TBNilles/LEANN
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: mlx when searching, added to embedding_server

Browse Source
This commit is contained in:
Andy Lee
2025-07-14 01:11:21 -07:00
parent 8b4654921b
commit 3da5b44d7f
8 changed files with 315 additions and 885 deletions
Download Patch File Download Diff File Expand all files Collapse all files

35
README.md
View File

@@ -303,6 +303,41 @@ Once the index is built, you can ask questions like:
</details>
## ⚡ Performance Comparison
### LEANN vs Faiss HNSW
We benchmarked LEANN against the popular Faiss HNSW implementation to demonstrate the significant memory and storage savings our approach provides:
```bash
# Run the comparison benchmark
python examples/compare_faiss_vs_leann.py
```
#### 🎯 Results Summary
| Metric | Faiss HNSW | LEANN HNSW | **Improvement** |
|--------|------------|-------------|-----------------|
| **Peak Memory** | 887.0 MB | 618.2 MB | **1.4x less** (268.8 MB saved) |
| **Storage Size** | 5.5 MB | 0.5 MB | **11.4x smaller** (5.0 MB saved) |
#### 📈 Key Takeaways
- **🧠 Memory Efficiency**: LEANN uses **30% less memory** during index building and querying
- **💾 Storage Optimization**: LEANN requires **91% less storage** for the same dataset
- **🔄 On-demand Computing**: Storage savings come from computing embeddings at query time instead of pre-storing them
- **⚖️ Fair Comparison**: Both systems tested on identical hardware with the same 2,573 document dataset
> **Note**: Results may vary based on dataset size, hardware configuration, and query patterns. The comparison excludes text storage to focus purely on index structures.
### Run the comparison
```bash
python examples/compare_faiss_vs_leann.py
```
*Benchmark results obtained on Apple Silicon with consistent environmental conditions*
## 📊 Benchmarks

60
packages/leann-backend-hnsw/leann_backend_hnsw/hnsw_embedding_server.py
View File

@@ -150,6 +150,7 @@ def create_hnsw_embedding_server(
model_name: str = "sentence-transformers/all-mpnet-base-v2",
custom_max_length_param: Optional[int] = None,
distance_metric: str = "mips",
use_mlx: bool = False,
):
"""
Create and start a ZMQ-based embedding server for HNSW backend.
@@ -167,9 +168,13 @@ def create_hnsw_embedding_server(
custom_max_length_param: Custom max sequence length
distance_metric: The distance metric to use
"""
print(f"Loading tokenizer for {model_name}...")
tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=True)
print(f"Tokenizer loaded successfully!")
if not use_mlx:
print(f"Loading tokenizer for {model_name}...")
tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=True)
print(f"Tokenizer loaded successfully!")
else:
print("Using MLX mode - tokenizer will be loaded separately")
tokenizer = None
# Device setup
mps_available = hasattr(torch.backends, "mps") and torch.backends.mps.is_available()
@@ -191,8 +196,17 @@ def create_hnsw_embedding_server(
# Load model to the appropriate device
print(f"Starting HNSW server on port {zmq_port} with model {model_name}")
print(f"Loading model {model_name}... (this may take a while if downloading)")
model = AutoModel.from_pretrained(model_name).to(device).eval()
print(f"Model {model_name} loaded successfully!")
if use_mlx:
# For MLX models, we need to use the MLX embedding computation
print("MLX model detected - using MLX backend for embeddings")
model = None # We'll handle MLX separately
tokenizer = None
else:
# Use standard transformers for non-MLX models
model = AutoModel.from_pretrained(model_name).to(device).eval()
tokenizer = AutoTokenizer.from_pretrained(model_name)
print(f"Model {model_name} loaded successfully!")
# Check port availability
import socket
@@ -312,8 +326,37 @@ def create_hnsw_embedding_server(
def print_elapsed(self):
return # Disabled for now
def _process_batch_mlx(texts_batch, ids_batch, missing_ids):
"""Process a batch of texts using MLX backend"""
try:
# Import MLX embedding computation from main API
from leann.api import compute_embeddings
# Compute embeddings using MLX
embeddings = compute_embeddings(texts_batch, model_name, use_mlx=True)
print(
f"[leann_backend_hnsw.hnsw_embedding_server LOG]: MLX embeddings computed for {len(texts_batch)} texts"
)
print(
f"[leann_backend_hnsw.hnsw_embedding_server LOG]: Embedding shape: {embeddings.shape}"
)
return embeddings
except Exception as e:
print(
f"[leann_backend_hnsw.hnsw_embedding_server LOG]: ERROR in MLX processing: {e}"
)
raise
def process_batch(texts_batch, ids_batch, missing_ids):
"""Process a batch of texts and return embeddings"""
# Handle MLX models separately
if use_mlx:
return _process_batch_mlx(texts_batch, ids_batch, missing_ids)
_is_e5_model = "e5" in model_name.lower()
_is_bge_model = "bge" in model_name.lower()
batch_size = len(texts_batch)
@@ -927,6 +970,12 @@ if __name__ == "__main__":
parser.add_argument(
"--distance-metric", type=str, default="mips", help="Distance metric to use"
)
parser.add_argument(
"--use-mlx",
action="store_true",
default=False,
help="Use MLX for model inference",
)
args = parser.parse_args()
@@ -942,4 +991,5 @@ if __name__ == "__main__":
model_name=args.model_name,
custom_max_length_param=args.custom_max_length,
distance_metric=args.distance_metric,
use_mlx=args.use_mlx,
)

