### Install dependencies and run demo

Source: https://github.com/paddlepaddle/paddleocr/blob/main/paddleocr-js/README.md

Initial setup commands to install npm dependencies and start the Vite demo application in development mode.

```bash
npm install
npm run dev:demo
```

--------------------------------

### Quick Start

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/inference_deployment/serving/paddleocr_official_api/typescript.en.md

A quick start example demonstrating how to use the `ocr` method to process an image URL.

```APIDOC
## Quick Start

```ts
import { Model, PaddleOCRClient } from "@paddleocr/api-sdk";

const client = new PaddleOCRClient();
const result = await client.ocr({
  fileUrl: "https://example.com/invoice.pdf",
  model: Model.PPOCRv5,
});
console.log(result.jobId, result.pages.length);
```

Use `filePath` for a local file. Pass exactly one of `fileUrl` or `filePath`.
```

--------------------------------

### Quick Start

Source: https://github.com/paddlepaddle/paddleocr/blob/main/paddleocr-js/packages/core/README_cn.md

A quick start guide to initialize the OCR engine and perform prediction.

```APIDOC
## Quick Start

```js
import { PaddleOCR } from "@paddleocr/paddleocr-js";

const ocr = await PaddleOCR.create({
  lang: "ch",
  ocrVersion: "PP-OCRv5",
  ortOptions: {
    backend: "auto"
  }
});

const [result] = await ocr.predict(fileOrBlob);
console.log(result.items);
```

`predict` returns an array of `OcrResult` (one for each input image). When passing a single `Blob` / `File`, you will also get an array of length 1. Use destructuring or `results[0]` to access the result.
```

--------------------------------

### Local Development Setup

Source: https://github.com/paddlepaddle/paddleocr/blob/main/api_sdk/typescript/README.md

Install dependencies and build the SDK for local development. These commands are used when contributing to the SDK.

```bash
npm install
npm run build
```

--------------------------------

### Download Example Dataset for Text Recognition

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/module_usage/text_recognition.en.md

Use this command to download the example dataset for text recognition training. Ensure you have `wget` installed.

```shell
wget https://paddle-model-ecology.bj.bcebos.com/paddlex/data/ocr_rec_dataset_examples.tar
tar -xf ocr_rec_dataset_examples.tar
```

--------------------------------

### Install Go SDK and Set Access Token

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/inference_deployment/serving/paddleocr_official_api/go.md

Install the Go SDK using 'go get' and set your AI Studio access token as an environment variable. This token is required for authentication.

```bash
go get github.com/PaddlePaddle/PaddleOCR/api_sdk/go
export PADDLEOCR_ACCESS_TOKEN="your-access-token"
```

--------------------------------

### Basic Parallel Inference Setup

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/inference_deployment/local_inference/parallel_inference.en.md

This example demonstrates a basic parallel inference setup using the PPStructureV3 pipeline across multiple GPUs. Each instance processes one input file.

```bash
python infer_mp.py \
    --pipeline PPStructureV3 \
    --input_dir input_images/ \
    --device 'gpu:0,1,2,3' \
    --output_dir output
```

--------------------------------

### Install PaddleOCR Go SDK

Source: https://github.com/paddlepaddle/paddleocr/blob/main/api_sdk/go/README.md

Use 'go get' to install the PaddleOCR Go SDK. Versioned releases are available as submodule tags.

```bash
go get github.com/PaddlePaddle/PaddleOCR/api_sdk/go
```

--------------------------------

### Quick Start with DocVLM Command Line

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/module_usage/doc_vlm.md

Use this command to quickly experience the DocVLM module's capabilities. Ensure PaddleOCR is installed.

```bash
paddleocr doc_vlm -i "{'image': 'https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/medal_table.png', 'query': '识别这份表格的内容, 以markdown格式输出'}"
```

--------------------------------

### Install PaddleOCR SDK and Set Access Token

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/inference_deployment/serving/paddleocr_official_api/typescript.en.md

Install the SDK using npm and set your access token as an environment variable. This is the initial setup required before using the SDK.

```bash
npm install @paddleocr/api-sdk
export PADDLEOCR_ACCESS_TOKEN="your-access-token"
```

--------------------------------

### Quick Start with Command Line

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/module_usage/doc_vlm.md

Experience PaddleOCR's DocVLM module with a single command line. This example uses a remote image and a query to extract table content in markdown format.

```APIDOC
## Quick Start with Command Line

Use a single command to quickly experience PaddleOCR's DocVLM:

