llguidance 1.3.0__cp39-abi3-win_amd64.whl → 1.5.0__cp39-abi3-win_amd64.whl

llguidance/_lib.pyi

@@ -570,6 +570,26 @@ class LLExecutor:
         State rollback is performed to maintain matcher consistency.
         """
 
+    def consume_token_par(
+        self,
+        matchers: List[Tuple[LLMatcher, TokenId]],
+    ) -> List[bool]:
+        """
+        Consume a single token for each matcher in parallel.
+
+        Args:
+            matchers: List of tuples containing:
+                - LLMatcher: The matcher object
+                - TokenId: The token ID to consume
+
+        Returns:
+            List[bool]: Success/failure for each matcher (in order).
+
+        Note:
+            Matchers that fail (return False) are left in an error state,
+            consistent with the behavior of consume_token on a single matcher.
+        """
+
 class JsonCompileOptions(TypedDict, total=False):
     # defaults to ","
     item_separator: Optional[str]

llguidance/mlx.py

@@ -8,7 +8,7 @@ from .numpy import get_bitmask_shape, allocate_token_bitmask, fill_next_token_bi
 from numpy.typing import NDArray
 
 
-@mx.custom_function # type: ignore[misc]
+@mx.custom_function # type: ignore[untyped-decorator]
 def apply_token_bitmask_kernel(data: mx.array, mask: mx.array) -> mx.array:
     source = """
         uint batch = thread_position_in_grid.y;  // Batch index

llguidance/numpy.py

@@ -80,3 +80,22 @@ def fill_next_token_bitmask_par_with_draft_tokens(executor: LLExecutor,
     batch, vocab = bitmask.shape
     assert bitmask.flags["C_CONTIGUOUS"], "Mask must be contiguous"
     executor.unsafe_compute_mask_ptr_with_draft_token(matchers, bitmask.ctypes.data, vocab * 4, batch)
+
+
+def consume_token_par(executor: LLExecutor,
+                      matchers: List[Tuple[LLMatcher, int]]) -> List[bool]:
+    """
+    Consume a single token for each matcher in parallel.
+
+    Args:
+        executor: The LLExecutor to use for parallel execution.
+        matchers: List of tuples containing (LLMatcher, token_id).
+
+    Returns:
+        List[bool]: Success/failure for each matcher (in order).
+
+    Note:
+        Matchers that fail (return False) are left in an error state,
+        consistent with the behavior of consume_token on a single matcher.
+    """
+    return executor.consume_token_par(matchers)

llguidance/torch.py

@@ -77,3 +77,22 @@ def fill_next_token_bitmask_par_with_draft_tokens(executor: LLExecutor,
     batch, vocab = bitmask.shape
     assert bitmask.is_contiguous(), "Mask must be contiguous"
     executor.unsafe_compute_mask_ptr_with_draft_token(matchers, bitmask.data_ptr(), vocab * 4, batch)
+
+
+def consume_token_par(executor: LLExecutor,
+                      matchers: List[Tuple[LLMatcher, int]]) -> List[bool]:
+    """
+    Consume a single token for each matcher in parallel.
+
+    Args:
+        executor: The LLExecutor to use for parallel execution.
+        matchers: List of tuples containing (LLMatcher, token_id).
+
+    Returns:
+        List[bool]: Success/failure for each matcher (in order).
+
+    Note:
+        Matchers that fail (return False) are left in an error state,
+        consistent with the behavior of consume_token on a single matcher.
+    """
+    return executor.consume_token_par(matchers)

{llguidance-1.3.0.dist-info → llguidance-1.5.0.dist-info}/METADATA

@@ -1,15 +1,21 @@
 Metadata-Version: 2.4
 Name: llguidance
-Version: 1.3.0
+Version: 1.5.0
 License-File: LICENSE
 Summary: Bindings for the Low-level Guidance (llguidance) Rust library for use within Guidance
 Author: Michal Moskal
 License-Expression: MIT
-Requires-Python: >=3.9
+Requires-Python: >=3.10
 Description-Content-Type: text/markdown; charset=UTF-8; variant=GFM
-Project-URL: repository, https://github.com/microsoft/llguidance
 Project-URL: issue_tracker, https://github.com/microsoft/llguidance/issues
+Project-URL: repository, https://github.com/microsoft/llguidance
 
