bookdatamaker 0.2.3__tar.gz → 0.2.4__tar.gz

{bookdatamaker-0.2.3/src/bookdatamaker.egg-info → bookdatamaker-0.2.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: bookdatamaker
-Version: 0.2.3
+Version: 0.2.4
 Summary: CLI tool for extracting text with DeepSeek OCR and generating datasets
 Author-email: zwh20081 <zwh20081@solart.pro>
 License: MIT
@@ -89,6 +89,49 @@ A powerful CLI tool for extracting text from documents using DeepSeek OCR and ge
 - 🌐 **Flexible Modes**: API or self-hosted for both stages
 - 📈 **Progress Tracking**: Real-time progress bars
 
+## Installation
+
+### From PyPI (Recommended)
+
+```bash
+pip install bookdatamaker
+```
+
+### From Source
+
+```bash
+git clone https://github.com/yourusername/bookdatamaker.git
+cd bookdatamaker
+pip install -r requirements.txt
+pip install -e .
+```
+
+### Optional: Local Inference Support
+
+```bash
+# For self-hosted OCR and LLM generation
+pip install bookdatamaker[local]  # From PyPI
+# OR
+pip install -e ".[local]"  # From source - installs transformers==4.46.3, torch, flash-attn, etc.
+```
+
+**Note**: The project requires `transformers==4.46.3` for optimal compatibility with DeepSeek-OCR. A warning will be displayed if a different version is detected.
+
+### System Requirements
+
+**For API Mode:**
+- Python 3.10+
+- API keys (OpenAI, DeepSeek, etc.)
+
+**For Local Mode:**
+- Python 3.10-3.12 (3.13 not supported due to vLLM compatibility)
+- NVIDIA GPU with CUDA support (or CPU, though slower)
+- 16GB+ VRAM recommended for GPU
+- transformers==4.46.3
+- Linux or WSL2 (recommended)
+
+---
+
 ## Quick Start
 
 ### Prerequisites
@@ -103,7 +146,7 @@ export DEEPSEEK_API_KEY=your_deepseek_key    # For API OCR mode
 
 ```bash
 # 1. Install
-pip install -r requirements.txt && pip install -e .
+pip install bookdatamaker
 
 # 2. Extract → Generate → Export
 bookdatamaker extract book.pdf -o ./extracted
@@ -115,7 +158,7 @@ bookdatamaker export-dataset dataset.db -o output.parquet
 
 ```bash
 # 1. Install with local dependencies
-pip install -r requirements.txt && pip install -e ".[local]"
+pip install bookdatamaker[local]
 
 # 2. Extract with local OCR
 bookdatamaker extract book.pdf --mode local --batch-size 8 -o ./extracted
@@ -131,27 +174,9 @@ bookdatamaker generate ./extracted \
 bookdatamaker export-dataset dataset.db -o output.parquet
 ```
 
-## Installation
-
-### Basic Installation
-
-```bash
-git clone https://github.com/yourusername/bookdatamaker.git
-cd bookdatamaker
-pip install -r requirements.txt
-pip install -e .
-```
-
-### Optional: Local Inference Support
-
-```bash
-# For self-hosted OCR and LLM generation
-pip install -e ".[local]"  # Installs transformers==4.46.3, torch, flash-attn, etc.
-```
-
-**Note**: The project requires `transformers==4.46.3` for optimal compatibility with DeepSeek-OCR. A warning will be displayed if a different version is detected.
+---
 
-### System Requirements
+## System Requirements
 
 **For API Mode:**
 - Python 3.10+

{bookdatamaker-0.2.3 → bookdatamaker-0.2.4}/README.md

@@ -37,6 +37,49 @@ A powerful CLI tool for extracting text from documents using DeepSeek OCR and ge
 - 🌐 **Flexible Modes**: API or self-hosted for both stages
 - 📈 **Progress Tracking**: Real-time progress bars
 
+## Installation
+
+### From PyPI (Recommended)
+
+```bash
+pip install bookdatamaker
+```
+
+### From Source
+
+```bash
+git clone https://github.com/yourusername/bookdatamaker.git
+cd bookdatamaker
+pip install -r requirements.txt
+pip install -e .
+```
+
+### Optional: Local Inference Support
+
+```bash
+# For self-hosted OCR and LLM generation
+pip install bookdatamaker[local]  # From PyPI
+# OR
+pip install -e ".[local]"  # From source - installs transformers==4.46.3, torch, flash-attn, etc.
+```
+
+**Note**: The project requires `transformers==4.46.3` for optimal compatibility with DeepSeek-OCR. A warning will be displayed if a different version is detected.
+
+### System Requirements
+
+**For API Mode:**
+- Python 3.10+
+- API keys (OpenAI, DeepSeek, etc.)
+
+**For Local Mode:**
+- Python 3.10-3.12 (3.13 not supported due to vLLM compatibility)
+- NVIDIA GPU with CUDA support (or CPU, though slower)
+- 16GB+ VRAM recommended for GPU
+- transformers==4.46.3
+- Linux or WSL2 (recommended)
+
+---
+
 ## Quick Start
 
 ### Prerequisites
@@ -51,7 +94,7 @@ export DEEPSEEK_API_KEY=your_deepseek_key    # For API OCR mode
 
