本文面向已完成工作空间创建、文件类别配置和审核规则库设置的用户,演示如何在现有配置基础上,通过 API 完成日常的文件上传、字段抽取和智能审核流程。
如果您还没有配置过 DocFlow,请先阅读 AP 审单场景(从零开始)。
01 场景说明
工作空间、文件类别、审核规则库属于一次性基础配置,完成后即可持续复用。在日常 AP 审单业务中,只需通过 API 重复以下三步:- 上传文件:将新的单据(发票、采购合同、入库单、验收单)上传至工作空间
- 获取抽取结果:等待系统完成分类识别与字段抽取,获取结构化数据
- 智能审核:绑定已有规则库,提交审核任务并获取审核结论

02 先决条件
在运行本文代码之前,您需要准备:- 认证信息:从 TextIn 控制台 获取
x-ti-app-id和x-ti-secret-code - workspace_id:已创建的工作空间 ID(查看方式见下方)
- repo_id:已配置的审核规则库 ID(查看方式见下方)
- 待处理文件:本次需要处理的单据,例如发票、采购合同、入库单、验收单
如何获取 workspace_id
第一步:在左侧工作空间列表中,将鼠标悬停在目标空间名称上,点击出现的「设置」按钮。
workspace_id,点击复制图标可直接复制。

如何获取 repo_id
第一步:进入目标工作空间后,点击右上角「智能审核」按钮。

repo_id。

03 代码结构说明
本示例只包含日常处理所需的三个步骤,代码量比从零开始版本少约 60%。API 调用函数
| 函数(Python) | 方法(Java) | 对应 API 端点 | 说明 |
|---|---|---|---|
upload_file | uploadFile | POST /file/upload | 上传待处理文件,返回 batch_number |
submit_review_task | submitReviewTask | POST /review/task/submit | 提交审核任务,返回审核 task_id |
逐步代码说明
步骤 1:上传待处理文件
步骤 1:上传待处理文件
将四份单据逐一上传至工作空间,系统根据已配置的文件类别自动完成分类识别,并返回
batch_number。- Python
- Java
upload_targets = [
os.path.join(FILES_DIR, "sample_invoice.pdf"),
os.path.join(FILES_DIR, "sample_contract.pdf"),
os.path.join(FILES_DIR, "sample_inbound.pdf"),
os.path.join(FILES_DIR, "sample_acceptance.pdf"),
]
batch_numbers = [upload_file(WORKSPACE_ID, p) for p in upload_targets]
String[] sampleFiles = {
FILES_DIR + "/sample_invoice.pdf",
FILES_DIR + "/sample_contract.pdf",
FILES_DIR + "/sample_inbound.pdf",
FILES_DIR + "/sample_acceptance.pdf"
};
List<String> batchNumbers = new ArrayList<>();
for (String path : sampleFiles) {
batchNumbers.add(uploadFile(WORKSPACE_ID, path));
}
步骤 2:获取抽取结果
步骤 2:获取抽取结果
对每个
batch_number 轮询 file/fetch 接口,等待识别完成后获取各文档的分类和字段抽取结果,并从结果中收集 task_id 供后续审核使用。- Python
- Java
raw_results = []
for batch_number in batch_numbers:
result = wait_for_result(WORKSPACE_ID, batch_number)
raw_results.append(result)
display_result(result)
extract_task_ids = [r.get("task_id") for r in raw_results if r.get("task_id")]
List<JsonObject> rawResults = new ArrayList<>();
for (String batchNumber : batchNumbers) {
JsonObject result = waitForResult(WORKSPACE_ID, batchNumber, 120, 3);
rawResults.add(result);
displayResult(result);
}
List<String> extractTaskIds = new ArrayList<>();
for (JsonObject r : rawResults) {
if (r.has("task_id") && !r.get("task_id").isJsonNull()) {
extractTaskIds.add(r.get("task_id").getAsString());
}
}
步骤 3:提交审核任务并获取结果
步骤 3:提交审核任务并获取结果
将所有文件的
task_id 传入 review/task/submit 接口,绑定已有规则库提交审核;审核完成后轮询获取每条规则的通过/不通过结论及 AI 依据。- Python
- Java
task_name = f"AP审核_{datetime.now().strftime('%Y%m%d_%H%M%S')}"
review_task_id = submit_review_task(WORKSPACE_ID, task_name, REPO_ID, extract_task_ids)
review_result = wait_for_review(WORKSPACE_ID, review_task_id)
display_review_result(review_result)
String taskName = "AP审核_" + new SimpleDateFormat("yyyyMMdd_HHmmss").format(new Date());
String reviewTaskId = submitReviewTask(WORKSPACE_ID, taskName, REPO_ID, extractTaskIds);
JsonObject reviewResult = waitForReview(WORKSPACE_ID, reviewTaskId, 300, 5);
displayReviewResult(reviewResult);
04 完整示例代码
- Python
- Java
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
DocFlow AP 审单场景示例(已完成配置版)
适用于工作空间、文件类别、审核规则库均已配置完毕的场景。
流程:
1. 上传待处理文件(发票、采购合同、入库单、验收单)
2. 轮询获取抽取结果,展示分类与字段抽取结果
3. 提交审核任务
4. 轮询获取审核结果,展示审核结论
依赖:
pip install requests
使用前请先填写下方配置项。
"""
import os
import time
from datetime import datetime
import requests
# ============================================================
# 配置项 — 请替换为您的实际值
# ============================================================
APP_ID = "your-app-id" # TextIn 控制台中的 x-ti-app-id
SECRET_CODE = "your-secret-code" # TextIn 控制台中的 x-ti-secret-code
WORKSPACE_ID = "your-workspace-id" # 已创建的工作空间 ID
REPO_ID = "your-repo-id" # 已配置的审核规则库 ID
