Skip to content

armgit / LEANN

Public Imported from GitHub
0

[MLsys2026 Best Paper]: https://arxiv.org/abs/2506.08276. RAG on Everything with LEANN. Enjoy 97% storage savings while running a fast, accurate, and 100% private RAG application on your personal device.

add email test code

yichuan52003091032074ffd7ec646d
diff --git a/.gitignore b/.gitignore
index 26740c5..eb58011 100755
--- a/.gitignore
+++ b/.gitignore
@@ -9,6 +9,8 @@ demo/indices/
outputs/
*.pkl
*.pdf
+*.idx
+*.map
.history/
scripts/
lm_eval.egg-info/
diff --git a/examples/google_history_reader_leann.py b/examples/google_history_reader_leann.py
index 2c7c7a8..f8acf90 100644
--- a/examples/google_history_reader_leann.py
+++ b/examples/google_history_reader_leann.py
@@ -198,7 +198,16 @@ async def query_leann_index(index_path: str, query: str):
top_k=5,
recompute_beighbor_embeddings=True,
complexity=128,
- beam_width=1
+ beam_width=1,
+ llm_config={
+ "type": "openai",
+ "model": "gpt-4o",
+ "api_key": os.getenv("OPENAI_API_KEY"),
+ },
+ llm_kwargs={
+ "temperature": 0.0,
+ "max_tokens": 1000
+ }
)
print(f"Leann: {chat_response}")
@@ -206,7 +215,7 @@ async def main():
# Default Chrome profile path
default_chrome_profile = os.path.expanduser("~/Library/Application Support/Google/Chrome/Default")
- INDEX_DIR = Path("./chrome_history_index_leann")
+ INDEX_DIR = Path("./chrome_history_index_leann_test")
INDEX_PATH = str(INDEX_DIR / "chrome_history.leann")
# Find all Chrome profile directories
diff --git a/packages/leann-core/src/leann/api.py b/packages/leann-core/src/leann/api.py
index 2e90d8d..b8f2783 100644
--- a/packages/leann-core/src/leann/api.py
+++ b/packages/leann-core/src/leann/api.py
@@ -39,7 +39,8 @@ def compute_embeddings(chunks: List[str], model_name: str) -> np.ndarray:
model = model.to("mps")
# Generate embeddings
- embeddings = model.encode(chunks, convert_to_numpy=True, show_progress_bar=True, batch_size=64)
+ # give use an warning if OOM here means we need to turn down the batch size
+ embeddings = model.encode(chunks, convert_to_numpy=True, show_progress_bar=True, batch_size=256)
return embeddings
diff --git a/packages/leann-core/src/leann/chat.py b/packages/leann-core/src/leann/chat.py
index df83a57..47b6e41 100644
--- a/packages/leann-core/src/leann/chat.py
+++ b/packages/leann-core/src/leann/chat.py
@@ -209,10 +209,13 @@ def get_llm(llm_config: Optional[Dict[str, Any]] = None) -> LLMInterface:
An instance of an LLMInterface subclass.
"""
if llm_config is None:
- logger.info("No LLM config provided, defaulting to simulated chat.")
- return SimulatedChat()
+ llm_config = {
+ "type": "openai",
+ "model": "gpt-4o",
+ "api_key": os.getenv("OPENAI_API_KEY")
+ }
- llm_type = llm_config.get("type", "simulated")
+ llm_type = llm_config.get("type", "openai")
model = llm_config.get("model")
logger.info(f"Attempting to create LLM of type='{llm_type}' with model='{model}'")