1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
|
"""
Index all of the FAISS datasets
"""
import os
import glob
import faiss
import time
import numpy as np
from app.utils.file_utils import load_recipe, load_csv_safe
from app.settings import app_cfg as cfg
class DefaultRecipe:
def __init__(self):
self.dim = 128
self.factory_type = 'Flat'
def build_all_faiss_databases():
datasets = []
for fn in glob.iglob(os.path.join(cfg.DIR_FAISS_METADATA, "*")):
name = os.path.basename(fn)
recipe_fn = os.path.join(cfg.DIR_FAISS_RECIPES, name + ".json")
if os.path.exists(recipe_fn):
build_faiss_database(name, load_recipe(recipe_fn))
else:
build_faiss_database(name, DefaultRecipe())
def build_faiss_database(name, recipe):
vec_fn = os.path.join(cfg.DIR_FAISS_METADATA, name, "vecs.csv")
index_fn = os.path.join(cfg.DIR_FAISS_INDEXES, name + ".index")
index = faiss.index_factory(recipe.dim, recipe.factory_type)
keys, rows = load_csv_safe(vec_fn)
feats = np.array([ list(map(float, row[3].split(","))) for row in rows ]).astype('float32')
n, d = feats.shape
print("{}: training {} x {} dim vectors".format(name, n, d))
print(recipe.factory_type)
add_start = time.time()
index.add(feats)
add_end = time.time()
add_time = add_end - add_start
print("{}: add time: {:.1f}s".format(name, add_time))
faiss.write_index(index, index_fn)
def load_faiss_databases():
faiss_datasets = {}
for fn in glob.iglob(os.path.join(cfg.DIR_FAISS_METADATA, "*")):
name = os.path.basename(fn)
index_fn = os.path.join(cfg.DIR_FAISS_INDEXES, name + ".index")
if os.path.exists(index_fn):
index = faiss.read_index(index_fn)
faiss_datasets[name] = index
return faiss_datasets
|
