Updated RankLLM integration

[clides]

Updated existing RankLLM integration for langchain-community, with the latest params, function names, etc.

Example with updated usage for RankLLM:

Install all the packages:

pip install langchain-community faiss-gpu torch transformers sentence-transformers huggingface-hub rank_llm

Install the document example:
https://github.com/hwchase17/chat-your-data/blob/master/state_of_the_union.txt

Set up the base vector store retriever:

from langchain_community.document_loaders import TextLoader
from langchain_community.vectorstores import FAISS
from langchain_community.embeddings import HuggingFaceEmbeddings
from langchain_text_splitters import RecursiveCharacterTextSplitter
import torch
import os

device = "cuda"

documents = TextLoader("state_of_the_union.txt").load()
text_splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=100)
texts = text_splitter.split_documents(documents)
for idx, text in enumerate(texts):
    text.metadata["id"] = idx

embedding = HuggingFaceEmbeddings(
    model_name="BAAI/bge-small-en", # or any model of your choice
    model_kwargs={'device': 'cuda'},
    encode_kwargs={'normalize_embeddings': True}
)

retriever = FAISS.from_documents(texts, embedding).as_retriever(search_kwargs={"k": 20})

Retrieval without reranking:

query = "What was done to Russia?"
docs = retriever.invoke(query)
pretty_print_docs(docs)

All the field arguments to RankLLMRerank:

model_path: str = Field(default="rank_zephyr")
top_n: int = Field(default=3)
window_size: int = Field(default=20)
context_size: int = Field(default=4096)
prompt_mode: str = Field(default="rank_GPT")
num_gpus: int = Field(default=1)
num_few_shot_examples: int = Field(default=0)
few_shot_file: Optional[str] = Field(default=None)
use_logits: bool = Field(default=False)
use_alpha: bool = Field(default=False)
variable_passages: bool = Field(default=False)
stride: int = Field(default=10)
use_azure_openai: bool = Field(default=False)
model_coordinator: Any = Field(default=None, exclude=True)

Retrieval with reranking (default RankLLM model is rank_zephyr):

torch.cuda.empty_cache()

compressor = RankLLMRerank(top_n=3, model_path="rank_zephyr")
compression_retriever = ContextualCompressionRetriever(
    base_compressor=compressor, base_retriever=retriever
)

del compressor

compressed_docs = compression_retriever.invoke(query)
pretty_print_docs(compressed_docs)

[eyurtsev]

Hi @clides thanks for the contribution. At the moment, this code isn't tested or documented at a level where it can be merged to langchain-community.

As a rule of thumb the amount of documentation should be similar to the amount of code, and there should be likely some examples with usage.

It looks like there are some breaking changes in this code? Is that correct?

[clides]

It looks like there are some breaking changes in this code? Is that correct?

Yep, there are some changes with RankLLM (ie new params, changed function/class names, etc)

[clides]

As a rule of thumb the amount of documentation should be similar to the amount of code, and there should be likely some examples with usage.

Added more comments as requested