Update retriever imports to use langchain_classic for v1 compatibility (#1196)

The langchain.retrievers module was removed in LangChain 1.0+. 
All retriever imports have been updated to use langchain_classic as
documented in the v1 migration guide.

Fixes #1195

## Summary

Updated outdated retriever imports that don't work in LangChain 1.0+.
Changed `from langchain.retrievers` to `from
langchain_classic.retrievers` as per the v1 migration guide.

## Changes

31 files updated across integrations:
- 9 document_transformers files
- 7 retrievers files
- 12 providers files
- 2 vectorstores files
- 1 document_loaders file

Before:
```python
from langchain.retrievers import ContextualCompressionRetriever
from langchain.retrievers.multi_query import MultiQueryRetriever
```

After:
```python
from langchain_classic.retrievers.contextual_compression import ContextualCompressionRetriever
from langchain_classic.retrievers.multi_query import MultiQueryRetriever
```

All changes follow the pattern in the [v1 migration
guide](https://docs.langchain.com/oss/migrate/langchain-v1#langchain-classic).

Co-authored-by: bart0401 <bart0401@users.noreply.github.com>
Co-authored-by: Mason Daugherty <mason@langchain.dev>

Author: bart0401

src/oss/python/integrations/callbacks/uptrain.mdx

@@ -62,9 +62,9 @@ NOTE: that you can also install `faiss-gpu` instead of `faiss-cpu` if you want t
 from getpass import getpass
 
 from langchain.chains import RetrievalQA
-from langchain.retrievers import ContextualCompressionRetriever
-from langchain.retrievers.document_compressors import FlashrankRerank
-from langchain.retrievers.multi_query import MultiQueryRetriever
+from langchain_classic.retrievers.contextual_compression import ContextualCompressionRetriever
+from langchain_classic.retrievers.document_compressors import FlashrankRerank
+from langchain_classic.retrievers.multi_query import MultiQueryRetriever
 from langchain_community.callbacks.uptrain_callback import UpTrainCallbackHandler
 from langchain_community.document_loaders import TextLoader
 from langchain_community.vectorstores import FAISS

src/oss/python/integrations/document_loaders/docugami.mdx

@@ -204,7 +204,7 @@ We can use a self-querying retriever to improve our query accuracy, using this a
 
 ```python
 from langchain.chains.query_constructor.schema import AttributeInfo
-from langchain.retrievers.self_query.base import SelfQueryRetriever
+from langchain_classic.retrievers.self_query.base import SelfQueryRetriever
 from langchain_chroma import Chroma
 
 EXCLUDE_KEYS = ["id", "xpath", "structure"]
@@ -322,7 +322,7 @@ CHUNK 21b4d9517f7ccdc0e3a028ce5043a2a0: page_content='1.1 Landlord.\n <Landlord>
 ```
 
 ```python
-from langchain.retrievers.multi_vector import MultiVectorRetriever, SearchType
+from langchain_classic.retrievers.multi_vector import MultiVectorRetriever, SearchType
 from langchain.storage import InMemoryStore
 from langchain_chroma import Chroma
 from langchain_openai import OpenAIEmbeddings

src/oss/python/integrations/document_transformers/cross_encoder_reranker.mdx

@@ -58,8 +58,8 @@ pretty_print_docs(docs)
 Now let's wrap our base retriever with a `ContextualCompressionRetriever`. `CrossEncoderReranker` uses `HuggingFaceCrossEncoder` to rerank the returned results.
 
 ```python
-from langchain.retrievers import ContextualCompressionRetriever
-from langchain.retrievers.document_compressors import CrossEncoderReranker
+from langchain_classic.retrievers.contextual_compression import ContextualCompressionRetriever
+from langchain_classic.retrievers.document_compressors import CrossEncoderReranker
 from langchain_community.cross_encoders import HuggingFaceCrossEncoder
 
 model = HuggingFaceCrossEncoder(model_name="BAAI/bge-reranker-base")

src/oss/python/integrations/document_transformers/dashscope_rerank.mdx

@@ -263,7 +263,7 @@ And with an unwavering resolve that freedom will always triumph over tyranny.
 Now let's wrap our base retriever with a `ContextualCompressionRetriever`. We'll use the `DashScopeRerank` to rerank the returned results.
 
 ```python
-from langchain.retrievers import ContextualCompressionRetriever
+from langchain_classic.retrievers.contextual_compression import ContextualCompressionRetriever
 from langchain_community.document_compressors.dashscope_rerank import DashScopeRerank
 
 compressor = DashScopeRerank()

src/oss/python/integrations/document_transformers/google_cloud_vertexai_rerank.mdx

@@ -64,7 +64,7 @@ Your 1 documents have been split into 266 chunks
 