BASE_URL = "https://docflow.textin.com"
# 待处理文件目录(默认指向内置示例文件,可替换为您自己的文件路径)
FILES_DIR = os.path.join(
os.path.dirname(os.path.abspath(__file__)),
"..", "sample_files", "ap_review"
)
# ============================================================
# 工具辅助函数
# ============================================================
def _headers() -> dict:
return {
"x-ti-app-id": APP_ID,
"x-ti-secret-code": SECRET_CODE,
}
def _check(resp: requests.Response, action: str) -> dict:
"""校验响应状态,返回解析后的 JSON;失败时抛出 RuntimeError。"""
data = resp.json()
if data.get("code") != 200:
raise RuntimeError(f"{action} 失败(code={data.get('code')}): {data}")
return data
def _mime(file_path: str) -> str:
"""根据文件扩展名返回 MIME 类型。"""
ext = os.path.splitext(file_path)[1].lower()
return {
".png": "image/png",
".jpg": "image/jpeg",
".jpeg": "image/jpeg",
".pdf": "application/pdf",
".docx": "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
".xlsx": "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
}.get(ext, "application/octet-stream")
def display_result(file_result: dict):
"""格式化输出文件的分类结果和字段抽取结果。"""
print("\n" + "=" * 60)
print(f"文件名 : {file_result.get('name')}")
print(f"分类结果 : {file_result.get('category') or '未分类'}")
data = file_result.get("data") or {}
fields = data.get("fields") or []
if fields:
print("\n── 基本信息字段 ────────────────────────")
for f in fields:
print(f" {f.get('key', ''):<25s}: {f.get('value', '')}")
tables = data.get("tables") or []
for table in tables:
tname = table.get("tableName", "")
t_items = table.get("items") or []
if t_items:
print(f"\n── 表格[{tname}] ──────────────────────")
for row_idx, row in enumerate(t_items, start=1):
cells = " | ".join(f"{c.get('key')}={c.get('value', '')}" for c in row)
print(f" 第{row_idx}行: {cells}")
def display_review_result(review_result: dict):
"""格式化输出审核任务的结论和各规则审核结果。"""
STATUS_MAP = {
0: "未审核", 1: "审核通过", 2: "审核失败",
3: "审核中", 4: "审核不通过", 5: "识别中",
6: "排队中", 7: "识别失败",
}
RISK_MAP = {10: "高风险", 20: "中风险", 30: "低风险"}
stats = review_result.get("statistics", {})
print("\n" + "=" * 60)
print(f"审核任务状态 : {STATUS_MAP.get(review_result.get('status'), '未知')}")
print(f"规则通过数 : {stats.get('pass_count', 0)}")
print(f"规则不通过数 : {stats.get('failure_count', 0)}")
for group in review_result.get("groups", []):
print(f"\n── 规则组:{group.get('group_name')} ───────────────────")
for rt in group.get("review_tasks", []):
result_text = STATUS_MAP.get(rt.get("review_result"), "未知")
risk_text = RISK_MAP.get(rt.get("risk_level"), "未知")
icon = "✓" if rt.get("review_result") == 1 else "✗"
print(f" {icon} [{risk_text}] {rt.get('rule_name')}: {result_text}")
reasoning = rt.get("reasoning", "")
if reasoning:
print(f" 依据: {reasoning[:100]}{'...' if len(reasoning) > 100 else ''}")
# ============================================================
# 步骤 1:上传待处理文件
# REST API: POST /api/app-api/sip/platform/v2/file/upload
# ============================================================
def upload_file(workspace_id: str, file_path: str) -> str:
"""上传待处理文件至工作空间,返回 batch_number。"""
url = f"{BASE_URL}/api/app-api/sip/platform/v2/file/upload"
with open(file_path, "rb") as f:
resp = requests.post(
url,
params={"workspace_id": workspace_id},
files={"file": (os.path.basename(file_path), f, _mime(file_path))},
headers=_headers(),
timeout=60,
)
batch_number = _check(resp, "上传文件")["result"]["batch_number"]
print(f"[步骤1] 文件上传成功 name={os.path.basename(file_path)}"
f" batch_number={batch_number}")
return batch_number
# ============================================================
# 步骤 2:轮询等待抽取结果
# REST API: GET /api/app-api/sip/platform/v2/file/fetch
# ============================================================
def wait_for_result(