+<div align="right">
+  <a href="https://discord.gg/cjPfAK43dz"><img src="https://img.shields.io/badge/Discord-Join%20Us-5865F2?logo=discord&logoColor=white" alt="Discord"></a>
+  <a href="mailto:guidanceai@microsoft.com"><img src="https://img.shields.io/badge/Email-guidanceai%40microsoft.com-0078D4?logo=microsoft-outlook&logoColor=white" alt="Email"></a>
+  <img src="https://img.shields.io/badge/Hours-10am--2pm%20Pacific-gray" alt="Hours">
+</div>
+
 # Low-level Guidance (llguidance)
 
 <p align="center">
@@ -35,7 +41,7 @@ Project-URL: issue_tracker, https://github.com/microsoft/llguidance/issues
 ## About
 
 This library implements constrained decoding (also called constrained sampling or
-structured outputs) for Large Langauge Models (LLMs).
+structured outputs) for Large Language Models (LLMs).
 It can enforce arbitrary context-free grammar on the output of LLM
 and is fast - on the order of 50μs of CPU time per token
 (for 128k tokenizer) with negligible startup costs.
@@ -111,14 +117,14 @@ Thus, with 16 cores and a 10ms forward pass, llguidance can handle batch sizes u
 
 ## Building
 
-- [install rust](https://www.rust-lang.org/tools/install); 1.75 or later
+- [install rust](https://www.rust-lang.org/tools/install); 1.87 or later
 
 If you just need the C or Rust library (`llguidance`), 
 check the [parser](./parser/README.md) directory.
 
 For Python bindings:
 
-- install python 3.9 or later; very likely you'll need a virtual env/conda
+- install python 3.10 or later; very likely you'll need a virtual env/conda
 - run `./scripts/install-deps.sh`
 - to build and after any changes, run `./scripts/test-guidance.sh`
 

llguidance-1.5.0.dist-info/RECORD

@@ -0,0 +1,20 @@
+llguidance\__init__.py,sha256=tlrM-GzMdClgSIEblgA4ixP10h51VDJY5ejIP5FFJJw,620
+llguidance\_grammar_from.py,sha256=kEgwhVYBq3NE_O9MGjIRpci7dgOneLC9CvVGWpKEO84,2452
+llguidance\_lib.pyd,sha256=cnSFoKOc8xH8BUF2z-ywPDoj0TqqnEnSqPy29qia19g,7291904
+llguidance\_lib.pyi,sha256=ov5iAQT137lnMSQuLZ9TyfgP9TsOEbhsO7R-6V7Czp8,25530
+llguidance\_struct_tag.py,sha256=wEJd9KGZTpS-R1ZXyyRFmr-48aYOZw_tI4PhIRy9h4g,4548
+llguidance\_tokenizer.py,sha256=9jonF41IUypGtV6lVUl9l9FGAyn9HNnDb9HKNDywZuQ,1120
+llguidance\_util.py,sha256=6RMYO61KUynAym7KcbDn9VEMGcGdv7rj2yo6fITC-u8,274
+llguidance\cli.py,sha256=-vXUfe8u-qdFyaFjwDNd0OjUZ0UvKERl7Zr-Ea9KvKc,2461
+llguidance\gbnf_to_lark.py,sha256=YsFNyVXrzG3-cjqlcrcuoG1RdquVtZJz43X_lSAXL7Y,17187
+llguidance\hf.py,sha256=3LNNqC4_KafPe6bKP-Ba0aGsOVg3MqvuMvY6QYth0fY,1980
+llguidance\llamacpp.py,sha256=2LdNjiNku-8xGcswGEQga57FUsRgzdf4Z4BrLVBUfAg,2352
+llguidance\mlx.py,sha256=iOWqkhYItQRDOcQtOnH2tmtYmtiQGD_Y_RIYjYe4DZY,2507
+llguidance\numpy.py,sha256=NMJxvI_rG2m_kYLDXuSAnxJfjkea9hvcUIyvl70OhnM,4048
+llguidance\py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+llguidance\tiktoken.py,sha256=sE4QWj88dANfdM0Y50ZvFI7ku6zBFomUCE4bRTtewF4,1135
+llguidance\torch.py,sha256=Kr-mlF1FXQD4Iwuv_AvSpj_kXKX9FVTw-_un8JK9deU,4090
+llguidance-1.5.0.dist-info\METADATA,sha256=RSr9amkc8jFvwW_LqUP-0bL7soHZz7MwVrhRj9VekB0,10893
+llguidance-1.5.0.dist-info\WHEEL,sha256=OD0Is1kLHE07aD7XukSVDFU8ymUMI4Fdg0tKYWce3N0,95
+llguidance-1.5.0.dist-info\licenses\LICENSE,sha256=mQaUD2Gx8LUz-n2ZuvVReLKAj74RPqUd-_rYVyzNXys,1162
+llguidance-1.5.0.dist-info\RECORD,,