 ```python
 import pandas as pd
-from langchain.retrievers.contextual_compression import ContextualCompressionRetriever
+from langchain_classic.retrievers.contextual_compression import ContextualCompressionRetriever
 from langchain_google_community.vertex_rank import VertexAIRank
 
 # Instantiate the VertexAIReranker with the SDK manager

src/oss/python/integrations/document_transformers/infinity_rerank.mdx

@@ -280,7 +280,7 @@ Now let's wrap our base retriever with a `ContextualCompressionRetriever`. We'll
 
 ```python
 from infinity_client import Client
-from langchain.retrievers import ContextualCompressionRetriever
+from langchain_classic.retrievers.contextual_compression import ContextualCompressionRetriever
 from langchain_community.document_compressors.infinity_rerank import InfinityRerank
 
 client = Client(base_url="http://localhost:7997")

src/oss/python/integrations/document_transformers/jina_rerank.mdx

@@ -65,7 +65,7 @@ pretty_print_docs(docs)
 Now let's wrap our base retriever with a ContextualCompressionRetriever, using Jina Reranker as a compressor.
 
 ```python
-from langchain.retrievers import ContextualCompressionRetriever
+from langchain_classic.retrievers.contextual_compression import ContextualCompressionRetriever
 from langchain_community.document_compressors import JinaRerank
 
 compressor = JinaRerank()

src/oss/python/integrations/document_transformers/openvino_rerank.mdx

@@ -291,7 +291,7 @@ Metadata: {'source': '../../how_to/state_of_the_union.txt', 'id': 40}
 Now let's wrap our base retriever with a `ContextualCompressionRetriever`, using `OpenVINOReranker` as a compressor.
 
 ```python
-from langchain.retrievers import ContextualCompressionRetriever
+from langchain_classic.retrievers.contextual_compression import ContextualCompressionRetriever
 from langchain_community.document_compressors.openvino_rerank import OpenVINOReranker
 
 model_name = "BAAI/bge-reranker-large"

src/oss/python/integrations/document_transformers/rankllm-reranker.mdx

@@ -270,7 +270,7 @@ RankZephyr performs listwise reranking for improved retrieval quality but requir
 
 ```python
 import torch
-from langchain.retrievers.contextual_compression import ContextualCompressionRetriever
+from langchain_classic.retrievers.contextual_compression import ContextualCompressionRetriever
 from langchain_community.document_compressors.rankllm_rerank import RankLLMRerank
 
 torch.cuda.empty_cache()
@@ -567,7 +567,7 @@ One America.
 Retrieval + Reranking with RankGPT
 
 ```python
-from langchain.retrievers.contextual_compression import ContextualCompressionRetriever
+from langchain_classic.retrievers.contextual_compression import ContextualCompressionRetriever
 from langchain_community.document_compressors.rankllm_rerank import RankLLMRerank
 
 compressor = RankLLMRerank(top_n=3, model="gpt", gpt_model="gpt-4o-mini")

src/oss/python/integrations/document_transformers/volcengine_rerank.mdx

@@ -295,7 +295,7 @@ To disable this warning, you can either:
 Now let's wrap our base retriever with a `ContextualCompressionRetriever`. We'll use the `VolcengineRerank` to rerank the returned results.
 
 ```python
-from langchain.retrievers import ContextualCompressionRetriever
+from langchain_classic.retrievers.contextual_compression import ContextualCompressionRetriever
 from langchain_community.document_compressors.volcengine_rerank import VolcengineRerank
 
 compressor = VolcengineRerank()