qdrant

Name: qdrant
Author: giuseppe-trisciuoglio

1.6kinstalls36forks

How to install qdrant?

giuseppe-trisciuoglio/developer-kit/qdrant

$npx -y skills add giuseppe-trisciuoglio/developer-kit --skill qdrant

Installs into the current project.

Summary

Provides Qdrant vector database integration patterns with LangChain4j. Handles embedding storage, similarity search, and vector management for Java applications. Use when implementing vector-based retrieval for RAG systems, semantic search, or recommendation engines.

Files · 1

View on GitHub

SKILL.md

1	# Qdrant Vector Database Integration
2
3	## Overview
4
5	Qdrant is an AI-native vector database for semantic search and similarity retrieval. This skill provides patterns for integrating Qdrant with Java applications, focusing on Spring Boot and LangChain4j integration.
6
7	## When to Use
8
9	- Semantic search or recommendation systems in Spring Boot applications
10	- RAG pipelines with Java and LangChain4j
11	- Vector database integration for AI/ML applications
12	- High-performance similarity search with filtered queries
13
14	## Instructions
15
16	### 1. Deploy Qdrant with Docker
17
18	```bash
19	docker run -p 6333:6333 -p 6334:6334 \
20	-v "$(pwd)/qdrant_storage:/qdrant/storage:z" \
21	qdrant/qdrant
22	```
23
24	Access: REST API at `http://localhost:6333`, gRPC at `http://localhost:6334`.
25
26	### 2. Add Dependencies
27
28	Maven:
29	```xml
30	<dependency>
31	<groupId>io.qdrant</groupId>
32	<artifactId>client</artifactId>
33	<version>1.15.0</version>
34	</dependency>
35	```
36
37	Gradle:
38	```gradle
39	implementation 'io.qdrant:client:1.15.0'
40	```
41
42	### 3. Initialize Client
43
44	```java
45	QdrantClient client = new QdrantClient(
46	QdrantGrpcClient.newBuilder("localhost").build());
47	```
48
49	For production with API key:
50	```java
51	QdrantClient client = new QdrantClient(
52	QdrantGrpcClient.newBuilder("localhost", 6334, false)
53	.withApiKey("YOUR_API_KEY")
54	.build());
55	```
56
57	### 4. Create Collection
58
59	```java
60	client.createCollectionAsync("search-collection",
61	VectorParams.newBuilder()
62	.setDistance(Distance.Cosine)
63	.setSize(384)
64	.build()
65	).get();
66	```
67
68	Validation: Verify the collection was created by checking `client.getCollectionAsync("search-collection").get()`.
69
70	### 5. Upsert Vectors
71
72	```java
73	List<PointStruct> points = List.of(
74	PointStruct.newBuilder()
75	.setId(id(1))
76	.setVectors(vectors(0.05f, 0.61f, 0.76f, 0.74f))
77	.putAllPayload(Map.of("title", value("Spring Boot Documentation")))
78	.build()
79	);
80	client.upsertAsync("search-collection", points).get();
81	```
82
83	Validation: Check that `client.upsertAsync(...).get()` completes without throwing.