workspace_id: str,
batch_number: str,
timeout: int = 120,
interval: int = 3,
) -> dict:
"""
轮询直至文件识别完成,返回文件结果对象(含 task_id)。
recognition_status: 0=待识别, 1=成功, 2=失败
"""
url = f"{BASE_URL}/api/app-api/sip/platform/v2/file/fetch"
params = {"workspace_id": workspace_id, "batch_number": batch_number}
deadline = time.time() + timeout
print(f"[步骤2] 等待处理结果(batch_number={batch_number})...", end="", flush=True)
while time.time() < deadline:
resp = requests.get(url, params=params, headers=_headers(), timeout=30)
data = _check(resp, "获取处理结果")
files = data.get("result", {}).get("files", [])
if files:
status = files[0].get("recognition_status")
if status == 1:
print(" 完成")
return files[0]
elif status == 2:
raise RuntimeError(f"文件处理失败: {files[0].get('failure_causes')}")
print(".", end="", flush=True)
time.sleep(interval)
raise TimeoutError(f"等待处理结果超时({timeout}s)")
# ============================================================
# 步骤 3:提交审核任务
# REST API: POST /api/app-api/sip/platform/v2/review/task/submit
# ============================================================
def submit_review_task(
workspace_id: str,
name: str,
repo_id: str,
extract_task_ids: list,
) -> str:
"""提交审核任务,返回审核任务 task_id。"""
url = f"{BASE_URL}/api/app-api/sip/platform/v2/review/task/submit"
payload = {
"workspace_id": workspace_id,
"name": name,
"repo_id": repo_id,
"extract_task_ids": extract_task_ids,
}
resp = requests.post(url, json=payload, headers=_headers(), timeout=30)
task_id = _check(resp, "提交审核任务")["result"]["task_id"]
print(f"[步骤3] 审核任务提交成功 task_id={task_id}")
return task_id
# ============================================================
# 步骤 4:轮询等待审核结果
# REST API: POST /api/app-api/sip/platform/v2/review/task/result
# ============================================================
def wait_for_review(
workspace_id: str,
task_id: str,
timeout: int = 300,
interval: int = 5,
) -> dict:
"""
轮询直至审核任务完成,返回审核结果对象。
终态: 1=审核通过, 2=审核失败, 4=审核不通过, 7=识别失败
"""
url = f"{BASE_URL}/api/app-api/sip/platform/v2/review/task/result"
payload = {"workspace_id": workspace_id, "task_id": task_id}
deadline = time.time() + timeout
print(f"[步骤4] 等待审核结果(task_id={task_id})...", end="", flush=True)
while time.time() < deadline:
resp = requests.post(url, json=payload, headers=_headers(), timeout=30)
data = _check(resp, "获取审核结果")
result = data.get("result", {})
if result.get("status") in (1, 2, 4, 7):
print(" 完成")
return result
print(".", end="", flush=True)
time.sleep(interval)
raise TimeoutError(f"等待审核结果超时({timeout}s)")
# ============================================================
# 主流程
# ============================================================
def main():
print("=" * 60)
print(" DocFlow AP 审单场景示例(已完成配置版)")
print("=" * 60)
print(f"工作空间: {WORKSPACE_ID}")
print(f"规则库: {REPO_ID}")
# ----------------------------------------------------------
# 步骤 1:上传待处理文件
# ----------------------------------------------------------
print("\n开始上传待处理文件...")
upload_targets = [
os.path.join(FILES_DIR, "sample_invoice.pdf"),
os.path.join(FILES_DIR, "sample_contract.pdf"),
os.path.join(FILES_DIR, "sample_inbound.pdf"),
os.path.join(FILES_DIR, "sample_acceptance.pdf"),
]
batch_numbers = [upload_file(WORKSPACE_ID, p) for p in upload_targets]
# ----------------------------------------------------------
# 步骤 2:轮询获取抽取结果并展示
# ----------------------------------------------------------
print("\n开始获取处理结果...")
raw_results = []
for batch_number in batch_numbers:
result = wait_for_result(WORKSPACE_ID, batch_number)
raw_results.append(result)
display_result(result)
# ----------------------------------------------------------
# 步骤 3:提交审核任务
# ----------------------------------------------------------
print("\n开始审核...")
