mirror of
https://github.com/hwchase17/langchain
synced 2024-11-10 01:10:59 +00:00
bf5193bb99
Upgrade to using a literal for specifying the extra which is the
recommended approach in pydantic 2.
This works correctly also in pydantic v1.
```python
from pydantic.v1 import BaseModel
class Foo(BaseModel, extra="forbid"):
x: int
Foo(x=5, y=1)
```
And
```python
from pydantic.v1 import BaseModel
class Foo(BaseModel):
x: int
class Config:
extra = "forbid"
Foo(x=5, y=1)
```
## Enum -> literal using grit pattern:
```
engine marzano(0.1)
language python
or {
`extra=Extra.allow` => `extra="allow"`,
`extra=Extra.forbid` => `extra="forbid"`,
`extra=Extra.ignore` => `extra="ignore"`
}
```
Resorted attributes in config and removed doc-string in case we will
need to deal with going back and forth between pydantic v1 and v2 during
the 0.3 release. (This will reduce merge conflicts.)
## Sort attributes in Config:
```
engine marzano(0.1)
language python
function sort($values) js {
return $values.text.split(',').sort().join("\n");
}
class_definition($name, $body) as $C where {
$name <: `Config`,
$body <: block($statements),
$values = [],
$statements <: some bubble($values) assignment() as $A where {
$values += $A
},
$body => sort($values),
}
```
116 lines
3.9 KiB
Python
116 lines
3.9 KiB
Python
import logging
|
|
from typing import Any, Dict, List, Optional
|
|
|
|
from langchain_core.callbacks import CallbackManagerForLLMRun
|
|
from langchain_core.language_models import BaseLanguageModel
|
|
from langchain_core.language_models.llms import LLM
|
|
from langchain_core.messages import AIMessage
|
|
from langchain_core.utils import get_from_dict_or_env, pre_init
|
|
|
|
logger = logging.getLogger(__name__)
|
|
|
|
|
|
class OpaquePrompts(LLM):
|
|
"""LLM that uses OpaquePrompts to sanitize prompts.
|
|
|
|
Wraps another LLM and sanitizes prompts before passing it to the LLM, then
|
|
de-sanitizes the response.
|
|
|
|
To use, you should have the ``opaqueprompts`` python package installed,
|
|
and the environment variable ``OPAQUEPROMPTS_API_KEY`` set with
|
|
your API key, or pass it as a named parameter to the constructor.
|
|
|
|
Example:
|
|
.. code-block:: python
|
|
|
|
from langchain_community.llms import OpaquePrompts
|
|
from langchain_community.chat_models import ChatOpenAI
|
|
|
|
op_llm = OpaquePrompts(base_llm=ChatOpenAI())
|
|
"""
|
|
|
|
base_llm: BaseLanguageModel
|
|
"""The base LLM to use."""
|
|
|
|
class Config:
|
|
extra = "forbid"
|
|
|
|
@pre_init
|
|
def validate_environment(cls, values: Dict) -> Dict:
|
|
"""Validates that the OpaquePrompts API key and the Python package exist."""
|
|
try:
|
|
import opaqueprompts as op
|
|
except ImportError:
|
|
raise ImportError(
|
|
"Could not import the `opaqueprompts` Python package, "
|
|
"please install it with `pip install opaqueprompts`."
|
|
)
|
|
if op.__package__ is None:
|
|
raise ValueError(
|
|
"Could not properly import `opaqueprompts`, "
|
|
"opaqueprompts.__package__ is None."
|
|
)
|
|
|
|
api_key = get_from_dict_or_env(
|
|
values, "opaqueprompts_api_key", "OPAQUEPROMPTS_API_KEY", default=""
|
|
)
|
|
if not api_key:
|
|
raise ValueError(
|
|
"Could not find OPAQUEPROMPTS_API_KEY in the environment. "
|
|
"Please set it to your OpaquePrompts API key."
|
|
"You can get it by creating an account on the OpaquePrompts website: "
|
|
"https://opaqueprompts.opaque.co/ ."
|
|
)
|
|
return values
|
|
|
|
def _call(
|
|
self,
|
|
prompt: str,
|
|
stop: Optional[List[str]] = None,
|
|
run_manager: Optional[CallbackManagerForLLMRun] = None,
|
|
**kwargs: Any,
|
|
) -> str:
|
|
"""Call base LLM with sanitization before and de-sanitization after.
|
|
|
|
Args:
|
|
prompt: The prompt to pass into the model.
|
|
|
|
Returns:
|
|
The string generated by the model.
|
|
|
|
Example:
|
|
.. code-block:: python
|
|
|
|
response = op_llm.invoke("Tell me a joke.")
|
|
"""
|
|
import opaqueprompts as op
|
|
|
|
_run_manager = run_manager or CallbackManagerForLLMRun.get_noop_manager()
|
|
|
|
# sanitize the prompt by replacing the sensitive information with a placeholder
|
|
sanitize_response: op.SanitizeResponse = op.sanitize([prompt])
|
|
sanitized_prompt_value_str = sanitize_response.sanitized_texts[0]
|
|
|
|
# TODO: Add in callbacks once child runs for LLMs are supported by LangSmith.
|
|
# call the LLM with the sanitized prompt and get the response
|
|
llm_response = self.base_llm.bind(stop=stop).invoke(
|
|
sanitized_prompt_value_str,
|
|
)
|
|
if isinstance(llm_response, AIMessage):
|
|
llm_response = llm_response.content
|
|
|
|
# desanitize the response by restoring the original sensitive information
|
|
desanitize_response: op.DesanitizeResponse = op.desanitize(
|
|
llm_response,
|
|
secure_context=sanitize_response.secure_context,
|
|
)
|
|
return desanitize_response.desanitized_text
|
|
|
|
@property
|
|
def _llm_type(self) -> str:
|
|
"""Return type of LLM.
|
|
|
|
This is an override of the base class method.
|
|
"""
|
|
return "opaqueprompts"
|