Information extraction
This topic describes the syntax of the Tongyi Data Mining information extraction API and provides a request example.
Request syntax
POST /zhiwen-chat/extraction HTTP/1.1
Request parameters
Name | Type | Required | Description | Example value |
|---|---|---|---|---|
capabilityType | string | Yes | The capability type. Bidding document extraction: BID_EXTRACTION. Resume extraction: RESUME_EXTRACTION. Entity extraction: ENTITY_EXTRACTION. | ENTITY_EXTRACTION |
userPrompt | string | Yes | The prompt. | You are an AI assistant that specializes in named entity recognition (NER). Your task is to accurately extract all names of people, places, and organizations from the given text and organize them in the specified dictionary format. |
fileIdList | array | Yes | The list of document IDs. | [ "file_zhiwen_XXX" ] |
stream | boolean | Yes | Specifies whether to return the response in a stream. | true |
Response parameters
Name | Type | Description |
|---|---|---|
output | object | |
-choices | array | |
-finishReason | string | Termination reason |
-message | object | The message body. |
-role | string | The role. |
-content | string | The message content. |
requestId | string | The request ID. |
usage | object | Information about token usage. |
-inputTokens | int | The number of input tokens. |
-outputTokens | int | The number of output tokens. |
-totalTokens | int | The total number of tokens. |
Examples
Call example
import os
import requests
'''
API name: Information Extraction
API path: https://dashscope.aliyuncs.com/api/v2/apps/zhiwen-chat/extraction
Environment requirements: Python >= 3.7
'''
url = "https://dashscope.aliyuncs.com/api/v2/apps/zhiwen-chat/extraction"
request_headers = {
"Authorization": os.environ.get("DASHSCOPE_API_KEY"), # If you have not configured an environment variable, replace this with your API key.
"Content-Type": "application/json"
}
# Information extraction
# Bidding document extraction: BID_EXTRACTION. Resume extraction: RESUME_EXTRACTION. Entity extraction: ENTITY_EXTRACTION.
request_body = {
"capabilityType": "ENTITY_EXTRACTION",
"userPrompt": "You are an AI assistant that specializes in named entity recognition (NER). Your task is to accurately extract all names of people, places, and organizations from the given text and organize them in the specified dictionary format.",
"stream": True,
"fileIdList": [
"file_zhiwen_XXX" # Replace this with your document ID.
]
}
with requests.post(url, json=request_body, headers=request_headers) as response:
if request_body["stream"]:
for line in response.iter_lines():
if line:
line = line.decode('utf-8')
if line.startswith('data:'):
print("Received:", line[5:].strip())
else:
print("Received:", response.json())
import com.fasterxml.jackson.databind.ObjectMapper;
import org.apache.http.HttpEntity;
import org.apache.http.client.config.RequestConfig;
import org.apache.http.client.methods.CloseableHttpResponse;
import org.apache.http.client.methods.HttpPost;
import org.apache.http.entity.StringEntity;
import org.apache.http.impl.client.CloseableHttpClient;
import org.apache.http.impl.client.HttpClients;
import org.apache.http.util.EntityUtils;
import java.io.BufferedReader;
import java.io.IOException;
import java.io.InputStream;
import java.io.InputStreamReader;
import java.util.HashMap;
import java.util.Map;
/**
* API name: Information Extraction
* API path: https://dashscope.aliyuncs.com/api/v2/apps/zhiwen-chat/extraction
* Environment requirements: Java 8 or later
*/
public class ExtractionDemo {
// Step 1. API address and configuration
private static final String API_KEY = System.getenv("DASHSCOPE_API_KEY"); // Specify the API key of Alibaba Cloud Model Studio.
private static final String API_URL = "https://dashscope.aliyuncs.com/api/v2/apps/zhiwen-chat/extraction";
private static final String CHARSET = "UTF-8";
private static final ObjectMapper objectMapper = new ObjectMapper();
public static void main(String[] args) throws IOException, InterruptedException {
// Step 2. Parameter settings
// The fileId field returned by the document upload API. The format is file_zhiwen_XXX.
String fileId = "file_zhiwen_XXX";
// Custom prompt: Modify as needed. For an example, see the debugging page of the Tongyi ZoneWit lightweight application.
