Merge branch 'solve-833' into main-plus-fixes

outlines/fsm/parsing.py

@@ -38,6 +38,7 @@
 from outlines.fsm.regex import (
     fsm_union,
     get_sub_fsms_from_seq,
+    get_token_transitions,
     make_deterministic_fsm,
     walk_fsm,
 )
@@ -569,9 +570,15 @@ def match(self, text, pos, last_fsm_state_seq: Optional[Tuple[int, ...]] = None)
 
         text_part = text[start_pos:]
 
+        text_transitions = get_token_transitions(
+            self.fsm.fsm_info.alphabet_symbol_mapping,
+            self.fsm.fsm_info.alphabet_anything_value,
+            text_part,
+        )
+
         state_seq = walk_fsm(
             self.fsm,
-            text_part,
+            text_transitions,
             start_state,
             full_match=self.match_whole,
         )

outlines/fsm/regex.py

@@ -87,14 +87,11 @@ def fsm_info(self):
                 ((k, z) for k, v in self.trans_key_to_states.items() for z in v),
                 dtype=np.dtype("int64, int64"),
             )
-            alphabet_symbol_mapping_items = np.fromiter(
-                (
-                    it
-                    for it in self.alphabet._symbol_mapping.items()
-                    if it[0] != anything_else
-                ),
-                dtype=np.dtype("U2, int64"),
-            )
+            alphabet_symbol_mapping_items = [
+                (k, v)
+                for k, v in self.alphabet._symbol_mapping.items()
+                if k != anything_else
+            ]
             nb_finals = np.fromiter(self.finals, dtype=np.dtype("int64"))
             self.__dict__["_fsm_info"] = create_fsm_info(
                 self.initial,
@@ -110,7 +107,7 @@ def fsm_info(self):
 
 nb_int_list_type = numba.types.ListType(numba.int64)
 nb_int_pair_type = numba.types.UniTuple(numba.int64, 2)
-nb_unichar_2_type = numba.types.UnicodeCharSeq(2)
+nb_unicode_type = numba.types.unicode_type
 
 
 @numba.njit(cache=True)
@@ -136,7 +133,7 @@ def create_fsm_info(
 
     # use 2-char strings so that we can represent incomplete utf-8 sequences
     # as 2-hex-digit pairs
-    alphabet_symbol_map = numba.typed.Dict.empty(nb_unichar_2_type, numba.int64)
+    alphabet_symbol_map = numba.typed.Dict.empty(nb_unicode_type, numba.int64)
     for symbol_and_trans_key in alphabet_symbol_mapping_items:
         alphabet_symbol_map[symbol_and_trans_key[0]] = symbol_and_trans_key[1]
 
@@ -199,7 +196,7 @@ def transition_trie_setdefault(
 
 
 def byte_symbol(byte: int) -> str:
-    return f"{byte:02X}" if byte >= 0x80 else chr(byte)
+    return f"\x00{byte:02X}" if byte >= 0x80 else chr(byte)
 
 
 def make_byte_level_fsm(fsm: FSM, keep_utf8=False) -> FSM:
@@ -415,21 +412,19 @@ def make_deterministic_fsm(fsm: FSM) -> Tuple[BetterFSM, Dict[int, int]]:
 @numba.njit(nogil=True, cache=True)
 def _walk_fsm(
     fsm_transitions: Dict[Tuple[int, int], int],
-    alphabet_symbol_mapping: Dict[str, int],
-    alphabet_anything_value: int,
     fsm_initial: int,
     fsm_finals: Set[int],
-    input_string: Sequence[str],
+    token_trans_key_seq: Sequence[int],
     start_state: int,
     full_match: bool = True,
 ) -> List[int]:
     state = start_state
     accepted_states: List[int] = numba.typed.List.empty_list(numba.int64)
     last_final_idx: int = numba.uint64(0)
 
