index_name
@property
index_name: str
Returns
str: The unique name identifier of the index.
Example Usage
index_name = index.index_name
print(f'Index name: {index_name}')
# Output: "my-vectors"
index_type
@property
index_type: str
'ivfflat', 'ivfpq', or 'ivfsq').
"ivf" is a deprecated index type that may be returned by older indexes. New indexes should use 'ivfflat', 'ivfpq', or 'ivfsq'.
Returns
str: The index type.
Example Usage
index_type = index.index_type
print(f'Index type: {index_type}')
# Output: "ivfflat", "ivfpq", or "ivfsq"
if index_type == 'ivfsq':
print('Using IVFSQ index for high accuracy and memory efficiency')
elif index_type == 'ivfpq':
print('Using IVFPQ index for memory efficiency')
elif index_type == 'ivfflat':
print('Using IVFFlat index for high accuracy')
else:
print('Unknown index type')
is_trained
def is_trained(self) -> bool
Returns
bool: True if the index has been trained, False otherwise.
Example Usage
trained = index.is_trained()
print(f'Index trained: {trained}')
if not trained:
print('Index needs training for optimal performance')
index.train()
is_training
def is_training(self) -> bool
Returns
bool: True if the index is currently training, False otherwise.
Example Usage
import time
index.train() # Start training
while index.is_training():
print('Index is still training...')
time.sleep(5) # Wait before checking again
print('Index training complete')
index_config
@property
index_config: Dict[str, Any]
Returns
Dict[str, Any]: Configuration dictionary containing index parameters.
Example Usage
config = index.index_config
print(f'Index configuration: {config}')
# Access common properties
print(f"Dimension: {config.get('dimension')}")
print(f"Metric: {config.get('metric')}")
print(f"Number of lists: {config.get('n_lists')}")
# Type-specific properties
if config.get('type') == 'ivfpq':
print(f"PQ dimension: {config.get('pq_dim')}")
print(f"PQ bits: {config.get('pq_bits')}")
Configuration Properties
Common Properties (All Index Types)
| Property | Type | Description |
|---|
type | str | Index type ('ivfflat', 'ivfpq', or 'ivfsq') |
dimension | int | Vector dimensionality |
metric | str | Distance metric ('cosine', 'euclidean', etc.) |
n_lists | int | Number of inverted lists for clustering |
IVFPQ-Specific Properties
| Property | Type | Description |
|---|
pq_dim | int | Product quantization dimension |
pq_bits | int | Number of bits for quantization |
