_ValueError, "less than %u of input", MIN_TLSH_LEN);
	228	if (self->required_data < MIN_DATA_LENGTH) {
	229	return PyErr_Format(PyExc_ValueError, "less than %u of input", MIN_DATA_LENGTH);
189	230	}
190	231	self->finalized = true;
191	232	self->tlsh.final();