-    for i, symbol in enumerate(input_string):
-        trans_key = alphabet_symbol_mapping.get(symbol, alphabet_anything_value)
-
+    # Iterate over token transition key sequence. The transition key
+    # sequence represents the FSM traversal rules of the tokens symbols.
+    for i, trans_key in enumerate(token_trans_key_seq):
         new_state = fsm_transitions.get((state, trans_key))
 
         if new_state is None:
@@ -453,7 +448,7 @@ def _walk_fsm(
 
 def walk_fsm(
     fsm: BetterFSM,
-    input_string: Sequence[str],
+    token_trans_key_seq: Sequence[int],
     start_state: int,
     full_match: bool = True,
 ) -> List[int]:
@@ -463,13 +458,11 @@ def walk_fsm(
     accepted_states: List[int] = []
     last_final_idx: int = 0
 
-    alphabet_symbol_mapping = fsm.alphabet._symbol_mapping
-    alphabet_anything_value = fsm.alphabet.anything_value
     fsm_transitions = fsm.flat_transition_map
 
-    for i, symbol in enumerate(input_string):
-        trans_key = alphabet_symbol_mapping.get(symbol, alphabet_anything_value)
-
+    # Iterate over token transition key sequence. The transition key
+    # sequence represents the FSM traversal rules of the tokens symbols.
+    for i, trans_key in enumerate(token_trans_key_seq):
         new_state = fsm_transitions.get((state, trans_key))
 
         if new_state is None:
@@ -655,24 +648,25 @@ def state_scan_tokens(
     alphabet_anything_value: int,
     fsm_initial: int,
     fsm_finals: Set[int],
-    vocabulary: List[Tuple[Sequence[str], Sequence[int]]],
+    vocabulary: List[Tuple[str, Sequence[int]]],
+    token_trans_key_seqs: List[Sequence[int]],
     start_state: int,
 ) -> Set[Tuple[int, int]]:
     res = set()
 
-    for token, token_ids in vocabulary:
+    for (token, token_ids), token_trans_key_seq in zip(
+        vocabulary, token_trans_key_seqs
+    ):
         state_seq = _walk_fsm(
             fsm_transitions,
-            alphabet_symbol_mapping,
-            alphabet_anything_value,
             fsm_initial,
             fsm_finals,
-            token,
+            token_trans_key_seq,
             start_state,
             False,
         )
 
-        if state_seq is not None and len(state_seq) < len(token):
+        if state_seq is not None and len(state_seq) < len(token_trans_key_seq):
             continue
 
         for token_id in token_ids:
@@ -681,9 +675,51 @@ def state_scan_tokens(
     return res
 
 
+@numba.njit(cache=True, nogil=True)
+def get_token_transitions(
+    alphabet_symbol_mapping: Dict[str, int],
+    alphabet_anything_value: int,
+    token_str: str,
+) -> Sequence[int]:
+    trans_key_seq = []
+    i = 0
+    while i < len(token_str):
+        if token_str[i] == "\x00" and i != len(token_str) - 1:
+            symbol = token_str[i : i + 3]
+            i += 3
+        else:
+            symbol = token_str[i]
+            i += 1
+
+        trans_key_seq.append(
+            alphabet_symbol_mapping.get(symbol, alphabet_anything_value)
+        )
+
+    trans_key_seq_array = np.empty(len(trans_key_seq), dtype=np.int64)
+    for j in range(len(trans_key_seq)):
+        trans_key_seq_array[j] = trans_key_seq[j]
+    return trans_key_seq_array
+
+
+@numba.njit(cache=True, nogil=True)
+def get_tokens_trans_keys(
+    alphabet_symbol_mapping: Dict[str, int],
+    alphabet_anything_value: int,
+    vocabulary: List[Tuple[str, Sequence[int]]],
+) -> List[Sequence[int]]:
+    tokens_trans_keys = numba.typed.List.empty_list(numba.int64[:])
+    for token_str, _ in vocabulary:
+        trans_key_seq_array = get_token_transitions(
+            alphabet_symbol_mapping, alphabet_anything_value, token_str
+        )
+        tokens_trans_keys.append(trans_key_seq_array)
+
+    return tokens_trans_keys
+
+
 def create_fsm_index_end_to_end(
     fsm_info: FSMInfo,
-    vocabulary: List[Tuple[Sequence[str], Sequence[int]]],
+    vocabulary: List[Tuple[str, Sequence[int]]],
 ) -> Dict[int, Set[Tuple[int, int]]]:
     """Create an FSM state-to-vocabulary map/index through end-to-end token parsing."""
 