task_name = f"AP审核_{datetime.now().strftime('%Y%m%d_%H%M%S')}"
extract_task_ids = [r.get("task_id") for r in raw_results if r.get("task_id")]
review_task_id = submit_review_task(WORKSPACE_ID, task_name, REPO_ID, extract_task_ids)
# ----------------------------------------------------------
# 步骤 4:轮询获取审核结果并展示
# ----------------------------------------------------------
review_result = wait_for_review(WORKSPACE_ID, review_task_id)
display_review_result(review_result)
if __name__ == "__main__":
main()
package com.docflow;
import com.google.gson.*;
import okhttp3.*;
import java.io.File;
import java.io.IOException;
import java.text.SimpleDateFormat;
import java.util.*;
import java.util.concurrent.TimeUnit;
/**
* DocFlow AP 审单场景示例(已完成配置版)
*
* <p>适用于工作空间、文件类别、审核规则库均已配置完毕的场景。
* <ol>
* <li>上传待处理文件(发票、采购合同、入库单、验收单)</li>
* <li>轮询获取抽取结果,展示分类与字段抽取结果</li>
* <li>提交审核任务</li>
* <li>轮询获取审核结果,展示审核结论</li>
* </ol>
*
* <p>依赖:OkHttp 4.x、Gson(见 pom.xml)
*/
public class ApReviewConfigured {
// ============================================================
// 配置项 — 请替换为您的实际值
// ============================================================
private static final String APP_ID = "your-app-id"; // x-ti-app-id
private static final String SECRET_CODE = "your-secret-code"; // x-ti-secret-code
private static final String WORKSPACE_ID = "your-workspace-id"; // 已创建的工作空间 ID
private static final String REPO_ID = "your-repo-id"; // 已配置的审核规则库 ID
private static final String BASE_URL = "https://docflow.textin.com";
// 待处理文件目录
private static final String FILES_DIR =
new File("../sample_files/ap_review").getAbsolutePath();
// ============================================================
// 全局工具对象
// ============================================================
private static final OkHttpClient HTTP = new OkHttpClient.Builder()
.connectTimeout(30, TimeUnit.SECONDS)
.readTimeout(60, TimeUnit.SECONDS)
.writeTimeout(60, TimeUnit.SECONDS)
.build();
private static final Gson GSON = new GsonBuilder().disableHtmlEscaping().create();
private static final MediaType JSON_TYPE = MediaType.get("application/json; charset=utf-8");
// ============================================================
// 工具辅助函数
// ============================================================
private static Headers authHeaders() {
return new Headers.Builder()
.add("x-ti-app-id", APP_ID)
.add("x-ti-secret-code", SECRET_CODE)
.build();
}
private static JsonObject checkResponse(String body, String action) {
JsonObject obj = JsonParser.parseString(body).getAsJsonObject();
if (obj.get("code").getAsInt() != 200) {
throw new RuntimeException(action + " 失败: " + body);
}
return obj;
}
private static String mimeType(String filename) {
String lower = filename.toLowerCase();
if (lower.endsWith(".png")) return "image/png";
if (lower.endsWith(".jpg") || lower.endsWith(".jpeg")) return "image/jpeg";
if (lower.endsWith(".pdf")) return "application/pdf";
if (lower.endsWith(".docx")) return "application/vnd.openxmlformats-officedocument.wordprocessingml.document";
if (lower.endsWith(".xlsx")) return "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet";
return "application/octet-stream";
}
public static void displayResult(JsonObject fileResult) {
System.out.println("\n" + "=".repeat(60));
System.out.println("文件名 : " + str(fileResult, "name"));
System.out.println("分类结果 : " + str(fileResult, "category", "未分类"));
if (!fileResult.has("data") || fileResult.get("data").isJsonNull()) return;
JsonObject data = fileResult.getAsJsonObject("data");
JsonArray fields = jsonArray(data, "fields");
if (fields != null && fields.size() > 0) {
System.out.println("\n── 基本信息字段 ────────────────────────");
for (JsonElement e : fields) {
JsonObject f = e.getAsJsonObject();
System.out.printf(" %-25s: %s%n", str(f, "key"), str(f, "value"));
}
}
JsonArray tables = jsonArray(data, "tables");
if (tables != null) {
for (JsonElement te : tables) {
JsonObject table = te.getAsJsonObject();
String tname = str(table, "tableName");
JsonArray tItems = jsonArray(table, "items");
if (tItems != null && tItems.size() > 0) {
System.out.println("\n── 表格[" + tname + "] ──────────────────────");
for (int i = 0; i < tItems.size(); i++) {
JsonArray row = tItems.get(i).getAsJsonArray();
