2
0

chunker.pyx 1.9 KB

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253545556575859606162
  1. # -*- coding: utf-8 -*-
  2. API_VERSION = 2
  3. from libc.stdlib cimport free
  4. cdef extern from "_chunker.c":
  5. ctypedef int uint32_t
  6. ctypedef struct _Chunker "Chunker":
  7. pass
  8. _Chunker *chunker_init(int window_size, int chunk_mask, int min_size, int max_size, uint32_t seed)
  9. void chunker_set_fd(_Chunker *chunker, object f, int fd)
  10. void chunker_free(_Chunker *chunker)
  11. object chunker_process(_Chunker *chunker)
  12. uint32_t *buzhash_init_table(uint32_t seed)
  13. uint32_t c_buzhash "buzhash"(unsigned char *data, size_t len, uint32_t *h)
  14. uint32_t c_buzhash_update "buzhash_update"(uint32_t sum, unsigned char remove, unsigned char add, size_t len, uint32_t *h)
  15. cdef class Chunker:
  16. cdef _Chunker *chunker
  17. def __cinit__(self, window_size, chunk_mask, min_size, max_size, seed):
  18. self.chunker = chunker_init(window_size, chunk_mask, min_size, max_size, seed & 0xffffffff)
  19. def chunkify(self, fd, fh=-1):
  20. """
  21. Cut a file into chunks.
  22. :param fd: Python file object
  23. :param fh: OS-level file handle (if available),
  24. defaults to -1 which means not to use OS-level fd.
  25. """
  26. chunker_set_fd(self.chunker, fd, fh)
  27. return self
  28. def __dealloc__(self):
  29. if self.chunker:
  30. chunker_free(self.chunker)
  31. def __iter__(self):
  32. return self
  33. def __next__(self):
  34. return chunker_process(self.chunker)
  35. def buzhash(unsigned char *data, unsigned long seed):
  36. cdef uint32_t *table
  37. cdef uint32_t sum
  38. table = buzhash_init_table(seed & 0xffffffff)
  39. sum = c_buzhash(data, len(data), table)
  40. free(table)
  41. return sum
  42. def buzhash_update(uint32_t sum, unsigned char remove, unsigned char add, size_t len, unsigned long seed):
  43. cdef uint32_t *table
  44. table = buzhash_init_table(seed & 0xffffffff)
  45. sum = c_buzhash_update(sum, remove, add, len, table)
  46. free(table)
  47. return sum