String userPrompt = "You are an AI assistant that specializes in named entity recognition (NER). Your task is to accurately extract all names of people, places, and organizations from the given text and organize them in the specified dictionary format.";
// Streaming output: true
boolean isStreaming = true;
// Step 3. Build the request body
Map<String, Object> requestBodyMap = new HashMap<>();
requestBodyMap.put("capabilityType", "ENTITY_EXTRACTION"); // Bidding document extraction: BID_EXTRACTION. Resume extraction: RESUME_EXTRACTION. Entity extraction: ENTITY_EXTRACTION.
requestBodyMap.put("userPrompt", userPrompt);
requestBodyMap.put("stream", isStreaming);
requestBodyMap.put("fileIdList", new String[]{fileId});
String requestBody = objectMapper.writeValueAsString(requestBodyMap);
// Step 4. Configure the request timeout period
RequestConfig requestConfig = RequestConfig.custom().setConnectTimeout(30 * 1000).build();
// Step 5. Execute the request. Select a streaming or non-streaming request as needed.
try (CloseableHttpClient httpClient = HttpClients.custom()
.setDefaultRequestConfig(requestConfig)
.build()) {
// Construct an HTTP POST request.
HttpPost httpPost = new HttpPost(API_URL);
buildRequestHeaders(httpPost, requestBody);
// Execute the request. This example uses streaming output. You can configure the stream parameter as needed.
buildStreamingRequest(httpClient, httpPost); // Streaming request
// buildNonStreamingRequest(httpClient, httpPost); // Non-streaming request
}
}
/**
* Build common request headers.
*
* @param httpPost The HTTP request object.
* @param requestBody The request body string.
*/
private static void buildRequestHeaders(HttpPost httpPost, String requestBody) {
httpPost.setHeader("Authorization", API_KEY);
httpPost.setHeader("Content-Type", "application/json");
httpPost.setEntity(new StringEntity(requestBody, CHARSET));
}
/**
* Build a streaming request.
* @param client The HTTP client.
* @param httpPost The HTTP request object.
*/
private static void buildStreamingRequest(CloseableHttpClient client, HttpPost httpPost) {
System.out.println("Streaming request:");
try (CloseableHttpResponse response = client.execute(httpPost)) {
HttpEntity entity = response.getEntity();
int statusCode = response.getStatusLine().getStatusCode();
if (statusCode == 200 && entity != null) {
try (InputStream inputStream = entity.getContent();
BufferedReader reader = new BufferedReader(new InputStreamReader(inputStream))) {
String line;
while ((line = reader.readLine()) != null) {
if (line.startsWith("data:")) {
String json = line.substring(5).trim();
System.out.println(json);
}
}
}
} else {
String errorBody = EntityUtils.toString(response.getEntity());
System.err.println("Request failed: " + errorBody);
}
} catch (IOException e) {
System.err.println("Request failed: " + e.getMessage());
e.printStackTrace();
}
}
/**
* Build a non-streaming request.
* @param client The HTTP client.
* @param httpPost The HTTP request object.
* @throws IOException if a network request or data read error occurs.
*/
private static void buildNonStreamingRequest(CloseableHttpClient client, HttpPost httpPost) throws IOException {
System.out.println("Non-streaming request:");
try (CloseableHttpResponse response = client.execute(httpPost)) {
HttpEntity entity = response.getEntity();
if (entity == null) {
System.err.println("The response body is empty.");
return;
}
int statusCode = response.getStatusLine().getStatusCode();
if (statusCode == 200 && entity != null){
String responseBody = EntityUtils.toString(entity);
System.out.println(responseBody);
}else{
String errorBody = EntityUtils.toString(entity);
System.err.println("Request failed: " + errorBody);
}
}catch (IOException e) {
System.err.println("Request failed: " + e.getMessage());
e.printStackTrace();
}
}
}
Normal response example
JSON format
{
"output": {
"choices": [{
"finishReason": "stop",
"message": {
"role": "assistant",
"content": "OK. I will extract all names of people, places, and organizations from the given text and organize them in the specified dictionary format: Document information extraction result."
}
}]
},
"requestId": "60a9c0f4-9414-9c94-81f3-f487f7769c9a",
"usage": {
"inputTokens": 100,
"outputTokens": 100,
"totalTokens": 200
}
}
Error codes
For more information, see Error codes - Tongyi Data Mining.