84
85	### 6. Search Vectors
86
87	```java
88	List<ScoredPoint> results = client.queryAsync(
89	QueryPoints.newBuilder()
90	.setCollectionName("search-collection")
91	.setLimit(5)
92	.setQuery(nearest(0.2f, 0.1f, 0.9f, 0.7f))
93	.build()
94	).get();
95	```
96
97	Filtered search:
98	```java
99	List<ScoredPoint> results = client.searchAsync(
100	SearchPoints.newBuilder()
101	.setCollectionName("search-collection")
102	.addAllVector(List.of(0.62f, 0.12f, 0.53f, 0.12f))
103	.setFilter(Filter.newBuilder()
104	.addMust(range("category", Range.newBuilder().setEq("docs").build()))
105	.build())
106	.setLimit(5)
107	.build()).get();
108	```
109
110	## LangChain4j Integration
111
112	For RAG pipelines, use LangChain4j's high-level abstractions:
113
114	```java
115	EmbeddingStore<TextSegment> embeddingStore = QdrantEmbeddingStore.builder()
116	.collectionName("rag-collection")
117	.host("localhost")
118	.port(6334)
119	.apiKey("YOUR_API_KEY")
120	.build();
121	```
122
123	Spring Boot configuration with LangChain4j:
124	```java
125	@Bean
126	public EmbeddingStore<TextSegment> embeddingStore() {
127	return QdrantEmbeddingStore.builder()
128	.collectionName("rag-collection")
129	.host(host)
130	.port(port)
131	.build();
132	}
133
134	@Bean
135	public EmbeddingModel embeddingModel() {
136	return new AllMiniLmL6V2EmbeddingModel();
137	}
138	```
139
140	## Spring Boot Integration
141
142	Inject the client via configuration:
143
144	```java
145	@Configuration
146	public class QdrantConfig {
147	@Value("${qdrant.host:localhost}")
148	private String host;
149
150	@Value("${qdrant.port:6334}")
151	private int port;
152
153	@Bean
154	public QdrantClient qdrantClient() {
155	return new QdrantClient(
156	QdrantGrpcClient.newBuilder(host, port, false).build());
157	}
158	}
159	```
160
161	## Examples
162
163	### REST Search Endpoint
164
165	```java
166	@RestController
167	@RequestMapping("/api/search")
168	public class SearchController {
169	private final VectorSearchService searchService;
170
171	public SearchController(VectorSearchService searchService) {
172	this.searchService = searchService;
173	}
174
175	@GetMapping
176	public List<ScoredPoint> search(@RequestParam String query) {
177	List<Float> queryVector = embeddingModel.embed(query).content().vectorAsList();
178	return searchService.search("documents", queryVector);
179	}
180	}
181	```
182
183	## Best Practices
184
185	- Distance metric: Cosine for normalized text embeddings, Euclidean for non-normalized.
186	- Batch upserts: Use batch operations over individual point insertions.
187	- Connection pooling: Configure connection pooling for high-throughput production workloads.
188	- Error handling: W

