Name and Version
development
What is the problem this feature will solve?
examples/langchain_rag/app.py uses outdated LangChain import paths that are incompatible with currently pinned dependencies in examples/langchain_rag/requirements.txt.
Repro steps:
cd examples/langchain_ragpython -m venv .venv && source .venv/bin/activatepip install -r requirements.txtpython app.py
Current result: app fails at import time with errors like:
ModuleNotFoundError: No module named 'langchain.chains'ImportError: cannot import name 'CacheBackedEmbeddings' from 'langchain.embeddings'ModuleNotFoundError: No module named 'langchain.globals'ModuleNotFoundError: No module named 'langchain.storage'
Expected result: example should start without import-time failures (runtime config errors like missing DIAL_URL are acceptable).
What is the feature you are proposing to solve the problem?
Update examples/langchain_rag/app.py imports to match modern LangChain package split used by pinned versions:
RetrievalQA import: langchain.chains... -> langchain_classic.chains.retrieval_qa.baseCacheBackedEmbeddings import: langchain.embeddings -> langchain_classic.embeddingsset_debug import: langchain.globals -> langchain_core.globalsLocalFileStore import: langchain.storage -> langchain_classic.storage
Acceptance criteria:
python app.py no longer fails on import errors in examples/langchain_rag.
What alternatives have you considered?
No response
Name and Version
development
What is the problem this feature will solve?
examples/langchain_rag/app.pyuses outdated LangChain import paths that are incompatible with currently pinned dependencies inexamples/langchain_rag/requirements.txt.Repro steps:
cd examples/langchain_ragpython -m venv .venv && source .venv/bin/activatepip install -r requirements.txtpython app.pyCurrent result: app fails at import time with errors like:
ModuleNotFoundError: No module named 'langchain.chains'ImportError: cannot import name 'CacheBackedEmbeddings' from 'langchain.embeddings'ModuleNotFoundError: No module named 'langchain.globals'ModuleNotFoundError: No module named 'langchain.storage'Expected result: example should start without import-time failures (runtime config errors like missing
DIAL_URLare acceptable).What is the feature you are proposing to solve the problem?
Update
examples/langchain_rag/app.pyimports to match modern LangChain package split used by pinned versions:RetrievalQAimport:langchain.chains...->langchain_classic.chains.retrieval_qa.baseCacheBackedEmbeddingsimport:langchain.embeddings->langchain_classic.embeddingsset_debugimport:langchain.globals->langchain_core.globalsLocalFileStoreimport:langchain.storage->langchain_classic.storageAcceptance criteria:
python app.pyno longer fails on import errors inexamples/langchain_rag.What alternatives have you considered?
No response