```bash
paddleocr doc_vlm -i "{'image': 'https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/medal_table.png', 'query': '识别这份表格的内容, 以markdown格式输出'}"
```

This example uses the `paddle_dynamic` inference engine. Ensure PaddlePaddle is installed as per the [PaddlePaddle Installation Guide](../paddlepaddle_installation.md).
```

--------------------------------

### Basic Service Deployment Server Output

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/inference_deployment/serving/serving.md

Example output when the PaddleX serving server starts successfully. It indicates the server is running and ready to accept requests.

```text
INFO:     Started server process [63108]
INFO:     Waiting for application startup.
INFO:     Application startup complete.
INFO:     Uvicorn running on http://0.0.0.0:8080 (Press CTRL+C to quit)
```

--------------------------------

### Prepare Serving Environment

Source: https://github.com/paddlepaddle/paddleocr/blob/main/deploy/paddleocr_vl_docker/hps/README_en.md

Copy the example environment file and run the prepare script to download the SDK and configure the Triton pipeline.

```bash
cp .env.example .env
# Edit HPS_PIPELINE_NAME in .env if needed
bash prepare.sh
```

--------------------------------

### Install and Authenticate

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/inference_deployment/serving/paddleocr_official_api/typescript.en.md

Instructions on how to install the SDK and authenticate using an access token.

```APIDOC
## Install And Authenticate

First obtain an access token from the [AI Studio Access Token page](https://aistudio.baidu.com/account/accessToken).

```bash
npm install @paddleocr/api-sdk
export PADDLEOCR_ACCESS_TOKEN="your-access-token"
```

The client reads `PADDLEOCR_ACCESS_TOKEN` by default and also accepts `token`:

```ts
import { PaddleOCRClient } from "@paddleocr/api-sdk";

const client = new PaddleOCRClient({
  token: process.env.PADDLEOCR_ACCESS_TOKEN,
});
```
```

--------------------------------

### Quick Start with PaddleOCR Orientation Classification

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/module_usage/doc_img_orientation_classification.en.md

This command quickly demonstrates the document image orientation classification module. It requires the PaddleOCR wheel package to be installed. The example uses a sample image URL.

```bash
paddleocr doc_img_orientation_classification -i https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/img_rot180_demo.jpg
```

--------------------------------

### Execute Packaging Script

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/inference_deployment/others/packaging.md

Examples of how to run the packaging script. Use `--file` to specify your main script and `--nvidia` to include NVIDIA dependencies.

```bash
python package.py --file main.py
# 将NVIDIA的CUDA、cuDNN相关依赖库打包至可执行文件的同级目录中。
python package.py --file main.py --nvidia
```

--------------------------------

### Install X-AnyLabeling via Pip

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/data_anno_synth/x_anylabeling.md

Install the X-AnyLabeling package using pip. This command is used for quick setup and integration.

```bash
pip install x-anylabeling
```

--------------------------------

### Install and Setup Pre-commit Hooks

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/community/code_and_doc.md

Install the 'pre-commit' tool using pip and then install the Git pre-commit hooks in your current repository. This helps enforce code style and checks before committing.

```bash
pip install pre-commit
pre-commit install
```

--------------------------------

### Prepare Data and Environment for Benchmark Training (Shell)

Source: https://github.com/paddlepaddle/paddleocr/blob/main/test_tipc/docs/benchmark_train.md

Use this script to set up the training data and install the required environment for benchmark testing. It takes a configuration file and mode as arguments.

```shell
bash test_tipc/prepare.sh test_tipc/configs/det_mv3_db_v2_0/train_infer_python.txt benchmark_train
```

--------------------------------

### Install PaddleOCR and GenAI Server Dependencies

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/pipeline_usage/PaddleOCR-VL.en.md

Install PaddleOCR with doc-parser support and the dependencies for the genai server using vLLM as an example.

```shell
python -m pip install "paddleocr[doc-parser]"
paddleocr install_genai_server_deps vllm
```

--------------------------------

### Prepare Environment and Configuration

Source: https://github.com/paddlepaddle/paddleocr/blob/main/deploy/paddleocr_vl_docker/hps/README.md

Copy the example environment file and modify it as needed. The prepare.sh script downloads the SDK and configures Triton.

```bash
cp .env.example .env
# Modify HPS_PIPELINE_NAME in .env as needed
bash prepare.sh
```