197
packages/leann-core/src/leann/api.py
View File

@@ -1,4 +1,3 @@
"""
This file contains the core API for the LEANN project, now definitively updated
with the correct, original embedding logic from the user's reference code.
@@ -18,7 +17,10 @@ from .interface import LeannBackendFactoryInterface
# --- The Correct, Verified Embedding Logic from old_code.py ---
def compute_embeddings(chunks: List[str], model_name: str, use_mlx: bool = False) -> np.ndarray:
def compute_embeddings(
chunks: List[str], model_name: str, use_mlx: bool = False
) -> np.ndarray:
"""Computes embeddings using sentence-transformers or MLX for consistent results."""
if use_mlx:
return compute_embeddings_mlx(chunks, model_name)
@@ -33,7 +35,9 @@ def compute_embeddings(chunks: List[str], model_name: str, use_mlx: bool = False
model = SentenceTransformer(model_name)
model = model.half()
print(f"INFO: Computing embeddings for {len(chunks)} chunks using SentenceTransformer model '{model_name}'...")
print(
f"INFO: Computing embeddings for {len(chunks)} chunks using SentenceTransformer model '{model_name}'..."
)
# use acclerater GPU or MAC GPU
if torch.cuda.is_available():
@@ -43,10 +47,13 @@ def compute_embeddings(chunks: List[str], model_name: str, use_mlx: bool = False
# Generate embeddings
# 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)
embeddings = model.encode(
chunks, convert_to_numpy=True, show_progress_bar=True, batch_size=256
)
return embeddings
def compute_embeddings_mlx(chunks: List[str], model_name: str) -> np.ndarray:
"""Computes embeddings using an MLX model."""
try:
@@ -54,10 +61,12 @@ def compute_embeddings_mlx(chunks: List[str], model_name: str) -> np.ndarray:
from mlx_lm.utils import load
except ImportError as e:
raise RuntimeError(
f"MLX or related libraries not available. Install with: pip install mlx mlx-lm"
f"MLX or related libraries not available. Install with: uv pip install mlx mlx-lm"
) from e
print(f"INFO: Computing embeddings for {len(chunks)} chunks using MLX model '{model_name}'...")
print(
f"INFO: Computing embeddings for {len(chunks)} chunks using MLX model '{model_name}'..."
)
# Load model and tokenizer
model, tokenizer = load(model_name)
@@ -67,27 +76,28 @@ def compute_embeddings_mlx(chunks: List[str], model_name: str) -> np.ndarray:
for chunk in chunks:
# Tokenize
token_ids = tokenizer.encode(chunk)
# Convert to MLX array and add batch dimension
input_ids = mx.array([token_ids])
# Get embeddings
embeddings = model(input_ids)
# Mean pooling (since we only have one sequence, just take the mean)
pooled = embeddings.mean(axis=1) # Shape: (1, hidden_size)
# Convert individual embedding to numpy via list (to handle bfloat16)
pooled_list = pooled[0].tolist() # Remove batch dimension and convert to list
pooled_numpy = np.array(pooled_list, dtype=np.float32)
all_embeddings.append(pooled_numpy)
# Stack numpy arrays
return np.stack(all_embeddings)
# --- Core API Classes (Restored and Unchanged) ---
@dataclass
class SearchResult:
id: str
@@ -95,23 +105,26 @@ class SearchResult:
text: str
metadata: Dict[str, Any] = field(default_factory=dict)
class PassageManager:
def __init__(self, passage_sources: List[Dict[str, Any]]):
self.offset_maps = {}
self.passage_files = {}
self.global_offset_map = {} # Combined map for fast lookup
for source in passage_sources:
if source["type"] == "jsonl":
passage_file = source["path"]
index_file = source["index_path"]
if not Path(index_file).exists():
raise FileNotFoundError(f"Passage index file not found: {index_file}")
with open(index_file, 'rb') as f:
raise FileNotFoundError(
f"Passage index file not found: {index_file}"
)
with open(index_file, "rb") as f:
offset_map = pickle.load(f)
self.offset_maps[passage_file] = offset_map
self.passage_files[passage_file] = passage_file
# Build global map for O(1) lookup
for passage_id, offset in offset_map.items():
self.global_offset_map[passage_id] = (passage_file, offset)
@@ -119,15 +132,25 @@ class PassageManager:
def get_passage(self, passage_id: str) -> Dict[str, Any]:
if passage_id in self.global_offset_map:
passage_file, offset = self.global_offset_map[passage_id]
with open(passage_file, 'r', encoding='utf-8') as f:
with open(passage_file, "r", encoding="utf-8") as f:
