mirror of
https://github.com/hwchase17/langchain
synced 2024-11-11 19:11:02 +00:00
a0c2281540
```python
"""python scripts/update_mypy_ruff.py"""
import glob
import tomllib
from pathlib import Path
import toml
import subprocess
import re
ROOT_DIR = Path(__file__).parents[1]
def main():
for path in glob.glob(str(ROOT_DIR / "libs/**/pyproject.toml"), recursive=True):
print(path)
with open(path, "rb") as f:
pyproject = tomllib.load(f)
try:
pyproject["tool"]["poetry"]["group"]["typing"]["dependencies"]["mypy"] = (
"^1.10"
)
pyproject["tool"]["poetry"]["group"]["lint"]["dependencies"]["ruff"] = (
"^0.5"
)
except KeyError:
continue
with open(path, "w") as f:
toml.dump(pyproject, f)
cwd = "/".join(path.split("/")[:-1])
completed = subprocess.run(
"poetry lock --no-update; poetry install --with typing; poetry run mypy . --no-color",
cwd=cwd,
shell=True,
capture_output=True,
text=True,
)
logs = completed.stdout.split("\n")
to_ignore = {}
for l in logs:
if re.match("^(.*)\:(\d+)\: error:.*\[(.*)\]", l):
path, line_no, error_type = re.match(
"^(.*)\:(\d+)\: error:.*\[(.*)\]", l
).groups()
if (path, line_no) in to_ignore:
to_ignore[(path, line_no)].append(error_type)
else:
to_ignore[(path, line_no)] = [error_type]
print(len(to_ignore))
for (error_path, line_no), error_types in to_ignore.items():
all_errors = ", ".join(error_types)
full_path = f"{cwd}/{error_path}"
try:
with open(full_path, "r") as f:
file_lines = f.readlines()
except FileNotFoundError:
continue
file_lines[int(line_no) - 1] = (
file_lines[int(line_no) - 1][:-1] + f" # type: ignore[{all_errors}]\n"
)
with open(full_path, "w") as f:
f.write("".join(file_lines))
subprocess.run(
"poetry run ruff format .; poetry run ruff --select I --fix .",
cwd=cwd,
shell=True,
capture_output=True,
text=True,
)
if __name__ == "__main__":
main()
```
85 lines
2.8 KiB
Python
85 lines
2.8 KiB
Python
"""Experimental implementation of lm-format-enforcer wrapped LLM."""
|
|
|
|
from __future__ import annotations
|
|
|
|
from typing import TYPE_CHECKING, Any, List, Optional
|
|
|
|
from langchain.schema import LLMResult
|
|
from langchain_community.llms.huggingface_pipeline import HuggingFacePipeline
|
|
from langchain_core.callbacks.manager import CallbackManagerForLLMRun
|
|
|
|
from langchain_experimental.pydantic_v1 import Field
|
|
|
|
if TYPE_CHECKING:
|
|
import lmformatenforcer
|
|
|
|
|
|
def import_lmformatenforcer() -> lmformatenforcer:
|
|
"""Lazily import of the lmformatenforcer package."""
|
|
try:
|
|
import lmformatenforcer
|
|
except ImportError:
|
|
raise ImportError(
|
|
"Could not import lmformatenforcer python package. "
|
|
"Please install it with `pip install lm-format-enforcer`."
|
|
)
|
|
return lmformatenforcer
|
|
|
|
|
|
class LMFormatEnforcer(HuggingFacePipeline):
|
|
"""LMFormatEnforcer wrapped LLM using HuggingFace Pipeline API.
|
|
|
|
This pipeline is experimental and not yet stable.
|
|
"""
|
|
|
|
json_schema: Optional[dict] = Field(
|
|
description="The JSON Schema to complete.", default=None
|
|
)
|
|
regex: Optional[str] = Field(
|
|
description="The regular expression to complete.", default=None
|
|
)
|
|
|
|
def _generate(
|
|
self,
|
|
prompts: List[str],
|
|
stop: Optional[List[str]] = None,
|
|
run_manager: Optional[CallbackManagerForLLMRun] = None,
|
|
**kwargs: Any,
|
|
) -> LLMResult:
|
|
lmformatenforcer = import_lmformatenforcer()
|
|
import lmformatenforcer.integrations.transformers as hf_integration
|
|
|
|
# We integrate lmformatenforcer by adding a prefix_allowed_tokens_fn.
|
|
# It has to be done on each call, because the prefix function is stateful.
|
|
if "prefix_allowed_tokens_fn" in self.pipeline._forward_params:
|
|
raise ValueError(
|
|
"prefix_allowed_tokens_fn param is forbidden with LMFormatEnforcer."
|
|
)
|
|
|
|
has_json_schema = self.json_schema is not None
|
|
has_regex = self.regex is not None
|
|
if has_json_schema == has_regex:
|
|
raise ValueError(
|
|
"You must specify exactly one of json_schema or a regex, but not both."
|
|
)
|
|
|
|
if has_json_schema:
|
|
parser = lmformatenforcer.JsonSchemaParser(self.json_schema)
|
|
else:
|
|
parser = lmformatenforcer.RegexParser(self.regex)
|
|
|
|
prefix_function = hf_integration.build_transformers_prefix_allowed_tokens_fn(
|
|
self.pipeline.tokenizer, parser
|
|
)
|
|
self.pipeline._forward_params["prefix_allowed_tokens_fn"] = prefix_function
|
|
|
|
result = super()._generate(
|
|
prompts,
|
|
stop=stop,
|
|
run_manager=run_manager,
|
|
**kwargs,
|
|
)
|
|
|
|
del self.pipeline._forward_params["prefix_allowed_tokens_fn"]
|
|
return result
|