Couchbase

1import os
2from kern.agent import Agent
3from kern.knowledge.embedder.openai import OpenAIEmbedder
4from kern.knowledge.knowledge import Knowledge
5from kern.vectordb.couchbase import CouchbaseSearch
6from couchbase.auth import PasswordAuthenticator
7from couchbase.management.search import SearchIndex
8from couchbase.options import ClusterOptions, KnownConfigProfiles
9
10# Couchbase connection settings
11username = os.getenv("COUCHBASE_USER")
12password = os.getenv("COUCHBASE_PASSWORD")
13connection_string = os.getenv("COUCHBASE_CONNECTION_STRING")
14
15# Create cluster options with authentication
16auth = PasswordAuthenticator(username, password)
17cluster_options = ClusterOptions(auth)
18cluster_options.apply_profile(KnownConfigProfiles.WanDevelopment)
19
20# Define the vector search index
21search_index = SearchIndex(
22    name="vector_search",
23    source_type="gocbcore",
24    idx_type="fulltext-index",
25    source_name="recipe_bucket",
26    plan_params={"index_partitions": 1, "num_replicas": 0},
27    params={
28        "doc_config": {
29            "docid_prefix_delim": "",
30            "docid_regexp": "",
31            "mode": "scope.collection.type_field",
32            "type_field": "type",
33        },
34        "mapping": {
35            "default_analyzer": "standard",
36            "default_datetime_parser": "dateTimeOptional",
37            "index_dynamic": True,
38            "store_dynamic": True,
39            "default_mapping": {"dynamic": True, "enabled": False},
40            "types": {
41                "recipe_scope.recipes": {
42                    "dynamic": False,
43                    "enabled": True,
44                    "properties": {
45                        "content": {
46                            "enabled": True,
47                            "fields": [
48                                {
49                                    "docvalues": True,
50                                    "include_in_all": False,
51                                    "include_term_vectors": False,
52                                    "index": True,
53                                    "name": "content",
54                                    "store": True,
55                                    "type": "text",
56                                }
57                            ],
58                        },
59                        "embedding": {
60                            "enabled": True,
61                            "dynamic": False,
62                            "fields": [
63                                {
64                                    "vector_index_optimized_for": "recall",
65                                    "docvalues": True,
66                                    "dims": 1536,
67                                    "include_in_all": False,
68                                    "include_term_vectors": False,
69                                    "index": True,
70                                    "name": "embedding",
71                                    "similarity": "dot_product",
72                                    "store": True,
73                                    "type": "vector",
74                                }
75                            ],
76                        },
77                        "meta": {
78                            "dynamic": True,
79                            "enabled": True,
80                            "properties": {
81                                "name": {
82                                    "enabled": True,
83                                    "fields": [
84                                        {
85                                            "docvalues": True,
86                                            "include_in_all": False,
87                                            "include_term_vectors": False,
88                                            "index": True,
89                                            "name": "name",
90                                            "store": True,
91                                            "analyzer": "keyword",
92                                            "type": "text",
93                                        }
94                                    ],
95                                }
96                            },
97                        },
98                    },
99                }
100            },
101        },
102    },
103)
104vector_db = CouchbaseSearch(
105    bucket_name="recipe_bucket",
106    scope_name="recipe_scope",
107    collection_name="recipes",
108    couchbase_connection_string=connection_string,
109    cluster_options=cluster_options,
110    search_index=search_index,
111    embedder=OpenAIEmbedder(
112        dimensions=1536,
113    ),
114    wait_until_index_ready=60,
115    overwrite=True,
116)
117
118knowledge = Knowledge(
119    name="Couchbase Knowledge Base",
120    description="This is a knowledge base that uses a Couchbase DB",
121    vector_db=vector_db,
122)
123
124knowledge.insert(
125    name="Recipes",
126    url="https://kern-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf",
127    metadata={"doc_type": "recipe_book"},
128)
129agent = Agent(
130    knowledge=knowledge,
131    search_knowledge=True,
132    read_chat_history=True,
133)
134
135agent.print_response("List down the ingredients to make Massaman Gai", markdown=True)
136
137vector_db.delete_by_name("Recipes")
138# or
139vector_db.delete_by_metadata({"doc_type": "recipe_book"})