StringBuilder sb = new StringBuilder(" 第").append(i + 1).append("行: ");
for (int j = 0; j < row.size(); j++) {
JsonObject cell = row.get(j).getAsJsonObject();
if (j > 0) sb.append(" | ");
sb.append(str(cell, "key")).append("=").append(str(cell, "value"));
}
System.out.println(sb);
}
}
}
}
}
public static void displayReviewResult(JsonObject reviewResult) {
Map<Integer, String> statusMap = new LinkedHashMap<>();
statusMap.put(0, "未审核"); statusMap.put(1, "审核通过"); statusMap.put(2, "审核失败");
statusMap.put(3, "审核中"); statusMap.put(4, "审核不通过"); statusMap.put(5, "识别中");
statusMap.put(6, "排队中"); statusMap.put(7, "识别失败");
Map<Integer, String> riskMap = new LinkedHashMap<>();
riskMap.put(10, "高风险"); riskMap.put(20, "中风险"); riskMap.put(30, "低风险");
int status = reviewResult.get("status").getAsInt();
JsonObject stats = reviewResult.has("statistics")
? reviewResult.getAsJsonObject("statistics") : new JsonObject();
System.out.println("\n" + "=".repeat(60));
System.out.println("审核任务状态 : " + statusMap.getOrDefault(status, "未知"));
System.out.println("规则通过数 : " + (stats.has("pass_count") ? stats.get("pass_count").getAsInt() : 0));
System.out.println("规则不通过数 : " + (stats.has("failure_count") ? stats.get("failure_count").getAsInt() : 0));
JsonArray groups = jsonArray(reviewResult, "groups");
if (groups != null) {
for (JsonElement ge : groups) {
JsonObject group = ge.getAsJsonObject();
System.out.println("\n── 规则组:" + str(group, "group_name") + " ───────────────────");
JsonArray tasks = jsonArray(group, "review_tasks");
if (tasks != null) {
for (JsonElement te : tasks) {
JsonObject rt = te.getAsJsonObject();
int rv = rt.has("review_result") ? rt.get("review_result").getAsInt() : 0;
int riskLevel = rt.has("risk_level") ? rt.get("risk_level").getAsInt() : 0;
String icon = rv == 1 ? "✓" : "✗";
System.out.printf(" %s [%s] %s: %s%n",
icon, riskMap.getOrDefault(riskLevel, "未知"),
str(rt, "rule_name"), statusMap.getOrDefault(rv, "未知"));
String reasoning = str(rt, "reasoning");
if (!reasoning.isEmpty()) {
System.out.println(" 依据: " + (reasoning.length() > 100
? reasoning.substring(0, 100) + "..." : reasoning));
}
}
}
}
}
}
// ============================================================
// 步骤 1:上传待处理文件
// REST API: POST /api/app-api/sip/platform/v2/file/upload
// ============================================================
public static String uploadFile(String workspaceId, String filePath) throws IOException {
File file = new File(filePath);
HttpUrl url = HttpUrl.parse(BASE_URL + "/api/app-api/sip/platform/v2/file/upload")
.newBuilder()
.addQueryParameter("workspace_id", workspaceId)
.build();
MultipartBody body = new MultipartBody.Builder()
.setType(MultipartBody.FORM)
.addFormDataPart("file", file.getName(),
RequestBody.create(file, MediaType.get(mimeType(file.getName()))))
.build();
Request req = new Request.Builder()
.url(url).headers(authHeaders())
.post(body)
.build();
try (Response resp = HTTP.newCall(req).execute()) {
JsonObject data = checkResponse(resp.body().string(), "上传文件[" + file.getName() + "]");
String batchNumber = data.getAsJsonObject("result").get("batch_number").getAsString();
System.out.println("[步骤1] 文件上传成功 name=" + file.getName()
+ " batch_number=" + batchNumber);
return batchNumber;
}
}
// ============================================================
// 步骤 2:轮询等待抽取结果
// REST API: GET /api/app-api/sip/platform/v2/file/fetch
// ============================================================
public static JsonObject waitForResult(
String workspaceId, String batchNumber,
int timeoutSec, int intervalSec) throws IOException, InterruptedException {
HttpUrl url = HttpUrl.parse(BASE_URL + "/api/app-api/sip/platform/v2/file/fetch")
.newBuilder()
.addQueryParameter("workspace_id", workspaceId)
.addQueryParameter("batch_number", batchNumber)
.build();
long deadline = System.currentTimeMillis() + (long) timeoutSec * 1000;
System.out.print("[步骤2] 等待处理结果(batch_number=" + batchNumber + ")...");
while (System.currentTimeMillis() < deadline) {
Request req = new Request.Builder()
.url(url).headers(authHeaders()).get().build();
try (Response resp = HTTP.newCall(req).execute()) {
JsonObject data = checkResponse(resp.body().string(), "获取处理结果");