--------------------------------

### Example Usage of Packaging Script

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/inference_deployment/others/packaging.en.md

Demonstrates how to run the packaging script with and without the --nvidia flag to include NVIDIA dependencies.

```bash
python package.py --file main.py
# Packages NVIDIA CUDA and cuDNN related dependencies into the same directory as the executable.
python package.py --file main.py --nvidia
```

--------------------------------

### Training log output example

Source: https://github.com/paddlepaddle/paddleocr/blob/main/ppstructure/layout/README_ch.md

Example of the expected log output when training starts successfully, showing model weight loading and epoch metrics.

```text
[08/15 04:02:30] ppdet.utils.checkpoint INFO: Finish loading model weights: /root/.cache/paddle/weights/LCNet_x1_0_pretrained.pdparams
[08/15 04:02:46] ppdet.engine INFO: Epoch: [0] [   0/1929] learning_rate: 0.040000 loss_vfl: 1.216707 loss_bbox: 1.142163 loss_dfl: 0.544196 loss: 2.903065 eta: 17 days, 13:50:26 batch_cost: 15.7452 data_cost: 2.9112 ips: 1.5243 images/s
[08/15 04:03:19] ppdet.engine INFO: Epoch: [0] [  20/1929] learning_rate: 0.064000 loss_vfl: 1.180627 loss_bbox: 0.939552 loss_dfl: 0.442436 loss: 2.628206 eta: 2 days, 12:18:53 batch_cost: 1.5770 data_cost: 0.0008 ips: 15.2184 images/s
[08/15 04:03:47] ppdet.engine INFO: Epoch: [0] [  40/1929] learning_rate: 0.088000 loss_vfl: 0.543321 loss_bbox: 1.071401 loss_dfl: 0.457817 loss: 2.057003 eta: 2 days, 0:07:03 batch_cost: 1.3190 data_cost: 0.0007 ips: 18.1954 images/s
[08/15 04:04:12] ppdet.engine INFO: Epoch: [0] [  60/1929] learning_rate: 0.112000 loss_vfl: 0.630989 loss_bbox: 0.859183 loss_dfl: 0.384702 loss: 1.883143 eta: 1 day, 19:01:29 batch_cost: 1.2177 data_cost: 0.0006 ips: 19.7087 images/s
```

--------------------------------

### Docker Compose Output Example

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/pipeline_usage/PaddleOCR-VL-AMD-GPU.en.md

This is an example of the expected output after successfully starting the PaddleOCR-VL service using Docker Compose. It indicates the Uvicorn server is running.

```text
paddleocr-vl-api             | INFO:     Started server process [1]
paddleocr-vl-api             | INFO:     Waiting for application startup.
paddleocr-vl-api             | INFO:     Application startup complete.
paddleocr-vl-api             | INFO:     Uvicorn running on http://0.0.0.0:8080 (Press CTRL+C to quit)
```

--------------------------------

### Install PaddleOCR-VL Dependencies Manually

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/pipeline_usage/PaddleOCR-VL-Iluvatar-GPU.en.md

Installs the necessary PaddlePaddle inference engine and PaddleOCR packages for manual setup. Ensure you are within an activated virtual environment.

```shell
python -m pip install paddlepaddle==3.2.0 -i https://www.paddlepaddle.org.cn/packages/stable/cpu/ 
python -m pip install paddle-iluvatar-gpu==3.2.0 -i https://www.paddlepaddle.org.cn/packages/stable/maca/ 
python -m pip install -U "paddleocr[doc-parser]"
```

--------------------------------

### Copy Environment Example

Source: https://github.com/paddlepaddle/paddleocr/blob/main/deploy/paddleocr_vl_docker/hps/README.md

Copy the .env.example file to .env to configure service variables.

```bash
cp .env.example .env
```

--------------------------------

### Install All doc2md Dependencies

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/doc2md.md

This command installs all necessary dependencies for the `doc2md` functionality within `paddleocr` in one go. It simplifies the setup process for full `doc2md` support.

```bash
pip install "paddleocr[doc2md]"
```

--------------------------------

### Pipeline Configuration Example

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/pipeline_usage/PP-ChatOCRv4.en.md