People who liked this also viewed

6 picks

Type

qdrant

1.6kinstalls36forks

View on GitHub

How to install qdrant?

giuseppe-trisciuoglio/developer-kit/qdrant

$npx -y skills add giuseppe-trisciuoglio/developer-kit --skill qdrant

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Qdrant Vector Database Integration
2
3	## Overview
4
5	Qdrant is an AI-native vector database for semantic search and similarity retrieval. This skill provides patterns for integrating Qdrant with Java applications, focusing on Spring Boot and LangChain4j integration.
6
7	## When to Use
8
9	- Semantic search or recommendation systems in Spring Boot applications
10	- RAG pipelines with Java and LangChain4j
11	- Vector database integration for AI/ML applications
12	- High-performance similarity search with filtered queries
13
14	## Instructions
15
16	### 1. Deploy Qdrant with Docker
17
18	```bash
19	docker run -p 6333:6333 -p 6334:6334 \
20	-v "$(pwd)/qdrant_storage:/qdrant/storage:z" \
21	qdrant/qdrant
22	```
23
24	Access: REST API at `http://localhost:6333`, gRPC at `http://localhost:6334`.
25
26	### 2. Add Dependencies
27
28	Maven:
29	```xml
30	<dependency>
31	<groupId>io.qdrant</groupId>
32	<artifactId>client</artifactId>
33	<version>1.15.0</version>
34	</dependency>
35	```
36
37	Gradle:
38	```gradle
39	implementation 'io.qdrant:client:1.15.0'
40	```
41
42	### 3. Initialize Client
43
44	```java
45	QdrantClient client = new QdrantClient(
46	QdrantGrpcClient.newBuilder("localhost").build());
47	```
48
49	For production with API key:
50	```java
51	QdrantClient client = new QdrantClient(
52	QdrantGrpcClient.newBuilder("localhost", 6334, false)
53	.withApiKey("YOUR_API_KEY")
54	.build());
55	```
56
57	### 4. Create Collection
58
59	```java
60	client.createCollectionAsync("search-collection",
61	VectorParams.newBuilder()
62	.setDistance(Distance.Cosine)
63	.setSize(384)
64	.build()
65	).get();
66	```
67
68	Validation: Verify the collection was created by checking `client.getCollectionAsync("search-collection").get()`.
69
70	### 5. Upsert Vectors
71
72	```java
73	List<PointStruct> points = List.of(
74	PointStruct.newBuilder()
75	.setId(id(1))
76	.setVectors(vectors(0.05f, 0.61f, 0.76f, 0.74f))
77	.putAllPayload(Map.of("title", value("Spring Boot Documentation")))
78	.build()
79	);
80	client.upsertAsync("search-collection", points).get();
81	```
82
83	Validation: Check that `client.upsertAsync(...).get()` completes without throwing.
84
85	### 6. Search Vectors
86
87	```java
88	List<ScoredPoint> results = client.queryAsync(
89	QueryPoints.newBuilder()
90	.setCollectionName("search-collection")
91	.setLimit(5)
92	.setQuery(nearest(0.2f, 0.1f, 0.9f, 0.7f))
93	.build()
94	).get();
95	```
96
97	Filtered search:
98	```java
99	List<ScoredPoint> results = client.searchAsync(
100	SearchPoints.newBuilder()
101	.setCollectionName("search-collection")
102	.addAllVector(List.of(0.62f, 0.12f, 0.53f, 0.12f))
103	.setFilter(Filter.newBuilder()
104	.addMust(range("category", Range.newBuilder().setEq("docs").build()))
105	.build())
106	.setLimit(5)
107	.build()).get();
108	```
109
110	## LangChain4j Integration
111
112	For RAG pipelines, use LangChain4j's high-level abstractions:
113
114	```java
115	EmbeddingStore<TextSegment> embeddingStore = QdrantEmbeddingStore.builder()
116	.collectionName("rag-collection")
117	.host("localhost")
118	.port(6334)
119	.apiKey("YOUR_API_KEY")
120	.build();
121	```
122
123	Spring Boot configuration with LangChain4j:
124	```java
125	@Bean
126	public EmbeddingStore<TextSegment> embeddingStore() {
127	return QdrantEmbeddingStore.builder()
128	.collectionName("rag-collection")
129	.host(host)
130	.port(port)
131	.build();
132	}
133
134	@Bean
135	public EmbeddingModel embeddingModel() {
136	return new AllMiniLmL6V2EmbeddingModel();
137	}
138	```
139
140	## Spring Boot Integration
141
142	Inject the client via configuration:
143
144	```java
145	@Configuration
146	public class QdrantConfig {
147	@Value("${qdrant.host:localhost}")
148	private String host;
149
150	@Value("${qdrant.port:6334}")
151	private int port;
152
153	@Bean
154	public QdrantClient qdrantClient() {
155	return new QdrantClient(
156	QdrantGrpcClient.newBuilder(host, port, false).build());
157	}
158	}
159	```
160
161	## Examples
162
163	### REST Search Endpoint
164
165	```java
166	@RestController
167	@RequestMapping("/api/search")
168	public class SearchController {
169	private final VectorSearchService searchService;
170
171	public SearchController(VectorSearchService searchService) {
172	this.searchService = searchService;
173	}
174
175	@GetMapping
176	public List<ScoredPoint> search(@RequestParam String query) {
177	List<Float> queryVector = embeddingModel.embed(query).content().vectorAsList();
178	return searchService.search("documents", queryVector);
179	}
180	}
181	```
182
183	## Best Practices
184
185	- Distance metric: Cosine for normalized text embeddings, Euclidean for non-normalized.
186	- Batch upserts: Use batch operations over individual point insertions.
187	- Connection pooling: Configure connection pooling for high-throughput production workloads.
188	- Error handling: W

People who liked this also viewed

6 picks

Type