mirror of
https://github.com/hwchase17/langchain.git
synced 2025-06-23 15:19:33 +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),
}
```
123 lines
4.0 KiB
Python
123 lines
4.0 KiB
Python
from __future__ import annotations
|
|
|
|
from copy import deepcopy
|
|
from enum import Enum
|
|
from typing import TYPE_CHECKING, Any, Dict, Optional, Sequence
|
|
|
|
from langchain.retrievers.document_compressors.base import BaseDocumentCompressor
|
|
from langchain_core.callbacks.manager import Callbacks
|
|
from langchain_core.documents import Document
|
|
from langchain_core.pydantic_v1 import Field, PrivateAttr, root_validator
|
|
from langchain_core.utils import get_from_dict_or_env
|
|
|
|
if TYPE_CHECKING:
|
|
from rank_llm.data import Candidate, Query, Request
|
|
else:
|
|
# Avoid pydantic annotation issues when actually instantiating
|
|
# while keeping this import optional
|
|
try:
|
|
from rank_llm.data import Candidate, Query, Request
|
|
except ImportError:
|
|
pass
|
|
|
|
|
|
class RankLLMRerank(BaseDocumentCompressor):
|
|
"""Document compressor using Flashrank interface."""
|
|
|
|
client: Any = None
|
|
"""RankLLM client to use for compressing documents"""
|
|
top_n: int = Field(default=3)
|
|
"""Top N documents to return."""
|
|
model: str = Field(default="zephyr")
|
|
"""Name of model to use for reranking."""
|
|
step_size: int = Field(default=10)
|
|
"""Step size for moving sliding window."""
|
|
gpt_model: str = Field(default="gpt-3.5-turbo")
|
|
"""OpenAI model name."""
|
|
_retriever: Any = PrivateAttr()
|
|
|
|
class Config:
|
|
arbitrary_types_allowed = True
|
|
extra = "forbid"
|
|
|
|
@root_validator(pre=True)
|
|
def validate_environment(cls, values: Dict) -> Dict:
|
|
"""Validate python package exists in environment."""
|
|
|
|
if not values.get("client"):
|
|
client_name = values.get("model", "zephyr")
|
|
|
|
try:
|
|
model_enum = ModelType(client_name.lower())
|
|
except ValueError:
|
|
raise ValueError(
|
|
"Unsupported model type. Please use 'vicuna', 'zephyr', or 'gpt'."
|
|
)
|
|
|
|
try:
|
|
if model_enum == ModelType.VICUNA:
|
|
from rank_llm.rerank.vicuna_reranker import VicunaReranker
|
|
|
|
values["client"] = VicunaReranker()
|
|
elif model_enum == ModelType.ZEPHYR:
|
|
from rank_llm.rerank.zephyr_reranker import ZephyrReranker
|
|
|
|
values["client"] = ZephyrReranker()
|
|
elif model_enum == ModelType.GPT:
|
|
from rank_llm.rerank.rank_gpt import SafeOpenai
|
|
from rank_llm.rerank.reranker import Reranker
|
|
|
|
openai_api_key = get_from_dict_or_env(
|
|
values, "open_api_key", "OPENAI_API_KEY"
|
|
)
|
|
|
|
agent = SafeOpenai(
|
|
model=values["gpt_model"],
|
|
context_size=4096,
|
|
keys=openai_api_key,
|
|
)
|
|
values["client"] = Reranker(agent)
|
|
|
|
except ImportError:
|
|
raise ImportError(
|
|
"Could not import rank_llm python package. "
|
|
"Please install it with `pip install rank_llm`."
|
|
)
|
|
|
|
return values
|
|
|
|
def compress_documents(
|
|
self,
|
|
documents: Sequence[Document],
|
|
query: str,
|
|
callbacks: Optional[Callbacks] = None,
|
|
) -> Sequence[Document]:
|
|
request = Request(
|
|
query=Query(text=query, qid=1),
|
|
candidates=[
|
|
Candidate(doc={"text": doc.page_content}, docid=index, score=1)
|
|
for index, doc in enumerate(documents)
|
|
],
|
|
)
|
|
|
|
rerank_results = self.client.rerank(
|
|
request,
|
|
rank_end=len(documents),
|
|
window_size=min(20, len(documents)),
|
|
step=10,
|
|
)
|
|
|
|
final_results = []
|
|
for res in rerank_results.candidates:
|
|
doc = documents[int(res.docid)]
|
|
doc_copy = Document(doc.page_content, metadata=deepcopy(doc.metadata))
|
|
final_results.append(doc_copy)
|
|
|
|
return final_results[: self.top_n]
|
|
|
|
|
|
class ModelType(Enum):
|
|
VICUNA = "vicuna"
|
|
ZEPHYR = "zephyr"
|
|
GPT = "gpt"
|