Example of a pipeline configuration file, showing where to replace model directories for fine-tuned weights.

```yaml
......
SubModules:
    TextDetection:
    module_name: text_detection
    model_name: PP-OCRv5_server_det
    model_dir: null # Replace with the fine-tuned text detection model weights directory
    limit_side_len: 960
    limit_type: max
    thresh: 0.3
    box_thresh: 0.6
    unclip_ratio: 1.5

    TextRecognition:
    module_name: text_recognition
    model_name: PP-OCRv5_server_rec
    model_dir: null # Replace with the fine-tuned text recognition model weights directory
        batch_size: 1
    batch_size: 1
            score_thresh: 0
......
```

--------------------------------

### Download and Install TensorRT (Manual)

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/inference_deployment/local_inference/high_performance_inference.en.md

Steps to download, extract, and install TensorRT, then configure the library path for use.

```bash
# Download the TensorRT tar file
wget https://developer.nvidia.com/downloads/compute/machine-learning/tensorrt/secure/8.6.1/tars/TensorRT-8.6.1.6.Linux.x86_64-gnu.cuda-11.8.tar.gz
# Extract the TensorRT tar file
tar xvf TensorRT-8.6.1.6.Linux.x86_64-gnu.cuda-11.8.tar.gz
# Install the TensorRT wheel package
python -m pip install TensorRT-8.6.1.6/python/tensorrt-8.6.1-cp310-none-linux_x86_64.whl
# Add the absolute path of the TensorRT `lib` directory to LD_LIBRARY_PATH
export LD_LIBRARY_PATH="$LD_LIBRARY_PATH:TensorRT-8.6.1.6/lib"
```

--------------------------------

### Install PaddleOCR-VL Dependencies Manually

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/pipeline_usage/PaddleOCR-VL-Iluvatar-GPU.md

This command installs the necessary PaddlePaddle framework and PaddleOCR-VL packages for manual setup. Ensure you are within an activated virtual environment and have Python versions 3.9-3.13. It's crucial to install version 3.2.0 or higher for the PaddlePaddle framework.

```shell
python -m pip install paddlepaddle==3.2.0 -i https://www.paddlepaddle.org.cn/packages/stable/cpu/ 
python -m pip install paddle-iluvatar-gpu==3.2.0 -i https://www.paddlepaddle.org.cn/packages/stable/ixuca/ 
python -m pip install -U "paddleocr[doc-parser]"
```

--------------------------------

### Device Selection Examples

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/pipeline_usage/PP-StructureV3.en.md

Examples of specifying different hardware accelerators for inference. If not set, the pipeline defaults to local GPU device 0 or CPU.

```text
npu:0
```

```text
xpu:0
```

```text
mlu:0
```

```text
dcu:0
```

```text
metax_gpu:0
```

```text
iluvatar_gpu:0
```

--------------------------------

### Install PaddleOCR-VL Dependencies Manually

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/pipeline_usage/PaddleOCR-VL-MetaX-GPU.en.md

Installs the necessary PaddlePaddle inference engine and PaddleOCR packages for manual setup. Ensure PaddlePaddle version is 3.2.0 or above. Uses specific repositories for CPU and MACA.

```shell
python -m pip install paddlepaddle==3.2.0 -i https://www.paddlepaddle.org.cn/packages/stable/cpu/ 
python -m pip install paddle-metax-gpu==3.2.0 -i https://www.paddlepaddle.org.cn/packages/stable/maca/
python -m pip install -U "paddleocr[doc-parser]"
```

--------------------------------

### Quick Start with Transformers Engine

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/module_usage/doc_img_orientation_classification.en.md

This command demonstrates using the document image orientation classification module with the 'transformers' inference engine. Ensure the Transformers environment is configured before running.

```bash
# Use the transformers engine for inference
paddleocr doc_img_orientation_classification -i https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/img_rot180_demo.jpg \
    --engine transformers
```

--------------------------------

### Quick Start Formula Recognition Command

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/module_usage/formula_recognition.md

Use this command for a quick, one-line experience of formula recognition. It requires PaddleOCR to be installed.

```bash
paddleocr formula_recognition -i https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/general_formula_rec_001.png
```

--------------------------------

### Install PaddleX Serving Plugin

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/inference_deployment/serving/serving.md

Use the PaddleX CLI to install the serving deployment plugin. This is a prerequisite for running the PaddleX server.

```bash
paddlex --install serving
```

--------------------------------

### Install DBNet.paddle using Conda or Manual Setup

Source: https://github.com/paddlepaddle/paddleocr/blob/main/benchmark/PaddleOCR_DBNet/README.MD

