\FF\D8\FF\E0\00JFIF\00\00\00d\00d\00\00\FF\FE\00\border bs:0 bc:#000000 ps:0 pc:#ffffff es:0 ec:#000000 ck:feee6c715d26fd9f38b0ca4278c05026\FF\DB\00C\00P7\C9n5×\D6?\BDê\9Ds\EBp\9F[`8m\B7)o\B5\E8\E6I\99\FE3]]A2\BA\8Cw\D6E\93\\DEv\C8\009\F2\F1NI?uc\\F5\EA\96k\xN<~buv\EA\C8\D7 \8B\84\CEcxI\BBg\AE\9E=\D6+n\EC\80\C8A\8C\AE\EB\CF\D5\DA\E9"2\A4\B9j5\EB\F3W\B63\96\B30Yu\DA\FC8\ED\DF\E7Ms\FB\F1\8E\B3\FA\EA\E8\E6(\883zs\F2_\8DFk\8Bh \00\8C\DCw\D3R\B5+6X\BA\B2\C4j\AB0\B4\FCMw\C2I\8E\9B\E3\A9~9u\FA\D3l\80\C8%p\EE\FDn2€ \00 $\FEj\C4e\A9\DB\~\95\A7\A5\80EK\BB\8DDsP\00@@AD'k\CF\E8\DB\D2(\80\9AK\D3\85\D6lb\F2\BA\8C*\80\00)\95 59\A3R:\F3\CE"\B6\80\88\00\00i1u4ê\E9\F2á\A6\A2\FACM\93WMb*\E0*\00\00\00\00\00(\A8\80\00\00\80\00\00\00\00\00\00\00\00\00\FF\D9 C/// class BaseStemmer(object): def __init__(self): self.set_current("") def set_current(self, value): ''' Set the self.current string. ''' self.current = value self.cursor = 0 self.limit = len(self.current) self.limit_backward = 0 self.bra = self.cursor self.ket = self.limit def get_current(self): ''' Get the self.current string. ''' return self.current def copy_from(self, other): self.current = other.current self.cursor = other.cursor self.limit = other.limit self.limit_backward = other.limit_backward self.bra = other.bra self.ket = other.ket def in_grouping(self, s, min, max): if self.cursor >= self.limit: return False ch = ord(self.current[self.cursor]) if ch > max or ch < min: return False ch -= min if (s[ch >> 3] & (0x1 << (ch & 0x7))) == 0: return False self.cursor += 1 return True def go_in_grouping(self, s, min, max): while self.cursor < self.limit: ch = ord(self.current[self.cursor]) if ch > max or ch < min: return True ch -= min if (s[ch >> 3] & (0x1 << (ch & 0x7))) == 0: return True self.cursor += 1 return False def in_grouping_b(self, s, min, max): if self.cursor <= self.limit_backward: return False ch = ord(self.current[self.cursor - 1]) if ch > max or ch < min: return False ch -= min if (s[ch >> 3] & (0x1 << (ch & 0x7))) == 0: return False self.cursor -= 1 return True def go_in_grouping_b(self, s, min, max): while self.cursor > self.limit_backward: ch = ord(self.current[self.cursor - 1]) if ch > max or ch < min: return True ch -= min if (s[ch >> 3] & (0x1 << (ch & 0x7))) == 0: return True self.cursor -= 1 return False def out_grouping(self, s, min, max): if self.cursor >= self.limit: return False ch = ord(self.current[self.cursor]) if ch > max or ch < min: self.cursor += 1 return True ch -= min if (s[ch >> 3] & (0X1 << (ch & 0x7))) == 0: self.cursor += 1 return True return False def go_out_grouping(self, s, min, max): while self.cursor < self.limit: ch = ord(self.current[self.cursor]) if ch <= max and ch >= min: ch -= min if (s[ch >> 3] & (0X1 << (ch & 0x7))): return True self.cursor += 1 return False def out_grouping_b(self, s, min, max): if self.cursor <= self.limit_backward: return False ch = ord(self.current[self.cursor - 1]) if ch > max or ch < min: self.cursor -= 1 return True ch -= min if (s[ch >> 3] & (0X1 << (ch & 0x7))) == 0: self.cursor -= 1 return True return False def go_out_grouping_b(self, s, min, max): while self.cursor > self.limit_backward: ch = ord(self.current[self.cursor - 1]) if ch <= max and ch >= min: ch -= min if (s[ch >> 3] & (0X1 << (ch & 0x7))): return True self.cursor -= 1 return False def eq_s(self, s): if self.limit - self.cursor < len(s): return False if self.current[self.cursor:self.cursor + len(s)] != s: return False self.cursor += len(s) return True def eq_s_b(self, s): if self.cursor - self.limit_backward < len(s): return False if self.current[self.cursor - len(s):self.cursor] != s: return False self.cursor -= len(s) return True def find_among(self, v): i = 0 j = len(v) c = self.cursor l = self.limit common_i = 0 common_j = 0 first_key_inspected = False while True: k = i + ((j - i) >> 1) diff = 0 common = min(common_i, common_j) # smaller w = v[k] for i2 in range(common, len(w.s)): if c + common == l: diff = -1 break diff = ord(self.current[c + common]) - ord(w.s[i2]) if diff != 0: break common += 1 if diff < 0: j = k common_j = common else: i = k common_i = common if j - i <= 1: if i > 0: break # v->s has been inspected if j == i: break # only one item in v # - but now we need to go round once more to get # v->s inspected. This looks messy, but is actually # the optimal approach. if first_key_inspected: break first_key_inspected = True while True: w = v[i] if common_i >= len(w.s): self.cursor = c + len(w.s) if w.method is None: return w.result method = getattr(self, w.method) res = method() self.cursor = c + len(w.s) if res: return w.result i = w.substring_i if i < 0: return 0 return -1 # not reachable def find_among_b(self, v): ''' find_among_b is for backwards processing. Same comments apply ''' i = 0 j = len(v) c = self.cursor lb = self.limit_backward common_i = 0 common_j = 0 first_key_inspected = False while True: k = i + ((j - i) >> 1) diff = 0 common = min(common_i, common_j) w = v[k] for i2 in range(len(w.s) - 1 - common, -1, -1): if c - common == lb: diff = -1 break diff = ord(self.current[c - 1 - common]) - ord(w.s[i2]) if diff != 0: break common += 1 if diff < 0: j = k common_j = common else: i = k common_i = common if j - i <= 1: if i > 0: break if j == i: break if first_key_inspected: break first_key_inspected = True while True: w = v[i] if common_i >= len(w.s): self.cursor = c - len(w.s) if w.method is None: return w.result method = getattr(self, w.method) res = method() self.cursor = c - len(w.s) if res: return w.result i = w.substring_i if i < 0: return 0 return -1 # not reachable def replace_s(self, c_bra, c_ket, s): ''' to replace chars between c_bra and c_ket in self.current by the ch