{llguidance-1.3.0.dist-info → llguidance-1.5.0.dist-info}/WHEEL

@@ -1,4 +1,4 @@
 Wheel-Version: 1.0
-Generator: maturin (1.9.6)
+Generator: maturin (1.11.5)
 Root-Is-Purelib: false
 Tag: cp39-abi3-win_amd64

llguidance-1.3.0.dist-info/RECORD

@@ -1,20 +0,0 @@
-llguidance-1.3.0.dist-info/METADATA,sha256=2y5QnhlB2fabweFtiAIpeabETYQh9xdTFSjyJjyElO0,10436
-llguidance-1.3.0.dist-info/WHEEL,sha256=dtDRtmb8ugdlQQiOTIfQg6qwccecUkp34WgpWvjGLGA,94
-llguidance-1.3.0.dist-info/licenses/LICENSE,sha256=mQaUD2Gx8LUz-n2ZuvVReLKAj74RPqUd-_rYVyzNXys,1162
-llguidance/__init__.py,sha256=tlrM-GzMdClgSIEblgA4ixP10h51VDJY5ejIP5FFJJw,620
-llguidance/_grammar_from.py,sha256=kEgwhVYBq3NE_O9MGjIRpci7dgOneLC9CvVGWpKEO84,2452
-llguidance/_lib.pyd,sha256=f1re0RSpRXHw6hDwX-2F-BI398uGWzDcha3PzVfoYwY,7515136
-llguidance/_lib.pyi,sha256=tc0mb-ZKH3LjodEf_aIdLPXFLuEbqGI3flxQpxZn-Bw,24893
-llguidance/_struct_tag.py,sha256=wEJd9KGZTpS-R1ZXyyRFmr-48aYOZw_tI4PhIRy9h4g,4548
-llguidance/_tokenizer.py,sha256=9jonF41IUypGtV6lVUl9l9FGAyn9HNnDb9HKNDywZuQ,1120
-llguidance/_util.py,sha256=6RMYO61KUynAym7KcbDn9VEMGcGdv7rj2yo6fITC-u8,274
-llguidance/cli.py,sha256=-vXUfe8u-qdFyaFjwDNd0OjUZ0UvKERl7Zr-Ea9KvKc,2461
-llguidance/gbnf_to_lark.py,sha256=YsFNyVXrzG3-cjqlcrcuoG1RdquVtZJz43X_lSAXL7Y,17187
-llguidance/hf.py,sha256=3LNNqC4_KafPe6bKP-Ba0aGsOVg3MqvuMvY6QYth0fY,1980
-llguidance/llamacpp.py,sha256=2LdNjiNku-8xGcswGEQga57FUsRgzdf4Z4BrLVBUfAg,2352
-llguidance/mlx.py,sha256=Q2cNocnFB3xqCCg614iE0wPcumBFJBvo0pG5JTpShds,2494
-llguidance/numpy.py,sha256=eMIihG5rcPWFMMWjwl-ZPSkcEo8BXKZFUHaIHkxnFhw,3409
-llguidance/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-llguidance/tiktoken.py,sha256=sE4QWj88dANfdM0Y50ZvFI7ku6zBFomUCE4bRTtewF4,1135
-llguidance/torch.py,sha256=PzTI7wRiznXnH7u4_iEP6NL2wmL2cN8p-UlR0PE8JJ4,3451
-llguidance-1.3.0.dist-info/RECORD,,