Azure AI Contentsafety Java

1---
2name: azure-ai-contentsafety-java
3description: Build content moderation applications with Azure AI Content Safety SDK for Java. Use when implementing text/image analysis, blocklist management, or harm detection for hate, violence, sexual content, and self-harm.
4package: com.azure:azure-ai-contentsafety
5---
6 
7# Azure AI Content Safety SDK for Java
8 
9Build content moderation applications using the Azure AI Content Safety SDK for Java.
10 
11## Installation
12 
13```xml
14<dependency>
15    <groupId>com.azure</groupId>
16    <artifactId>azure-ai-contentsafety</artifactId>
17    <version>1.1.0-beta.1</version>
18</dependency>
19```
20 
21## Client Creation
22 
23### With API Key
24 
25```java
26import com.azure.ai.contentsafety.ContentSafetyClient;
27import com.azure.ai.contentsafety.ContentSafetyClientBuilder;
28import com.azure.ai.contentsafety.BlocklistClient;
29import com.azure.ai.contentsafety.BlocklistClientBuilder;
30import com.azure.core.credential.KeyCredential;
31 
32String endpoint = System.getenv("CONTENT_SAFETY_ENDPOINT");
33String key = System.getenv("CONTENT_SAFETY_KEY");
34 
35ContentSafetyClient contentSafetyClient = new ContentSafetyClientBuilder()
36    .credential(new KeyCredential(key))
37    .endpoint(endpoint)
38    .buildClient();
39 
40BlocklistClient blocklistClient = new BlocklistClientBuilder()
41    .credential(new KeyCredential(key))
42    .endpoint(endpoint)
43    .buildClient();
44```
45 
46### With DefaultAzureCredential
47 
48```java
49import com.azure.identity.DefaultAzureCredentialBuilder;
50 
51ContentSafetyClient client = new ContentSafetyClientBuilder()
52    .credential(new DefaultAzureCredentialBuilder().build())
53    .endpoint(endpoint)
54    .buildClient();
55```
56 
57## Key Concepts
58 
59### Harm Categories
60| Category | Description |
61|----------|-------------|
62| Hate | Discriminatory language based on identity groups |
63| Sexual | Sexual content, relationships, acts |
64| Violence | Physical harm, weapons, injury |
65| Self-harm | Self-injury, suicide-related content |
66 
67### Severity Levels
68- Text: 0-7 scale (default outputs 0, 2, 4, 6)
69- Image: 0, 2, 4, 6 (trimmed scale)
70 
71## Core Patterns
72 
73### Analyze Text
74 
75```java
76import com.azure.ai.contentsafety.models.*;
77 
78AnalyzeTextResult result = contentSafetyClient.analyzeText(
79    new AnalyzeTextOptions("This is text to analyze"));
80 
81for (TextCategoriesAnalysis category : result.getCategoriesAnalysis()) {
82    System.out.printf("Category: %s, Severity: %d%n",
83        category.getCategory(),
84        category.getSeverity());
85}
86```
87 
88### Analyze Text with Options
89 
90```java
91AnalyzeTextOptions options = new AnalyzeTextOptions("Text to analyze")
92    .setCategories(Arrays.asList(
93        TextCategory.HATE,
94        TextCategory.VIOLENCE))
95    .setOutputType(AnalyzeTextOutputType.EIGHT_SEVERITY_LEVELS);
96 
97AnalyzeTextResult result = contentSafetyClient.analyzeText(options);
98```
99 
100### Analyze Text with Blocklist
101 
102```java
103AnalyzeTextOptions options = new AnalyzeTextOptions("I h*te you and want to k*ll you")
104    .setBlocklistNames(Arrays.asList("my-blocklist"))
105    .setHaltOnBlocklistHit(true);
106 
107AnalyzeTextResult result = contentSafetyClient.analyzeText(options);
108 
109if (result.getBlocklistsMatch() != null) {
110    for (TextBlocklistMatch match : result.getBlocklistsMatch()) {
111        System.out.printf("Blocklist: %s, Item: %s, Text: %s%n",
112            match.getBlocklistName(),
113            match.getBlocklistItemId(),
114            match.getBlocklistItemText());
115    }
116}
117```
118 
119### Analyze Image
120 
121```java
122import com.azure.ai.contentsafety.models.*;
123import com.azure.core.util.BinaryData;
124import java.nio.file.Files;
125import java.nio.file.Paths;
126 
127// From file
128byte[] imageBytes = Files.readAllBytes(Paths.get("image.png"));
129ContentSafetyImageData imageData = new ContentSafetyImageData()
130    .setContent(BinaryData.fromBytes(imageBytes));
131 
132AnalyzeImageResult result = contentSafetyClient.analyzeImage(
133    new AnalyzeImageOptions(imageData));
134 
135for (ImageCategoriesAnalysis category : result.getCategoriesAnalysis()) {