JsonArray files = data.getAsJsonObject("result").getAsJsonArray("files");
if (files != null && files.size() > 0) {
JsonObject file = files.get(0).getAsJsonObject();
int status = file.get("recognition_status").getAsInt();
if (status == 1) { System.out.println(" 完成"); return file; }
if (status == 2) {
String cause = file.has("failure_causes")
? file.get("failure_causes").getAsString() : "未知原因";
throw new RuntimeException("文件处理失败: " + cause);
}
}
}
System.out.print(".");
Thread.sleep((long) intervalSec * 1000);
}
throw new RuntimeException("等待处理结果超时(" + timeoutSec + "s)");
}
// ============================================================
// 步骤 3:提交审核任务
// REST API: POST /api/app-api/sip/platform/v2/review/task/submit
// ============================================================
public static String submitReviewTask(
String workspaceId, String name, String repoId,
List<String> extractTaskIds) throws IOException {
String url = BASE_URL + "/api/app-api/sip/platform/v2/review/task/submit";
JsonObject payload = new JsonObject();
payload.addProperty("workspace_id", workspaceId);
payload.addProperty("name", name);
payload.addProperty("repo_id", repoId);
JsonArray ids = new JsonArray();
extractTaskIds.forEach(ids::add);
payload.add("extract_task_ids", ids);
Request req = new Request.Builder()
.url(url).headers(authHeaders())
.post(RequestBody.create(GSON.toJson(payload), JSON_TYPE))
.build();
try (Response resp = HTTP.newCall(req).execute()) {
JsonObject data = checkResponse(resp.body().string(), "提交审核任务");
String taskId = data.getAsJsonObject("result").get("task_id").getAsString();
System.out.println("[步骤3] 审核任务提交成功 task_id=" + taskId);
return taskId;
}
}
// ============================================================
// 步骤 4:轮询等待审核结果
// REST API: POST /api/app-api/sip/platform/v2/review/task/result
// ============================================================
public static JsonObject waitForReview(
String workspaceId, String taskId,
int timeoutSec, int intervalSec) throws IOException, InterruptedException {
String url = BASE_URL + "/api/app-api/sip/platform/v2/review/task/result";
JsonObject payload = new JsonObject();
payload.addProperty("workspace_id", workspaceId);
payload.addProperty("task_id", taskId);
long deadline = System.currentTimeMillis() + (long) timeoutSec * 1000;
System.out.print("[步骤4] 等待审核结果(task_id=" + taskId + ")...");
while (System.currentTimeMillis() < deadline) {
Request req = new Request.Builder()
.url(url).headers(authHeaders())
.post(RequestBody.create(GSON.toJson(payload), JSON_TYPE))
.build();
try (Response resp = HTTP.newCall(req).execute()) {
JsonObject data = checkResponse(resp.body().string(), "获取审核结果");
JsonObject result = data.getAsJsonObject("result");
int status = result.get("status").getAsInt();
if (status == 1 || status == 2 || status == 4 || status == 7) {
System.out.println(" 完成");
return result;
}
}
System.out.print(".");
Thread.sleep((long) intervalSec * 1000);
}
throw new RuntimeException("等待审核结果超时(" + timeoutSec + "s)");
}
// ============================================================
// 主流程
// ============================================================
public static void main(String[] args) throws Exception {
System.out.println("=".repeat(60));
System.out.println(" DocFlow AP 审单场景示例(已完成配置版)");
System.out.println("=".repeat(60));
System.out.println("工作空间: " + WORKSPACE_ID);
System.out.println("规则库: " + REPO_ID);
// 步骤 1:上传待处理文件
System.out.println("\n开始上传待处理文件...");
String[] sampleFiles = {
FILES_DIR + "/sample_invoice.pdf",
FILES_DIR + "/sample_contract.pdf",
FILES_DIR + "/sample_inbound.pdf",
FILES_DIR + "/sample_acceptance.pdf"
};
List<String> batchNumbers = new ArrayList<>();
for (String path : sampleFiles) {
batchNumbers.add(uploadFile(WORKSPACE_ID, path));
}
// 步骤 2:获取并展示抽取结果
System.out.println("\n开始获取处理结果...");
List<JsonObject> rawResults = new ArrayList<>();
for (String batchNumber : batchNumbers) {
JsonObject result = waitForResult(WORKSPACE_ID, batchNumber, 120, 3);
rawResults.add(result);
displayResult(result);
}
// 步骤 3:提交审核任务
System.out.println("\n开始审核...");
String taskName = "AP审核_" + new SimpleDateFormat("yyyyMMdd_HHmmss").format(new Date());
List<String> extractTaskIds = new ArrayList<>();
for (JsonObject r : rawResults) {