f.seek(offset)
return json.loads(f.readline())
raise KeyError(f"Passage ID not found: {passage_id}")
class LeannBuilder:
def __init__(self, backend_name: str, embedding_model: str = "facebook/contriever-msmarco", dimensions: Optional[int] = None, use_mlx: bool = False, **backend_kwargs):
def __init__(
self,
backend_name: str,
embedding_model: str = "facebook/contriever-msmarco",
dimensions: Optional[int] = None,
use_mlx: bool = False,
**backend_kwargs,
):
self.backend_name = backend_name
backend_factory: LeannBackendFactoryInterface | None = BACKEND_REGISTRY.get(backend_name)
backend_factory: LeannBackendFactoryInterface | None = BACKEND_REGISTRY.get(
backend_name
)
if backend_factory is None:
raise ValueError(f"Backend '{backend_name}' not found or not registered.")
self.backend_factory = backend_factory
@@ -138,14 +161,19 @@ class LeannBuilder:
self.chunks: List[Dict[str, Any]] = []
def add_text(self, text: str, metadata: Optional[Dict[str, Any]] = None):
if metadata is None: metadata = {}
passage_id = metadata.get('id', str(uuid.uuid4()))
if metadata is None:
metadata = {}
passage_id = metadata.get("id", str(uuid.uuid4()))
chunk_data = {"id": passage_id, "text": text, "metadata": metadata}
self.chunks.append(chunk_data)
def build_index(self, index_path: str):
if not self.chunks: raise ValueError("No chunks added.")
if self.dimensions is None: self.dimensions = len(compute_embeddings(["dummy"], self.embedding_model, self.use_mlx)[0])
if not self.chunks:
raise ValueError("No chunks added.")
if self.dimensions is None:
self.dimensions = len(
compute_embeddings(["dummy"], self.embedding_model, self.use_mlx)[0]
)
path = Path(index_path)
index_dir = path.parent
index_name = path.name
@@ -153,46 +181,76 @@ class LeannBuilder:
passages_file = index_dir / f"{index_name}.passages.jsonl"
offset_file = index_dir / f"{index_name}.passages.idx"
offset_map = {}
with open(passages_file, 'w', encoding='utf-8') as f:
with open(passages_file, "w", encoding="utf-8") as f:
for chunk in self.chunks:
offset = f.tell()
json.dump({"id": chunk["id"], "text": chunk["text"], "metadata": chunk["metadata"]}, f, ensure_ascii=False)
f.write('\n')
json.dump(
{
"id": chunk["id"],
"text": chunk["text"],
"metadata": chunk["metadata"],
},
f,
ensure_ascii=False,
)
f.write("\n")
offset_map[chunk["id"]] = offset
with open(offset_file, 'wb') as f: pickle.dump(offset_map, f)
with open(offset_file, "wb") as f:
pickle.dump(offset_map, f)
texts_to_embed = [c["text"] for c in self.chunks]
embeddings = compute_embeddings(texts_to_embed, self.embedding_model, self.use_mlx)
embeddings = compute_embeddings(
texts_to_embed, self.embedding_model, self.use_mlx
)
string_ids = [chunk["id"] for chunk in self.chunks]
current_backend_kwargs = {**self.backend_kwargs, 'dimensions': self.dimensions}
current_backend_kwargs = {**self.backend_kwargs, "dimensions": self.dimensions}
builder_instance = self.backend_factory.builder(**current_backend_kwargs)
builder_instance.build(embeddings, string_ids, index_path, **current_backend_kwargs)
builder_instance.build(
embeddings, string_ids, index_path, **current_backend_kwargs
)
leann_meta_path = index_dir / f"{index_name}.meta.json"
meta_data = {
"version": "1.0", "backend_name": self.backend_name, "embedding_model": self.embedding_model,
"dimensions": self.dimensions, "backend_kwargs": self.backend_kwargs, "use_mlx": self.use_mlx,
"passage_sources": [{"type": "jsonl", "path": str(passages_file), "index_path": str(offset_file)}]
"version": "1.0",
"backend_name": self.backend_name,
"embedding_model": self.embedding_model,
"dimensions": self.dimensions,
"backend_kwargs": self.backend_kwargs,
"use_mlx": self.use_mlx,
"passage_sources": [
{
"type": "jsonl",
"path": str(passages_file),
"index_path": str(offset_file),
}
],
}
# Add storage status flags for HNSW backend
if self.backend_name == "hnsw":
is_compact = self.backend_kwargs.get("is_compact", True)
is_recompute = self.backend_kwargs.get("is_recompute", True)
meta_data["is_compact"] = is_compact
meta_data["is_pruned"] = is_compact and is_recompute # Pruned only if compact and recompute
with open(leann_meta_path, 'w', encoding='utf-8') as f: json.dump(meta_data, f, indent=2)
meta_data["is_pruned"] = (
