Update the version to 0.1.29 (#406)

Signed-off-by: SimFG <bang.fu@zilliz.com>
zilliztech · Jun 2, 2023 · fd7e303 · fd7e303
1 parent c274d83
commit fd7e303
Show file tree

Hide file tree

Showing 5 changed files with 44 additions and 16 deletions.
diff --git a/docs/release_note.md b/docs/release_note.md
@@ -5,6 +5,20 @@ To read the following content, you need to understand the basic use of GPTCache,
 - [Readme doc](https://github.com/zilliztech/GPTCache)
 - [Usage doc](https://github.com/zilliztech/GPTCache/blob/main/docs/usage.md)
 
+## v0.1.29 (2023.6.2)
+
+1. Improve the GPTCache server by using FASTAPI
+
+**NOTE**: The api struct has been optimized, details: [Use GPTCache server](https://github.com/zilliztech/GPTCache/blob/dev/docs/usage.md#use-gptcache-server)
+
+2. Add the usearch vector store
+
+```python
+from gptcache.manager import manager_factory
+
+data_manager = manager_factory("sqlite,usearch", vector_params={"dimension": 10})
+```
+
 ## v0.1.28 (2023.5.29)
 To handle a large prompt, there are currently two options available:
 

diff --git a/gptcache/__init__.py b/gptcache/__init__.py
@@ -1,5 +1,5 @@
 """gptcache version"""
-__version__ = "0.1.28"
+__version__ = "0.1.29"
 
 from gptcache.config import Config
 from gptcache.core import Cache

diff --git a/gptcache/manager/vector_data/manager.py b/gptcache/manager/vector_data/manager.py
@@ -16,11 +16,8 @@
     "params": {"M": 8, "efConstruction": 64},
 }
 
-PGVECTOR_URL="postgresql://postgres:postgres@localhost:5432/postgres"
-PGVECTOR_INDEX_PARAMS = {
-    "index_type": "L2",
-    "params": {"lists": 100, "probes": 10}
-}
+PGVECTOR_URL = "postgresql://postgres:postgres@localhost:5432/postgres"
+PGVECTOR_INDEX_PARAMS = {"index_type": "L2", "params": {"lists": 100, "probes": 10}}
 
 COLLECTION_NAME = "gptcache"
 
@@ -169,6 +166,7 @@ def get(name, **kwargs):
             )
         elif name == "pgvector":
             from gptcache.manager.vector_data.pgvector import PGVector
+
             dimension = kwargs.get("dimension", DIMENSION)
             url = kwargs.get("url", PGVECTOR_URL)
             collection_name = kwargs.get("collection_name", COLLECTION_NAME)
@@ -178,13 +176,27 @@ def get(name, **kwargs):
                 top_k=top_k,
                 url=url,
                 collection_name=collection_name,
-                index_params=index_params
+                index_params=index_params,
             )
         elif name == "docarray":
             from gptcache.manager.vector_data.docarray_index import DocArrayIndex
 
             index_path = kwargs.pop("index_path", "./docarray_index.bin")
             vector_base = DocArrayIndex(index_file_path=index_path, top_k=top_k)
+        elif name == "usearch":
+            from gptcache.manager.vector_data.usearch import USearch
+
+            dimension = kwargs.get("dimension", DIMENSION)
+            index_path = kwargs.pop("index_path", "./index.usearch")
+            metric = kwargs.get("metric", "cos")
+            dtype = kwargs.get("dtype", "f32")
+            vector_base = USearch(
+                index_file_path=index_path,
+                dimension=dimension,
+                top_k=top_k,
+                metric=metric,
+                dtype=dtype,
+            )
         else:
             raise NotFoundError("vector store", name)
         return vector_base
diff --git a/gptcache/manager/vector_data/usearch.py b/gptcache/manager/vector_data/usearch.py
@@ -7,6 +7,7 @@
 from gptcache.utils import import_usearch
 
 import_usearch()
+
 from usearch.index import Index  # pylint: disable=C0413
 
 
@@ -33,11 +34,11 @@ class USearch(VectorBase):
 
     def __init__(
         self,
-        index_file_path: str = 'index.usearch',
+        index_file_path: str = "index.usearch",
         dimension: int = 64,
         top_k: int = 1,
-        metric: str = 'cos',
-        dtype: str = 'f32',
+        metric: str = "cos",
+        dtype: str = "f32",
         connectivity: int = 16,
         expansion_add: int = 128,
         expansion_search: int = 64,
@@ -57,16 +58,15 @@ def __init__(
             self._index.load(self._index_file_path)
 
     def mul_add(self, datas: List[VectorData]):
-        data_array, id_array = map(
-            list, zip(*((data.data, data.id) for data in datas)))
-        np_data = np.array(data_array).astype('float32')
+        data_array, id_array = map(list, zip(*((data.data, data.id) for data in datas)))
+        np_data = np.array(data_array).astype("float32")
         ids = np.array(id_array, dtype=np.longlong)
         self._index.add(ids, np_data)
 
     def search(self, data: np.ndarray, top_k: int = -1):
         if top_k == -1:
             top_k = self._top_k
-        np_data = np.array(data).astype('float32').reshape(1, -1)
+        np_data = np.array(data).astype("float32").reshape(1, -1)
         ids, dist, _ = self._index.search(np_data, top_k)
         return list(zip(dist[0], ids[0]))
 

diff --git a/tests/unit_tests/manager/test_usearch.py b/tests/unit_tests/manager/test_usearch.py
@@ -1,7 +1,8 @@
 import unittest
+
 import numpy as np
 
-from gptcache.manager.vector_data.usearch import USearch
+from gptcache.manager.vector_data import VectorBase
 from gptcache.manager.vector_data.base import VectorData
 
 
@@ -11,7 +12,8 @@ def test_normal(self):
         dim = 512
         top_k = 10
 
-        db = USearch(
+        db = VectorBase(
+            "usearch",
             index_file_path='./index.usearch',
             dimension=dim,
             top_k=top_k,