if (r.has("task_id") && !r.get("task_id").isJsonNull()) {
extractTaskIds.add(r.get("task_id").getAsString());
}
}
String reviewTaskId = submitReviewTask(WORKSPACE_ID, taskName, REPO_ID, extractTaskIds);
// 步骤 4:轮询获取审核结果
JsonObject reviewResult = waitForReview(WORKSPACE_ID, reviewTaskId, 300, 5);
displayReviewResult(reviewResult);
}
// ============================================================
// 私有工具方法
// ============================================================
private static String str(JsonObject obj, String key) { return str(obj, key, ""); }
private static String str(JsonObject obj, String key, String defaultVal) {
if (obj == null || !obj.has(key) || obj.get(key).isJsonNull()) return defaultVal;
return obj.get(key).getAsString();
}
private static JsonArray jsonArray(JsonObject obj, String key) {
if (obj == null || !obj.has(key) || obj.get(key).isJsonNull()) return null;
JsonElement e = obj.get(key);
return e.isJsonArray() ? e.getAsJsonArray() : null;
}
}
05 完整示例代码下载
完整可运行代码(含 Python、Java 两个版本)已内置在文档仓库的examples/ 目录下:
examples/
├── python/
│ ├── ap_review_configured.py # Python 完整示例(已完成配置版)
│ ├── requirements.txt
│ └── README.md
├── java/
│ ├── src/main/java/com/docflow/
│ │ └── ApReviewConfigured.java # Java 完整示例(已完成配置版)
│ ├── pom.xml
│ └── README.md
└── sample_files/
└── ap_review/
├── sample_invoice.pdf
├── sample_contract.pdf
├── sample_inbound.pdf
└── sample_acceptance.pdf
Python 示例
查看 Python 完整示例代码
Java 示例
查看 Java 完整示例代码
06 运行示例
- Python
- Java
环境要求:Python 3.8+1. 安装依赖2. 填写配置打开 3. 运行
cd examples/python
pip install -r requirements.txt
ap_review_configured.py,填写文件顶部的配置项:APP_ID = "your-app-id" # x-ti-app-id
SECRET_CODE = "your-secret-code" # x-ti-secret-code
WORKSPACE_ID = "your-workspace-id" # 已创建的工作空间 ID
REPO_ID = "your-repo-id" # 已配置的审核规则库 ID
python ap_review_configured.py
环境要求:JDK 11+,Maven 3.6+1. 填写配置打开 2. 编译并运行
src/main/java/com/docflow/ApReviewConfigured.java,填写文件顶部的配置项:private static final String APP_ID = "your-app-id";
private static final String SECRET_CODE = "your-secret-code";
private static final String WORKSPACE_ID = "your-workspace-id";
private static final String REPO_ID = "your-repo-id";
cd examples/java
mvn compile exec:java -Dexec.mainClass="com.docflow.ApReviewConfigured"
运行成功后,可登录 DocFlow Web 页面,在对应工作空间下直观查看每份文件的分类、字段抽取结果和智能审核结果,便于与代码输出对照验证。
预期控制台输出
============================================================
DocFlow AP 审单场景示例(已完成配置版)
============================================================
工作空间: <workspace_id>
规则库: <repo_id>
开始上传待处理文件...
[步骤1] 文件上传成功 name=sample_invoice.pdf batch_number=<batch_number>
[步骤1] 文件上传成功 name=sample_contract.pdf batch_number=<batch_number>
[步骤1] 文件上传成功 name=sample_inbound.pdf batch_number=<batch_number>
[步骤1] 文件上传成功 name=sample_acceptance.pdf batch_number=<batch_number>
开始获取处理结果...
[步骤2] 等待处理结果(batch_number=<batch_number>)..... 完成
============================================================
文件名 : sample_invoice.pdf
分类结果 : 国内票-数电票
── 基本信息字段 ────────────────────────
价税合计 : ¥50000.00
税前金额 : ¥44247.79
发票号码 : 25312000000123456789
开票日期 : 2025年12月01日
购买方名称 : 上海合合信息科技股份有限公司
销售方名称 : 上海一二三有限公司
...
[步骤2] 等待处理结果(batch_number=<batch_number>)... 完成
============================================================
文件名 : sample_contract.pdf
分类结果 : 采购合同
── 基本信息字段 ────────────────────────
合同编号 : CG-123456
甲方全称 : 上海合合信息科技股份有限公司
乙方全称 : 上海一二三有限公司
...
[步骤2] 等待处理结果(batch_number=<batch_number>)... 完成
============================================================
文件名 : sample_inbound.pdf
分类结果 : 入库单
── 基本信息字段 ────────────────────────
收货单编号 : GR-2025-03-0201
仓库/库位 : 上海总部仓库/A区-03
...
[步骤2] 等待处理结果(batch_number=<batch_number>)... 完成
============================================================
文件名 : sample_acceptance.pdf
分类结果 : 验收单
── 基本信息字段 ────────────────────────
乙方施工人 : 胡工
验收说明 : 根据双方签订的《合同》内容,乙方已完成了合同内约定的...
...
开始审核...
[步骤3] 审核任务提交成功 task_id=<task_id>
[步骤4] 等待审核结果(task_id=<task_id>).... 完成
============================================================
审核任务状态 : 审核不通过
规则通过数 : 4
规则不通过数 : 1
── 规则组:AP审单合规性检查 ───────────────────
✓ [高风险] 发票金额与合同金额一致性: 审核通过
✓ [高风险] 入库数量与合同数量一致性: 审核通过
✓ [中风险] 验收单签署完整性: 审核通过
✗ [中风险] 发票开票日期合理性: 审核不通过
依据: 发票开票日期(2025年12月01日)晚于入库日期,不符合先到货后开票的时序要求
✓ [低风险] 供应商名称一致性: 审核通过
...
