Pdf Ocr Tool — Pdf Ocr 工具

PDF OCR 工具 - Intelligent PDF to Markdown 转换器

Uses the Ollama GLM-OCR 模型 to intelligently recognize text, tables, and figures in PDF pages, 应用lying the most 应用ropriate prompts for OCR processing and 输出ting structured Markdown documents.

Features ✅ Smart Content 检测ion: Automatically identifies page content type (text/table/figure) ✅ Mixed Mode: Splits pages into multiple regions for processing different content types ✅ Multiple Processing Modes: Supports text, table, figure, mixed, and auto modes ✅ PDF Page-by-Page Processing: Converts PDF to images and processes each page ✅ Image OCR: Supports OCR for single images ✅ Custom Prompts: Adjustable OCR prompts based on requirements ✅ Flexible Configuration: Customizable Ollama host, port, and 模型 ✅ uv Package Management: Uses uv for Python dependency management 安装ation

• Prerequisites

# 安装 Ollama curl -fsSL https://ollama.com/安装.sh | sh ollama pull glm-ocr:q8_0

# 安装 poppler-utils (for PDF to image conversion) sudo apt 安装 poppler-utils # Debian/Ubuntu brew 安装 poppler # macOS

# 安装 uv package 管理器 curl -LsSf https://astral.sh/uv/安装.sh | sh

• 安装 with uv (Recommended)

cd 技能s/pdf-ocr-工具 uv venv source .venv/bin/activate uv 添加 请求s Pillow

• 安装 via ClawHub

npx ClawHub 安装 pdf-ocr-工具

• Manual 安装ation

# Clone or 下载 技能 git clone ~/.OpenClaw/workspace/技能s/pdf-ocr-工具

# 创建 virtual 环境 and 安装 dependencies cd ~/.OpenClaw/workspace/技能s/pdf-ocr-工具 uv venv source .venv/bin/activate uv 添加 请求s Pillow

# 运行 post-安装 script bash hooks/post-安装.sh

Usage Basic Usage # Auto-检测 content type (recommended) python ocr_工具.py --输入 document.pdf --输出 结果.md

# Specify processing mode python ocr_工具.py --输入 document.pdf --输出 结果.md --mode text python ocr_工具.py --输入 document.pdf --输出 结果.md --mode table python ocr_工具.py --输入 document.pdf --输出 结果.md --mode figure

# Mixed mode: split page into regions python ocr_工具.py --输入 document.pdf --输出 结果.md --granularity region

# Process a single image python ocr_工具.py --输入 image.png --输出 结果.md --mode mixed

Advanced Configuration # Specify Ollama host and port python ocr_工具.py --输入 document.pdf --输出 结果.md \ --host localhost --port 11434

# Use different 模型 python ocr_工具.py --输入 document.pdf --输出 结果.md \ --模型 glm-ocr:q8_0

# Custom prompt python ocr_工具.py --输入 image.png --输出 结果.md \ --prompt "Convert this table to Markdown 格式化, keeping rows and columns aligned"

# Save figure region images python ocr_工具.py --输入 document.pdf --输出 结果.md --save-images

环境 Configuration # 设置 default configuration 导出 OLLAMA_HOST="localhost" 导出 OLLAMA_PORT="11434" 导出 OCR_模型="glm-ocr:q8_0"

# 运行 python ocr_工具.py --输入 document.pdf --输出 结果.md

Processing Modes Mode Description Use Case auto Auto-检测 content type General use (default) text Pure text recognition Academic papers, articles, 报告s table Table recognition Data tables, financial 报告s figure 图表/figure recognition Statistical 图表s, flow图表s, diagrams mixed Mixed mode Pages with multiple content types Mixed Mode (Granularity)

When using --granularity region:

Page is split vertically into multiple regions (default: 3) Each region is independently analyzed for content type Cor响应ing prompts are used for OCR Final 结果s are combined into complete Markdown 输出 格式化 PDF 输出 Example # PDF to Markdown 结果 Total Pages: 15 模型: glm-ocr:q8_0 Mode: auto 生成d: 2026-02-27T01:00:00+08:00

---

Page 1

Type: mixed

Region 1 (text)

[OCR recognized text content]

Region 2 (table)

Column 1	Column 2
Data 1	Data 2

Region 3 (figure)

[图表 description] !图表

---

Image 输出 Example # image.png OCR 结果 模型: glm-ocr:q8_0 Mode: table

---

[OCR recognized 结果]

Prompt Templates

The 工具 includes four built-in prompt templates in the prompts/ directory:

Text Mode (prompts/text.md) Convert the text in this region to Markdown 格式化.

• Preserve paragraph structure and heading levels
• Handle 列出s correctly
• Preserve mathematical formulas
• MAIntAIn citations and references

Table Mode (prompts/table.md) Convert the table in this region to Markdown table 格式化.

• MAIntAIn row and column alignment
• Preserve all data and values
• Handle merged cells
• Preserve headers and units

Figure Mode (prompts/figure.md) Analyze the 图表 or image in this region:

• 图表 type (bar, line, pie, flow图表, etc.)
• Titles and axis labels
• Data trends and key observations
• 导入ant values and anomalies

Describe in Markdown 格式化.

Using in OpenClaw 导入 subprocess from pathlib 导入 Path

# Process PDF (auto mode) s