is_compact and is_recompute
) # Pruned only if compact and recompute
with open(leann_meta_path, "w", encoding="utf-8") as f:
json.dump(meta_data, f, indent=2)
class LeannSearcher:
def __init__(self, index_path: str, **backend_kwargs):
meta_path_str = f"{index_path}.meta.json"
if not Path(meta_path_str).exists(): raise FileNotFoundError(f"Leann metadata file not found at {meta_path_str}")
with open(meta_path_str, 'r', encoding='utf-8') as f: self.meta_data = json.load(f)
backend_name = self.meta_data['backend_name']
self.embedding_model = self.meta_data['embedding_model']
self.use_mlx = self.meta_data.get('use_mlx', False)
self.passage_manager = PassageManager(self.meta_data.get('passage_sources', []))
if not Path(meta_path_str).exists():
raise FileNotFoundError(f"Leann metadata file not found at {meta_path_str}")
with open(meta_path_str, "r", encoding="utf-8") as f:
self.meta_data = json.load(f)
backend_name = self.meta_data["backend_name"]
self.embedding_model = self.meta_data["embedding_model"]
self.use_mlx = self.meta_data.get("use_mlx", False)
self.passage_manager = PassageManager(self.meta_data.get("passage_sources", []))
backend_factory = BACKEND_REGISTRY.get(backend_name)
if backend_factory is None: raise ValueError(f"Backend '{backend_name}' not found.")
final_kwargs = {**self.meta_data.get('backend_kwargs', {}), **backend_kwargs}
if backend_factory is None:
raise ValueError(f"Backend '{backend_name}' not found.")
final_kwargs = {**self.meta_data.get("backend_kwargs", {}), **backend_kwargs}
self.backend_impl = backend_factory.searcher(index_path, **final_kwargs)
def search(self, query: str, top_k: int = 5, **search_kwargs) -> List[SearchResult]:
@@ -200,35 +258,56 @@ class LeannSearcher:
print(f" Query: '{query}'")
print(f" Top_k: {top_k}")
print(f" Search kwargs: {search_kwargs}")
query_embedding = compute_embeddings([query], self.embedding_model, self.use_mlx)
query_embedding = compute_embeddings(
[query], self.embedding_model, self.use_mlx
)
print(f" Generated embedding shape: {query_embedding.shape}")
print(f"🔍 DEBUG Query embedding first 10 values: {query_embedding[0][:10]}")
print(f"🔍 DEBUG Query embedding norm: {np.linalg.norm(query_embedding[0])}")
# Add use_mlx to search kwargs
search_kwargs["use_mlx"] = self.use_mlx
results = self.backend_impl.search(query_embedding, top_k, **search_kwargs)
print(f" Backend returned: labels={len(results.get('labels', [[]])[0])} results")
print(
f" Backend returned: labels={len(results.get('labels', [[]])[0])} results"
)
enriched_results = []
if 'labels' in results and 'distances' in results:
if "labels" in results and "distances" in results:
print(f" Processing {len(results['labels'][0])} passage IDs:")
for i, (string_id, dist) in enumerate(zip(results['labels'][0], results['distances'][0])):
for i, (string_id, dist) in enumerate(
zip(results["labels"][0], results["distances"][0])
):
try:
passage_data = self.passage_manager.get_passage(string_id)
enriched_results.append(SearchResult(
id=string_id, score=dist, text=passage_data['text'], metadata=passage_data.get('metadata', {})
))
print(f" {i+1}. passage_id='{string_id}' -> SUCCESS: {passage_data['text'][:60]}...")
except KeyError:
print(f" {i+1}. passage_id='{string_id}' -> ERROR: Passage not found in PassageManager!")
enriched_results.append(
SearchResult(
id=string_id,
score=dist,
text=passage_data["text"],
metadata=passage_data.get("metadata", {}),
)
)
print(
f" {i + 1}. passage_id='{string_id}' -> SUCCESS: {passage_data['text'][:60]}..."
)
except KeyError:
print(
f" {i + 1}. passage_id='{string_id}' -> ERROR: Passage not found in PassageManager!"
)
print(f" Final enriched results: {len(enriched_results)} passages")
return enriched_results
from .chat import get_llm
class LeannChat:
def __init__(self, index_path: str, llm_config: Optional[Dict[str, Any]] = None, **kwargs):
def __init__(
self, index_path: str, llm_config: Optional[Dict[str, Any]] = None, **kwargs
):
self.searcher = LeannSearcher(index_path, **kwargs)
self.llm = get_llm(llm_config)
@@ -248,7 +327,7 @@ class LeannChat:
while True:
try:
user_input = input("You: ").strip()
if user_input.lower() in ['quit', 'exit']:
if user_input.lower() in ["quit", "exit"]:
break
if not user_input:
continue