136    System.out.printf("Category: %s, Severity: %d%n",
137        category.getCategory(),
138        category.getSeverity());
139}
140```
141 
142### Analyze Image from URL
143 
144```java
145ContentSafetyImageData imageData = new ContentSafetyImageData()
146    .setBlobUrl("https://example.com/image.jpg");
147 
148AnalyzeImageResult result = contentSafetyClient.analyzeImage(
149    new AnalyzeImageOptions(imageData));
150```
151 
152## Blocklist Management
153 
154### Create or Update Blocklist
155 
156```java
157import com.azure.core.http.rest.RequestOptions;
158import com.azure.core.http.rest.Response;
159import com.azure.core.util.BinaryData;
160import java.util.Map;
161 
162Map<String, String> description = Map.of("description", "Custom blocklist");
163BinaryData resource = BinaryData.fromObject(description);
164 
165Response<BinaryData> response = blocklistClient.createOrUpdateTextBlocklistWithResponse(
166    "my-blocklist", resource, new RequestOptions());
167 
168if (response.getStatusCode() == 201) {
169    System.out.println("Blocklist created");
170} else if (response.getStatusCode() == 200) {
171    System.out.println("Blocklist updated");
172}
173```
174 
175### Add Block Items
176 
177```java
178import com.azure.ai.contentsafety.models.*;
179import java.util.Arrays;
180 
181List<TextBlocklistItem> items = Arrays.asList(
182    new TextBlocklistItem("badword1").setDescription("Offensive term"),
183    new TextBlocklistItem("badword2").setDescription("Another term")
184);
185 
186AddOrUpdateTextBlocklistItemsResult result = blocklistClient.addOrUpdateBlocklistItems(
187    "my-blocklist",
188    new AddOrUpdateTextBlocklistItemsOptions(items));
189 
190for (TextBlocklistItem item : result.getBlocklistItems()) {
191    System.out.printf("Added: %s (ID: %s)%n",
192        item.getText(),
193        item.getBlocklistItemId());
194}
195```
196 
197### List Blocklists
198 
199```java
200PagedIterable<TextBlocklist> blocklists = blocklistClient.listTextBlocklists();
201 
202for (TextBlocklist blocklist : blocklists) {
203    System.out.printf("Blocklist: %s, Description: %s%n",
204        blocklist.getName(),
205        blocklist.getDescription());
206}
207```
208 
209### Get Blocklist
210 
211```java
212TextBlocklist blocklist = blocklistClient.getTextBlocklist("my-blocklist");
213System.out.println("Name: " + blocklist.getName());
214```
215 
216### List Block Items
217 
218```java
219PagedIterable<TextBlocklistItem> items = 
220    blocklistClient.listTextBlocklistItems("my-blocklist");
221 
222for (TextBlocklistItem item : items) {
223    System.out.printf("ID: %s, Text: %s%n",
224        item.getBlocklistItemId(),
225        item.getText());
226}
227```
228 
229### Remove Block Items
230 
231```java
232List<String> itemIds = Arrays.asList("item-id-1", "item-id-2");
233 
234blocklistClient.removeBlocklistItems(
235    "my-blocklist",
236    new RemoveTextBlocklistItemsOptions(itemIds));
237```
238 
239### Delete Blocklist
240 
241```java
242blocklistClient.deleteTextBlocklist("my-blocklist");
243```
244 
245## Error Handling
246 
247```java
248import com.azure.core.exception.HttpResponseException;
249 
250try {
251    contentSafetyClient.analyzeText(new AnalyzeTextOptions("test"));
252} catch (HttpResponseException e) {
253    System.out.println("Status: " + e.getResponse().getStatusCode());
254    System.out.println("Error: " + e.getMessage());
255    // Common codes: InvalidRequestBody, ResourceNotFound, TooManyRequests
256}
257```
258 
259## Environment Variables
260 
261```bash
262CONTENT_SAFETY_ENDPOINT=https://<resource>.cognitiveservices.azure.com/
263CONTENT_SAFETY_KEY=<your-api-key>
264```
265 
266## Best Practices
267 
2681. **Blocklist Delay**: Changes take ~5 minutes to take effect
2692. **Category Selection**: Only request needed categories to reduce latency
2703. **Severity Thresholds**: Typically block severity >= 4 for strict moderation
2714. **Batch Processing**: Process multiple items in parallel for throughput
2725. **Caching**: Cache blocklist results where appropriate
273 
274## Trigger Phrases
275 
276- "content safety Java"
277- "content moderation Azure"
278- "analyze text safety"
279- "image moderation Java"
280- "blocklist management"
281- "hate speech detection"
282- "harmful content filter"
283