-
Notifications
You must be signed in to change notification settings - Fork 1
API Basics
mattmccj edited this page Jan 1, 2019
·
2 revisions
The basic commands are:
-
open(host, port, api_key)
- Open a socket for communication with the Nearist appliance.
host |
IP address of the Nearist appliance. |
port |
Port number for accessing the Nearist appliance. |
api_key |
Unique user access key which is required to access the appliance. |
-
close()
- Close the socket to the Nearist appliance. -
set_distance_mode(mode)
- Set the distance metric.
mode |
Distance metric from Common.DistanceMode . |
-
set_query_mode(mode)
- Set query mode
mode |
Query mode from Common.QueryMode . |
-
set_read_count(count)
- Set query result count forKNN_D
/KNN_A
query mode(s)
count |
The top 'K' values in KNN. |
-
set_threshold(threshold)
- Set query threshold forGT
,LT
, orKNN
query modes. InGT
orKNN_D
mode, the appliance will only return results whose distance is greater than the threshold value. InLT
orKNN_A
mode, the appliance will only return results whose distance is less than the threshold value.
threshold |
The threshold value. |
-
ds_load(vectors)
- Load dataset to Nearist appliance
vectors |
List of vectors (component lists). |
-
load_dataset_file(file_name, dataset_name)
- Load local dataset to Nearist appliance
file_name |
Local dataset file name. |
dataset_name |
Local dataset name. |
-
ds_load_random(vector_count, comp_count)
- Load random dataset to Nearist appliance
vector_count |
Vector count to generate. |
comp_count |
Vector's component count. |
-
query(vectors)
- Query for single/multiple vector(s)
vectors |
List of components for single query / List of vectors (component lists) for multipel query. |
-
query_from_file(file_name, dataset_name, output_name)
- Query local dataset to Nearist appliance
file_name |
Local dataset file name. |
dataset_name |
Local dataset name. |
output_name |
Results output file name. |
Additional documentation can be found in the function header comments in Client.py
.
Brute Force Benchmarks
ANN Benchmarks