镜像概览

完整引用
docker.io/dataelement/bisheng-unstructured
Registry
docker.io
命名空间
dataelement
仓库
bisheng-unstructured
已缓存 Tag 数
20
最近同步
07/06/2026, 06:25 PM

简介

Bisheng Unstructured 文档解析服务镜像,将 PDF/Word/HTML 转为结构化文本与表格,供 RAG 流水线与 Bisheng 工作流节点调用。

详细介绍

dataelement/bisheng-unstructured 封装 Unstructured.io 兼容 API,Bisheng 平台通过 HTTP 提交文件获取分块 Markdown 与元数据,减少自研 OCR 与版面分析成本。与本地 Tesseract 脚本相比版面还原更好,但 CPU 密集、大 PDF 需调高 timeout。生产应限制上传大小、隔离内网调用,并对敏感文档启用审计日志;GPU 加速非本镜像主要路径。

快速启动

docker run -d --name bisheng-unstructured \
  -p 8000:8000 \
  docker.io/dataelement/bisheng-unstructured:latest

推荐实践

docker run -d --name bisheng-unstructured \
  --restart unless-stopped \
  -p 8000:8000 \
  -e WORKERS=4 \
  -e MAX_UPLOAD_MB=50 \
  -v /tmp/unstructured:/tmp \
  docker.io/dataelement/bisheng-unstructured:1.0.0

核心参数说明

  • -p 8000:8000 — 文档解析 REST API
  • -e WORKERS — 并发处理进程数
  • -e MAX_UPLOAD_MB — 单文件上传上限
  • -v /tmp/unstructured:/tmp — 临时解压与转换目录
  • --restart unless-stopped — RAG 前置服务不中断
  • 固定 tag — 与 Bisheng 主版本 API 契约一致

Kubernetes

apiVersion: apps/v1
kind: Deployment
metadata:
  name: bisheng-unstructured
spec:
  replicas: 2
  selector:
    matchLabels:
      app: bisheng-unstructured
  template:
    metadata:
      labels:
        app: bisheng-unstructured
    spec:
      containers:
        - name: unstructured
          image: docker.io/dataelement/bisheng-unstructured:1.0.0
          ports:
            - containerPort: 8000
          env:
            - name: WORKERS
              value: "4"
          resources:
            requests:
              cpu: "1"
              memory: 2Gi
---
apiVersion: v1
kind: Service
metadata:
  name: bisheng-unstructured
spec:
  selector:
    app: bisheng-unstructured
  ports:
    - port: 8000
      targetPort: 8000

CPU 请求按并发调高;Bisheng 后端经 ClusterIP 调用,临时文件可用 emptyDir,敏感环境禁用公网 Ingress。

最新 Tag(20 条)

查看全部 20 个 Tag →
Tag架构推送时间大小Digest同步期望操作
docker.io/dataelement/bisheng-unstructured
:latest
linux/amd64
2025/05/20 10:356.5 GBsha256:e2878a87cfad
docker.io/dataelement/bisheng-unstructured
:v0.0.3.14
linux/amd64
2025/05/20 10:346.5 GBsha256:e2878a87cfad
docker.io/dataelement/bisheng-unstructured
:v0.0.3.13
linux/amd64
2024/11/22 11:506.5 GBsha256:a9346f71f6ab
docker.io/dataelement/bisheng-unstructured
:v0.0.3.12
linux/amd64
2024/11/11 11:536.5 GBsha256:42a38f036a47
docker.io/dataelement/bisheng-unstructured
:v0.0.3.11
linux/amd64
2024/10/12 17:186.4 GBsha256:6568e414f3f3
docker.io/dataelement/bisheng-unstructured
:v0.0.3.10
linux/amd64
2024/09/14 19:386.4 GBsha256:15c70e2eedc0
docker.io/dataelement/bisheng-unstructured
:0.0.3.1
linux/amd64
2024/09/03 17:296.0 GBsha256:631d4aa32868
docker.io/dataelement/bisheng-unstructured
:0.0.3.2
linux/amd64
2024/09/03 17:296.0 GBsha256:90943ec3f609
docker.io/dataelement/bisheng-unstructured
:0.0.2
linux/amd64
2024/09/03 17:296.0 GBsha256:cd33aac0c776
docker.io/dataelement/bisheng-unstructured
:0.0.3.3
linux/amd64
2024/09/03 17:296.0 GBsha256:18e44a84920f
docker.io/dataelement/bisheng-unstructured
:0.0.2.1
linux/amd64
2024/09/03 17:296.0 GBsha256:31db3a17367f
docker.io/dataelement/bisheng-unstructured
:0.0.3
linux/amd64
2024/09/03 17:296.0 GBsha256:d366061b64b6
docker.io/dataelement/bisheng-unstructured
:v0.0.3.9
linux/amd64
2024/09/02 23:536.4 GBsha256:40136df462fb
docker.io/dataelement/bisheng-unstructured
:v0.0.3.8
linux/amd64
2024/08/23 19:036.3 GBsha256:afc6a0a6a0e1
docker.io/dataelement/bisheng-unstructured
:v0.0.3.7
linux/amd64
2024/08/06 15:296.3 GBsha256:409355acc944
docker.io/dataelement/bisheng-unstructured
:v0.0.3.6
linux/amd64linux/arm64
2024/05/17 15:490.0 MBsha256:30d66e417e72
docker.io/dataelement/bisheng-unstructured
:0.0.3.6
linux/amd64
2024/05/17 10:506.3 GBsha256:71769fafae1a
docker.io/dataelement/bisheng-unstructured
:0.0.3.5
linux/amd64
2024/04/24 11:206.1 GBsha256:798d4a5bd164
docker.io/dataelement/bisheng-unstructured
:0.0.3.4
linux/amd64
2024/01/25 15:346.0 GBsha256:4a873f4d602d
docker.io/dataelement/bisheng-unstructured
:0.0.1
linux/amd64
2023/11/09 11:39709.8 MBsha256:f1406fea0394