Provides commands to set up the environment using a YAML file or manual installation via pip and conda. Requires Python 3.6 and specific dependencies.

```bash
conda env create -f environment.yml
git clone https://github.com/WenmuZhou/DBNet.paddle.git
cd DBNet.paddle/

# Or Manual Installation
conda create -n dbnet python=3.6
conda activate dbnet
conda install ipython pip
pip install -r requirement.txt
git clone https://github.com/WenmuZhou/DBNet.paddle.git
cd DBNet.paddle/
```

--------------------------------

### Python Service Call for Table Recognition

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/pipeline_usage/table_recognition_v2.md

Example of how to call the table recognition service using Python. Ensure you have the 'requests' library installed.

```python
import requests
import json

url = "http://127.0.0.1:8866/predict/table_recognition"

# Example with image URL
with open('table.jpg', 'rb') as f:
    image_data = f.read()

files = {
    'image': ('table.jpg', image_data)
}

# Optional parameters can be added here, e.g.:
# data = {
#     "use_angle_cls": "true",
#     "det_limit_side_len": "512",
#     "det_limit_type": "max",
# }

response = requests.post(url, files=files)

if response.status_code == 200:
    result = response.json()
    print(json.dumps(result, indent=4, ensure_ascii=False))
else:
    print(f"Error: {response.status_code}")
    print(response.text)
```

--------------------------------

### Install PaddleOCR.js SDK

Source: https://github.com/paddlepaddle/paddleocr/blob/main/paddleocr-js/packages/core/README_cn.md

Install the SDK using npm. This is the first step before using the library.

```bash
npm install @paddleocr/paddleocr-js
```

--------------------------------

### Start service with command-line arguments

Source: https://github.com/paddlepaddle/paddleocr/blob/main/deploy/hubserving/readme.md

Start PaddleHub Serving with modules, port, and worker configuration. CPU-only method; use configuration file for GPU support.

```bash
hub serving start --modules Module1==Version1, Module2==Version2, ... \
                  --port 8866 \
                  --use_multiprocess \
                  --workers
```

```bash
hub serving start -m ocr_system
```

--------------------------------

### Install PaddleSlim for OCR Model Quantization

Source: https://github.com/paddlepaddle/paddleocr/blob/main/deploy/slim/quantization/README_en.md

This command installs the PaddleSlim library, a dependency required for performing quantization-aware training and model compression on PaddleOCR models. It specifies version 2.3.2 to ensure compatibility with the provided examples and functionalities.

```bash
pip3 install paddleslim==2.3.2
```

--------------------------------

### Quick Start Table Classification

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/module_usage/table_classification.en.md

This command initiates table classification using the default paddle_static inference engine. Ensure PaddlePaddle is installed before running.

```bash
paddleocr table_classification -i https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/table_recognition.jpg
```

--------------------------------

### Start PP-OCRv3 Model Training

Source: https://github.com/paddlepaddle/paddleocr/blob/main/deploy/paddlecloud/README.md

These commands initiate the training process for the PP-OCRv3 model. The first command is for single GPU (or CPU if `use_gpu=false`), and the second is for multi-GPU distributed training.

```bash
# 这里以 GPU 训练为例，使用 CPU 进行训练的话，需要指定参数 Global.use_gpu=false
python3 tools/train.py -c configs/det/ch_PP-OCRv3/ch_PP-OCRv3_det_cml.yml -o Global.save_model_dir=./output/ Global.pretrained_model=./pre_train/ch_PP-OCRv3_det_distill_train/best_accuracy
```

```bash
# 启动训练，训练模型默认保存在output目录下，--gpus '0,1,2,3'表示使用0，1，2，3号GPU训练
python3 -m paddle.distributed.launch --log_dir=./debug/ --gpus '0,1,2,3' tools/train.py -c configs/det/ch_PP-OCRv3/ch_PP-OCRv3_det_cml.yml -o Global.save_model_dir=./output/ Global.pretrained_model=./pre_train/ch_PP-OCRv3_det_distill_train/best_accuracy
```

--------------------------------

### Prepare and test with ARM_CPU backend

Source: https://github.com/paddlepaddle/paddleocr/blob/main/test_tipc/docs/test_lite_arm_cpp.md