07 结果说明
抽取结果
处理完成后,每份文件将返回分类结果和字段抽取结果。字段抽取结果位于data.fields[],表格字段位于 data.items[][],每个字段包含 key、value 及坐标 position(可用于原文高亮回显)。
以下为各样本文件的实际接口返回(来自 file/fetch,省略了部分 position 坐标和字段):
sample_invoice.pdf
sample_invoice.pdf
{
"name": "sample_invoice.pdf",
"format": "pdf",
"category": "国内票-数电票",
"recognition_status": 1,
"duration_ms": 4832,
"data": {
"fields": [
{ "key": "价税合计", "value": "¥50000.00" },
{ "key": "税前金额", "value": "¥44247.79" },
{ "key": "税额", "value": "3347.79" },
{ "key": "发票号码", "value": "25312000000123456789" },
{ "key": "开票日期", "value": "2025年12月01日" },
{ "key": "购买方名称", "value": "上海合合信息科技股份有限公司" },
{ "key": "购买方纳税人识别号", "value": "91310110791485269J" },
{ "key": "销售方名称", "value": "上海一二三有限公司" },
{ "key": "销售方纳税人识别号", "value": "91310111111111111D" }
],
"items": [
[
{ "key": "商品名称", "value": "*计算机网络设备*交换机" },
{ "key": "规格型号", "value": "RG-S6510-48VS8CQ" },
{ "key": "单位", "value": "台" },
{ "key": "数量", "value": "1" },
{ "key": "单价", "value": "25752.21" },
{ "key": "金额", "value": "25752.21" },
{ "key": "税率", "value": "13%" },
{ "key": "税额", "value": "3347.79" }
]
],
"stamps": [],
"handwritings": []
}
}
sample_contract.pdf
sample_contract.pdf
{
"name": "sample_contract.pdf",
"format": "pdf",
"category": "采购合同",
"recognition_status": 1,
"data": {
"fields": [
{ "key": "合同编号", "value": "CG-123456" },
{ "key": "合同名称", "value": "合合信息采购合同" },
{ "key": "甲方全称", "value": "上海合合信息科技股份有限公司" },
{ "key": "乙方全称", "value": "上海一二三有限公司" },
{ "key": "含税总金额(大写)", "value": "伍万元整" },
{ "key": "税率", "value": "13%" },
{ "key": "账期", "value": "60日" },
{ "key": "签订日期", "value": "2025年9月25日" }
]
}
}
sample_inbound.pdf
sample_inbound.pdf
{
"name": "sample_inbound.pdf",
"format": "pdf",
"category": "入库单",
"recognition_status": 1,
"data": {
"fields": [
{ "key": "收货单编号", "value": "GR-2025-03-0201" },
{ "key": "仓库/库位", "value": "上海总部仓库/A区-03" }
]
}
}
sample_acceptance.pdf
sample_acceptance.pdf
{
"name": "sample_acceptance.pdf",
"format": "pdf",
"category": "验收单",
"recognition_status": 1,
"data": {
"fields": [
{ "key": "乙方施工人", "value": "胡工" },
{ "key": "验收说明", "value": "根据双方签订的《合同》内容,乙方已完成了合同内约定的软件到货、资料移交、安装、调试、测试、试运行等相关工作..." }
]
}
}
审核结果
审核完成后,可从review/task/result 接口获取以下信息:
status:任务整体状态(1=审核通过,4=审核不通过,2=审核失败)statistics:规则通过数、不通过数汇总groups[].review_tasks[]:每条规则的详细审核结果,包含:review_result:该规则的审核结论(1=通过,4=不通过)reasoning:AI 给出的审核依据说明anchors:依据在原文中的坐标位置(可用于高亮回显)
{
"task_id": "31415926",
"task_name": "AP审核",
"status": 1,
"statistics": { "pass_count": 6, "failure_count": 0, "error_count": 0 },
"groups": [
{
"group_name": "供应商一致性校验",
"review_tasks": [
{
"rule_name": "供应商一致性",
"risk_level": 10,
"review_result": 1,
"reasoning": "发票销售方名称"上海一二三有限公司"与采购合同乙方全称一致,纳税人识别号 91310111111111111D 匹配,审核通过。"
},
{
"rule_name": "发票与合同金额匹配",
"risk_level": 10,
"review_result": 1,
"reasoning": "发票价税合计 ¥50000.00 与采购合同含税总金额"伍万元整"一致,审核通过。"
},
{
"rule_name": "发票购买方与合同甲方一致",
"risk_level": 20,
"review_result": 1,
"reasoning": "发票购买方"上海合合信息科技股份有限公司"与采购合同甲方全称一致,审核通过。"
}
]
},
{
"group_name": "入库与合同一致性",
"review_tasks": [
{
"rule_name": "入库数量与合同数量匹配",
"risk_level": 10,
"review_result": 1,
"reasoning": "入库单实收数量与采购合同约定数量一致,审核通过。"
},
{
"rule_name": "入库单号关联合同号",
"risk_level": 20,
"review_result": 1,
"reasoning": "入库单对应合同号 CG-123456 与采购合同编号一致,审核通过。"
}
]
},
{
"group_name": "验收合规性",
"review_tasks": [
{
"rule_name": "验收单与入库单物料一致",
"risk_level": 20,
"review_result": 1,
"reasoning": "验收单合同编号与入库单对应合同号均为 CG-123456,物料关联一致,审核通过。"
}
]
}
]
}