@@ -699,6 +735,12 @@ def create_fsm_index_end_to_end(
         desc="Compiling FSM index for all state transitions",
     )
 
+    tokens_trans_key_seqs = get_tokens_trans_keys(
+        fsm_info.alphabet_symbol_mapping,
+        fsm_info.alphabet_anything_value,
+        vocabulary,
+    )
+
     while next_states:
         start_state = next_states.pop()
 
@@ -709,6 +751,7 @@ def create_fsm_index_end_to_end(
             fsm_info.initial,
             fsm_info.finals,
             vocabulary,
+            tokens_trans_key_seqs,
             start_state,
         )
 
@@ -771,7 +814,7 @@ def gpt2_unicode_to_bytes():
 @lru_cache
 def reduced_vocabulary(
     tokenizer: "Tokenizer",
-) -> Tuple[List[Tuple[Sequence[str], Sequence[int]]], Set[int]]:
+) -> Tuple[List[Tuple[str, Sequence[int]]], Set[int]]:
     """Create a map from decoded vocabulary tokens to lists of equivalent token ids."""
     empty_token_ids = set()
     vocabulary: Dict[Union[str, Tuple[str, ...]], List[int]] = {}
@@ -804,7 +847,7 @@ def reduced_vocabulary(
                         raise RuntimeError(
                             f"Cannot convert token `{token}` ({token_idx}) to bytes: {token_str}"
                         )
-                token_str = tuple(byte_symbol(b) for b in token_bytes)
+                token_str = "".join(byte_symbol(b) for b in token_bytes)
 
             vocabulary.setdefault(token_str, []).append(token_idx)
         else:
@@ -813,15 +856,14 @@ def reduced_vocabulary(
     vocabulary_nb = numba.typed.List.empty_list(
         numba.types.Tuple(
             (
-                nb_unichar_2_type[:],
+                nb_unicode_type,
                 numba.int64[:],
             )
         )
     )
-    for token_tuple, token_ids in vocabulary.items():
-        token_tuple_np = np.fromiter(token_tuple, dtype=np.dtype("U2"))
+    for token_str, token_ids in vocabulary.items():
         token_ids_np = np.fromiter(token_ids, dtype=np.dtype("int64"))
-        vocabulary_nb.append((token_tuple_np, token_ids_np))
+        vocabulary_nb.append((token_str, token_ids_np))
 
     return vocabulary_nb, empty_token_ids
 

outlines/integrations/llamacpp.py

@@ -26,7 +26,7 @@
 """
 
 import math
-from typing import TYPE_CHECKING, Dict, Optional, Set, Type, Union
+from typing import TYPE_CHECKING, Optional, Type, Union
 
 import numpy as np
 import torch
@@ -36,37 +36,12 @@
 from outlines.fsm.guide import CFGGuide, Guide, RegexGuide
 from outlines.fsm.json_schema import build_regex_from_schema
 from outlines.integrations.utils import convert_json_schema_to_str
+from outlines.models.llamacpp import LlamaCppTokenizer
 
 if TYPE_CHECKING:
     from llama_cpp import Llama
 
 
-class LlamaCppTokenizer:
-    def __init__(self, model: "Llama"):
-        self.eos_token_id = model.token_eos()
-        self.eos_token = model.tokenizer().decode([self.eos_token_id])
-        self.pad_token_id = self.eos_token_id
-        self.special_tokens: Set[int] = set()
-
-        self.vocabulary: Dict[str, int] = dict()
-
-        tokenizer = model.tokenizer()
-
-        self.decode = tokenizer.decode
-
-        # TODO: Remove when https://github.com/ggerganov/llama.cpp/pull/5613 is resolved
-        try:
-            self.vocabulary = model.tokenizer_.hf_tokenizer.get_vocab()
-        except AttributeError:
-            # ###
-            for t in range(model.n_vocab()):
-                token_piece = model.tokenizer().decode([t])
-                self.vocabulary[token_piece] = t
-
-    def convert_token_to_string(self, token: str) -> str:
-        return token
-
-
 class LogitsProcessor:
     """Bias LlamaCpp generation using a finite state machine.