Prepare test data, models, and Paddle-Lite prediction library for ARM_CPU testing using either download or compile method, then run tests on the device.

```shell
# 数据、模型、Paddle-Lite预测库准备
#预测库为下载方式
bash test_tipc/prepare_lite_cpp.sh ./test_tipc/configs/ch_PP-OCRv2_det/model_linux_gpu_normal_normal_lite_cpp_arm_cpu.txt download
#预测库为编译方式
bash test_tipc/prepare_lite_cpp.sh ./test_tipc/configs/ch_PP-OCRv2_det/model_linux_gpu_normal_normal_lite_cpp_arm_cpu.txt compile

# 手机端测试:
bash test_lite_arm_cpp.sh model_linux_gpu_normal_normal_lite_cpp_arm_cpu.txt
```

--------------------------------

### Python Client for PaddleOCR-VL

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/pipeline_usage/PaddleOCR-VL.md

Example of how to set up a Python client to interact with the PaddleOCR-VL service. It includes basic setup for making requests to the service URL.

```python
import base64
import requests
import pathlib

BASE_URL = "http://localhost:8080"

image_path = "./demo.jpg"

```

--------------------------------

### Initialize OpenAI Client and Submit to PP-DocBee

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/pipeline_usage/doc_understanding.en.md

This snippet shows how to initialize the OpenAI client with an API key and base URL, encode an image to base64, and then submit the image and a text prompt to the PP-DocBee model for analysis. It demonstrates a typical workflow for document understanding tasks.

```python
client = OpenAI(
    api_key='xxxxxxxxx',
    base_url=f'{API_BASE_URL}'
)

def encode_image(image_path):
  with open(image_path, "rb") as image_file:
    return base64.b64encode(image_file.read()).decode('utf-8')

image_path = "medal_table.png"

base64_image = encode_image(image_path)

response = client.chat.completions.create(
    model="pp-docbee",
    messages=[
        {
            "role": "system",
            "content": "You are a helpful assistant."
        },
        {
            "role": "user",
            "content":[
                {
                    "type": "text",
                    "text": "识别这份表格的内容,输出html格式的内容"
                },
                {
                    "type": "image_url",
                    "image_url": {"url": f"data:image/jpeg;base64,{base64_image}"}
                },
            ]
        },
    ],
)
content = response.choices[0].message.content
print('Reply:', content)
```

--------------------------------

### Node.js Service Call for Table Recognition

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/pipeline_usage/table_recognition_v2.md

Example of how to call the table recognition service using Node.js with the 'axios' library. Ensure 'axios' and 'fs' are installed.

```javascript
const axios = require('axios');
const fs = require('fs');

const url = 'http://127.0.0.1:8866/predict/table_recognition';

async function callTableRecognition() {
    try {
        const imageBuffer = fs.readFileSync('table.jpg');
        const base64Image = imageBuffer.toString('base64');

        const response = await axios.post(url, {
            image: base64Image
            // Optional parameters can be added here, e.g.:
            // use_angle_cls: 'true',
            // det_limit_side_len: '512',
            // det_limit_type: 'max'
        }, {
            headers: {
                'Content-Type': 'application/json'
            }
        });

        console.log(JSON.stringify(response.data, null, 4));
    } catch (error) {
        console.error('Error calling table recognition service:', error.response ? error.response.data : error.message);
    }
}

callTableRecognition();
```

--------------------------------

### Download Seal Text Detection Demo Dataset

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/module_usage/seal_text_detection.en.md

Downloads and extracts the example dataset for seal text detection. Ensure PaddleOCR dependencies are installed before proceeding.

```shell
wget https://paddle-model-ecology.bj.bcebos.com/paddlex/data/ocr_curve_det_dataset_examples.tar -P ./dataset
tar -xf ./dataset/ocr_curve_det_dataset_examples.tar -C ./dataset/
```

--------------------------------

### Run Benchmark with Fixture and Iterations

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/inference_deployment/cross_platform/ios_deployment.md

Utilize a pre-existing fixture for benchmarking, configuring warm-up and measured iterations.

```bash
./scripts/run_benchmark.sh --fixture ios_ocr_benchmark_reference --warmup 2 --measured-iterations 20
```

--------------------------------

### Install Training Dependencies

Source: https://github.com/paddlepaddle/paddleocr/blob/main/docs/version3.x/installation.en.md

Installs the remaining dependencies required for model training and export after cloning the repository. This command reads from the 'requirements.txt' file.

```bash
python -m pip install -r requirements.txt
```