Outputs for LLM-RAG

[sarthh7777]

INPUT COMMAND-
pip install -r requirements.txt
coverage run --source=llmrag -m unittest discover -s tests
coverage report -m

Result :
Device set to use cpu
C:\Users\dell.cache\chroma\onnx_models\all-MiniLM-L6-v2\onnx.tar.gz: 100%|███████████████████████████████████████████████| 79.3M/79.3M [00:17<00:00, 4.69MiB/s]
..Retrieved: [('Paris is the capital of France.', 0.28786033391952515)]
.Retrieved: [('Paris is the capital of France.', 0.3702659010887146)]
.

Ran 6 tests in 114.034s

OK
PS C:\Users\dell\Downloads\llmrag> coverage report -m
Name Stmts Miss Cover Missing

llmrag\chunking_init_.py 1 0 100%
llmrag\chunking\text_splitter.py 8 0 100%
llmrag\cli.py 52 52 0% 1-85
llmrag\embeddings_init_.py 3 0 100%
llmrag\embeddings\base_embedder.py 5 1 80% 6
llmrag\embeddings\sentence_transformers_embedder.py 12 0 100%
llmrag\generators_init_.py 0 0 100%
llmrag\generators\local_generator.py 7 3 57% 16, 29-35
llmrag\main.py 56 56 0% 1-72
llmrag\models_init_.py 3 0 100%
llmrag\models\base_model.py 5 1 80% 6
llmrag\models\transformers_model.py 8 0 100%
llmrag\pipelines_init_.py 1 0 100%
llmrag\pipelines\rag_pipeline.py 21 6 71% 19, 21, 61-64
llmrag\retrievers_init_.py 12 7 42% 7-14
llmrag\retrievers\base_vector_store.py 8 2 75% 6, 10
llmrag\retrievers\chroma_store.py 42 3 93% 48, 67-68
llmrag\retrievers\faiss_store.py 27 27 0% 1-31
llmrag\streamlit_app.py 29 29 0% 1-53

TOTAL

CONCLUSION-
ALL TESTS RAN SUCCESSFULLY

[onestardao]

this is a textbook ghost eval case — everything looks fine (test pass, coverage 100%), but the retrieval results are fixed-patterns, not semantically triggered.
we documented this as No.9 in our failure map — often caused by unweighted vector fallback or pre-initialized dummy memories.
ping me if you want to patch it out structurally, not just at surface level.