Spaces:
Runtime error
Runtime error
update app to use indexes directly to get embeddings
Browse files- app.py +32 -31
- big_id_to_image_emb_dict.pickle +0 -3
- big_indx_to_id_dict.pickle +2 -2
app.py
CHANGED
@@ -11,11 +11,11 @@ import click
|
|
11 |
|
12 |
|
13 |
def getRandID():
|
14 |
-
indx = random.randrange(0,
|
15 |
-
return
|
16 |
|
17 |
|
18 |
-
def
|
19 |
if indexType == "FlatIP(default)":
|
20 |
return image_index_IP
|
21 |
elif indexType == "FlatL2":
|
@@ -32,7 +32,7 @@ def chooseImageIndex(indexType):
|
|
32 |
return image_index_LSH
|
33 |
|
34 |
|
35 |
-
def
|
36 |
if indexType == "FlatIP(default)":
|
37 |
return dna_index_IP
|
38 |
elif indexType == "FlatL2":
|
@@ -49,35 +49,36 @@ def chooseDNAIndex(indexType):
|
|
49 |
return dna_index_LSH
|
50 |
|
51 |
|
52 |
-
def searchEmbeddings(id, key_type, query_type, index_type):
|
53 |
-
|
54 |
-
|
55 |
-
count = 0
|
56 |
-
num_neighbors = 10
|
57 |
-
|
58 |
-
index = faiss.IndexFlatIP(dim)
|
59 |
|
60 |
# get index
|
61 |
if query_type == "Image":
|
62 |
-
|
63 |
elif query_type == "DNA":
|
64 |
-
|
|
|
|
|
|
|
|
|
65 |
|
66 |
# search for query
|
67 |
if key_type == "Image":
|
68 |
-
|
69 |
elif key_type == "DNA":
|
70 |
-
|
71 |
-
|
72 |
-
|
73 |
|
74 |
-
|
75 |
-
|
|
|
76 |
for indx in I[0]:
|
77 |
-
id =
|
78 |
-
|
79 |
|
80 |
-
return
|
81 |
|
82 |
|
83 |
with gr.Blocks() as demo:
|
@@ -102,14 +103,8 @@ with gr.Blocks() as demo:
|
|
102 |
# with open("processid_to_index.pickle", "rb") as f:
|
103 |
# processid_to_index = pickle.load(f)
|
104 |
with open("big_indx_to_id_dict.pickle", "rb") as f:
|
105 |
-
|
106 |
-
|
107 |
-
# initialize both possible dicts
|
108 |
-
with open("big_id_to_image_emb_dict.pickle", "rb") as f:
|
109 |
-
id_to_image_emb_dict = pickle.load(f)
|
110 |
-
# with open("big_id_to_dna_emb_dict.pickle", "rb") as f:
|
111 |
-
# id_to_dna_emb_dict = pickle.load(f)
|
112 |
-
id_to_dna_emb_dict = None
|
113 |
|
114 |
with gr.Column():
|
115 |
with gr.Row():
|
@@ -124,12 +119,18 @@ with gr.Blocks() as demo:
|
|
124 |
index_type = gr.Radio(
|
125 |
choices=["FlatIP(default)", "FlatL2", "HNSWFlat", "IVFFlat", "LSH"], label="Index:", value="FlatIP(default)"
|
126 |
)
|
|
|
|
|
127 |
process_id = gr.Textbox(label="ID:", info="Enter a sample ID to search for")
|
128 |
process_id_list = gr.Textbox(label="Closest 10 matches:")
|
129 |
search_btn = gr.Button("Search")
|
130 |
id_btn.click(fn=getRandID, inputs=[], outputs=[rand_id, rand_id_indx])
|
131 |
|
132 |
-
search_btn.click(
|
|
|
|
|
|
|
|
|
133 |
|
134 |
|
135 |
demo.launch()
|
|
|
11 |
|
12 |
|
13 |
def getRandID():
|
14 |
+
indx = random.randrange(0, len(index_to_id_dict))
|
15 |
+
return index_to_id_dict[indx], indx
|
16 |
|
17 |
|
18 |
+
def get_image_index(indexType):
|
19 |
if indexType == "FlatIP(default)":
|
20 |
return image_index_IP
|
21 |
elif indexType == "FlatL2":
|
|
|
32 |
return image_index_LSH
|
33 |
|
34 |
|