3
packages/leann-core/src/leann/embedding_server_manager.py
View File

@@ -310,9 +310,12 @@ class EmbeddingServerManager:
command.extend(["--passages-file", str(kwargs["passages_file"])])
# if "distance_metric" in kwargs and kwargs["distance_metric"]:
# command.extend(["--distance-metric", kwargs["distance_metric"]])
if "use_mlx" in kwargs and kwargs["use_mlx"]:
command.extend(["--use-mlx"])
project_root = Path(__file__).parent.parent.parent.parent.parent
print(f"INFO: Running command from project root: {project_root}")
print(f"INFO: Command: {' '.join(command)}") # Debug: show actual command
self.server_process = subprocess.Popen(
command,

3
packages/leann-core/src/leann/searcher_base.py
View File

@@ -78,9 +78,10 @@ class BaseSearcher(LeannBackendSearcherInterface, ABC):
model_name=self.embedding_model,
passages_file=passages_source_file,
distance_metric=kwargs.get("distance_metric"),
use_mlx=kwargs.get("use_mlx", False),
)
if not server_started:
raise RuntimeError(f"Failed to start embedding server on port {kwargs.get('zmq_port')}")
raise RuntimeError(f"Failed to start embedding server on port {port}")
@abstractmethod
def search(self, query: np.ndarray, top_k: int, **kwargs) -> Dict[str, Any]:

2
pyproject.toml
View File

@@ -34,6 +34,8 @@ dependencies = [
"msgpack>=1.1.1",
"llama-index-vector-stores-faiss>=0.4.0",
"llama-index-embeddings-huggingface>=0.5.5",
"mlx>=0.26.3",
"mlx-lm>=0.26.0",
]
[project.optional-dependencies]

18
test/build_mlx_index.py
View File

@@ -12,7 +12,7 @@ else:
builder = LeannBuilder(
backend_name="hnsw",
embedding_model="mlx-community/Qwen3-Embedding-0.6B-4bit-DWQ",
use_mlx=True
use_mlx=True,
)
# 2. Add documents
@@ -22,7 +22,7 @@ else:
"It was designed by Apple's machine learning research team.",
"The mlx-community organization provides pre-trained models in MLX format.",
"It supports operations on multi-dimensional arrays.",
"Leann can now use MLX for its embedding models."
"Leann can now use MLX for its embedding models.",
]
for doc in docs:
builder.add_text(doc)
@@ -34,9 +34,11 @@ else:
print(f"Check the metadata file: {INDEX_PATH}.meta.json")
chat = LeannChat(index_path=INDEX_PATH)
# add query
query = "MLX is an array framework for machine learning on Apple silicon."
print(f"Query: {query}")
response = chat.ask(query, top_k=3, recompute_beighbor_embeddings=True, complexity=3, beam_width=1)
print(f"Response: {response}")
chat = LeannChat(index_path=INDEX_PATH)
# add query
query = "MLX is an array framework for machine learning on Apple silicon."
print(f"Query: {query}")
response = chat.ask(
query, top_k=3, recompute_beighbor_embeddings=True, complexity=3, beam_width=1
)
print(f"Response: {response}")

882
uv.lock generated
View File

File diff suppressed because it is too large Load Diff