Enable cythonized range contains

setup.py

@@ -22,7 +22,7 @@
 
 setup(
     name='pybloof',
-    version='0.7.1',
+    version='0.7.3',
     author='Jake Heinz',
     author_email='[email protected]',
     url="http://github.com/jhgg/pybloof",

src/_pybloof.pyx

@@ -8,6 +8,8 @@ import base64
 
 from libc.string cimport memcpy
 
+cdef unsigned int high = 0xFFFFFFFF
+
 cdef array.array char_array_template = array.array('b', [])
 
 cdef extern from "MurmurHash3.h" nogil:
@@ -47,16 +49,20 @@ cdef void _get_hash_buckets(key, unsigned long long * _bucket_indexes, unsigned
     for i in range(hash_count):
         _bucket_indexes[i] = llabs((hash1 + i * hash2) % max)\
 
+
+
 @cython.boundscheck(False)
-cdef void _get_hash_buckets_for_long(long long key, unsigned long long * _bucket_indexes, unsigned int hash_count,
+@cython.cdivision(True)
+cdef inline void _get_hash_buckets_for_long(long long key, unsigned long long * _bucket_indexes, unsigned int hash_count,
                                      unsigned long max):
     cdef unsigned long result[2]
     cdef unsigned long hash1, hash2
     cdef unsigned long i
 
+
     MurmurHash3_x64_128_long(key, 0, &result)
     hash1 = result[0]
-    MurmurHash3_x64_128_long(key, result[1] & 0xFFFFFFFF, result)
+    MurmurHash3_x64_128_long(key, result[1] & high, result)
     hash2 = result[0]
 
     for i in range(hash_count):
@@ -67,6 +73,34 @@ cdef char* fmt = '!III'
 cdef ssize_t header_size = sizeof(unsigned int) * 3
 DEF MAX_HASHES = 32
 
+
+@cython.boundscheck(False)
+cdef int _uniques_in_range(unsigned int start, unsigned int stop, int[:] bitarray,
+                           unsigned long long * _bucket_indexes, unsigned int size,
+                           unsigned int hash_count, int[:] flags, int[:] uniques):
+    cdef unsigned int i
+    cdef unsigned int bucket_index
+    cdef unsigned int idx
+    cdef unsigned int bit
+    cdef unsigned int off
+    cdef int is_in
+    idx = 0
+    off = 0
+    for item in range(start, stop):
+        is_in = 1
+        _get_hash_buckets_for_long(item, _bucket_indexes, hash_count, size)
+        for i in range(hash_count):
+            if not bitarray[_bucket_indexes[i]]:
+                is_in = 0
+                break
+        flags[idx] = is_in
+        if is_in:
+            uniques[off] = item
+            off += 1
+        idx += 1
+    return off
+
+
 cdef class _BloomFilter:
     cdef unsigned int _size
     cdef unsigned int _hashes
@@ -213,9 +247,35 @@ cdef class UIntBloomFilter(_BloomFilter):
 
         return True
 
+    @cython.boundscheck(False)
+    cdef _uniques_in_range(self, unsigned int start, unsigned int stop):
+        cdef unsigned long long _bucket_indexes[MAX_HASHES]
+        cdef unsigned int i
+        cdef unsigned int bucket_index
+        cdef unsigned int idx
+        cdef unsigned int bit
+        cdef int is_in
+        cdef array.array flags = array.array('i', [stop - start])
+        array.resize(flags, stop - start)
+        cdef array.array uniq = array.array('i', [stop - start])
+        array.resize(uniq, stop - start)
+        cdef array.array bitarray = array.array('i', [self._size])
+        array.resize(bitarray, self._size)
+        byte = self._bitarray.unpack()
+        for idx in range(self._size):
+            bitarray[idx] = byte[idx] == 255
+        off = _uniques_in_range(start, stop, bitarray, _bucket_indexes,
+                                self._size, self._hashes, flags, uniq)
+        return flags, uniq, off
+
     def __contains__(self, unsigned int item):
         return self.contains(item)
 
+    def uniques_in_range(self, start, stop):
+        flags, uniq, off = self._uniques_in_range(start, stop)
+        return list(flags), set(uniq[:off])
+
+
 cdef class StringBloomFilter(_BloomFilter):
     cpdef add(self, item):
         cdef unsigned long long _bucket_indexes[MAX_HASHES]