35 |
+
def get_dna_index(indexType):
|
36 |
if indexType == "FlatIP(default)":
|
37 |
return dna_index_IP
|
38 |
elif indexType == "FlatL2":
|
|
|
49 |
return dna_index_LSH
|
50 |
|
51 |
|
52 |
+
def searchEmbeddings(id, key_type, query_type, index_type, num_results: int = 10):
|
53 |
+
image_index = get_image_index(index_type)
|
54 |
+
dna_index = get_dna_index(index_type)
|
|
|
|
|
|
|
|
|
55 |
|
56 |
# get index
|
57 |
if query_type == "Image":
|
58 |
+
query = image_index.reconstruct(id_to_index_dict[id])
|
59 |
elif query_type == "DNA":
|
60 |
+
query = dna_index.reconstruct(id_to_index_dict[id])
|
61 |
+
else:
|
62 |
+
raise ValueError(f"Invalid query type: {query_type}")
|
63 |
+
query = query.astype(np.float32)
|
64 |
+
query = np.expand_dims(query, axis=0)
|
65 |
|
66 |
# search for query
|
67 |
if key_type == "Image":
|
68 |
+
index = image_index
|
69 |
elif key_type == "DNA":
|
70 |
+
index = dna_index
|
71 |
+
else:
|
72 |
+
raise ValueError(f"Invalid key type: {key_type}")
|
73 |
|
74 |
+
_, I = index.search(query, num_results)
|
75 |
+
|
76 |
+
closest_ids = []
|
77 |
for indx in I[0]:
|
78 |
+
id = index_to_id_dict[indx]
|
79 |
+
closest_ids.append(id)
|
80 |
|
81 |
+
return closest_ids
|
82 |
|
83 |
|
84 |
with gr.Blocks() as demo:
|
|
|
103 |
# with open("processid_to_index.pickle", "rb") as f:
|
104 |
# processid_to_index = pickle.load(f)
|
105 |
with open("big_indx_to_id_dict.pickle", "rb") as f:
|
106 |
+
index_to_id_dict = pickle.load(f)
|
107 |
+
id_to_index_dict = {v: k for k, v in index_to_id_dict.items()}
|
|
|
|
|
|
|
|
|
|
|
|
|
108 |
|
109 |
with gr.Column():
|
110 |
with gr.Row():
|
|
|
119 |
index_type = gr.Radio(
|
120 |
choices=["FlatIP(default)", "FlatL2", "HNSWFlat", "IVFFlat", "LSH"], label="Index:", value="FlatIP(default)"
|
121 |
)
|
122 |
+
num_results = gr.Number(label="Number of Results:", value=10, precision=0)
|
123 |
+
|
124 |
process_id = gr.Textbox(label="ID:", info="Enter a sample ID to search for")
|
125 |
process_id_list = gr.Textbox(label="Closest 10 matches:")
|
126 |
search_btn = gr.Button("Search")
|
127 |
id_btn.click(fn=getRandID, inputs=[], outputs=[rand_id, rand_id_indx])
|
128 |
|
129 |
+
search_btn.click(
|
130 |
+
fn=searchEmbeddings,
|
131 |
+
inputs=[process_id, key_type, query_type, index_type, num_results],
|
132 |
+
outputs=[process_id_list],
|
133 |
+
)
|
134 |
|
135 |
|
136 |
demo.launch()
|
big_id_to_image_emb_dict.pickle
DELETED
@@ -1,3 +0,0 @@
|
|
1 |
-
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:4fb3f21f2d38a91cb2cad8f40449f31c12d481944d93e9c61def2d3e8e6b78eb
|
3 |
-
size 274402415
|
|
|
|
|
|
|
|
big_indx_to_id_dict.pickle
CHANGED
@@ -1,3 +1,3 @@
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
-
oid sha256:
|
3 |
-
size
|
|
|
1 |
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:ee0a9044e054f640b704247a2fa2e74219180b78ded6ba07f551bfc222657fc5
|
3 |
+
size 885457
|