 ```bash
 # 1. Install
-pip install -r requirements.txt && pip install -e .
+pip install bookdatamaker
 
 # 2. Extract → Generate → Export
 bookdatamaker extract book.pdf -o ./extracted
@@ -63,7 +106,7 @@ bookdatamaker export-dataset dataset.db -o output.parquet
 
 ```bash
 # 1. Install with local dependencies
-pip install -r requirements.txt && pip install -e ".[local]"
+pip install bookdatamaker[local]
 
 # 2. Extract with local OCR
 bookdatamaker extract book.pdf --mode local --batch-size 8 -o ./extracted
@@ -79,27 +122,9 @@ bookdatamaker generate ./extracted \
 bookdatamaker export-dataset dataset.db -o output.parquet
 ```
 
-## Installation
-
-### Basic Installation
-
-```bash
-git clone https://github.com/yourusername/bookdatamaker.git
-cd bookdatamaker
-pip install -r requirements.txt
-pip install -e .
-```
-
-### Optional: Local Inference Support
-
-```bash
-# For self-hosted OCR and LLM generation
-pip install -e ".[local]"  # Installs transformers==4.46.3, torch, flash-attn, etc.
-```
-
-**Note**: The project requires `transformers==4.46.3` for optimal compatibility with DeepSeek-OCR. A warning will be displayed if a different version is detected.
+---
 
-### System Requirements
+## System Requirements
 
 **For API Mode:**
 - Python 3.10+

{bookdatamaker-0.2.3 → bookdatamaker-0.2.4}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "bookdatamaker"
-version = "0.2.3"
+version = "0.2.4"
 description = "CLI tool for extracting text with DeepSeek OCR and generating datasets"
 readme = "README.md"
 requires-python = ">=3.10,<3.13"

{bookdatamaker-0.2.3 → bookdatamaker-0.2.4}/src/bookdatamaker/cli.py

@@ -14,11 +14,16 @@ from bookdatamaker.ocr import OCRExtractor
 from bookdatamaker.utils import PageManager
 
 
-@click.group()
+@click.group(invoke_without_command=True)
 @click.version_option(version="0.1.0")
-def cli() -> None:
+@click.pass_context
+def cli(ctx: click.Context) -> None:
     """Book Data Maker - Extract text and generate datasets."""
     load_dotenv()
+    
+    # Show help if no command provided
+    if ctx.invoked_subcommand is None:
+        click.echo(ctx.get_help())
 
 
 @cli.command()

{bookdatamaker-0.2.3 → bookdatamaker-0.2.4}/src/bookdatamaker/ocr/extractor.py

@@ -8,7 +8,6 @@ import os
 
 import httpx
 from PIL import Image
-import torch
 
 
 class OCRExtractor:
@@ -75,6 +74,18 @@ class OCRExtractor:
         """Initialize local transformers model."""
         import importlib.metadata
         
+        # Import torch here (only when needed for local mode)
+        try:
+            import torch
+        except ImportError:
+            raise ImportError(
+                "PyTorch not found. Please install it for local mode:\n"
+                "  pip install bookdatamaker[local]"
+            )
+        
+        # Store torch reference for later use
+        self.torch = torch
+        
         # Check transformers version
         try:
             transformers_version = importlib.metadata.version("transformers")
@@ -217,10 +228,10 @@ class OCRExtractor:
         )
         
         # Move inputs to GPU
-        inputs = {k: v.cuda() if isinstance(v, torch.Tensor) else v 
+        inputs = {k: v.cuda() if isinstance(v, self.torch.Tensor) else v 
                   for k, v in inputs.items()}
         
-        with torch.no_grad():
+        with self.torch.no_grad():
             outputs = self.model.generate(**inputs, max_new_tokens=8192)
         
         # Decode output

{bookdatamaker-0.2.3 → bookdatamaker-0.2.4/src/bookdatamaker.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: bookdatamaker
-Version: 0.2.3
+Version: 0.2.4
 Summary: CLI tool for extracting text with DeepSeek OCR and generating datasets
 Author-email: zwh20081 <zwh20081@solart.pro>
 License: MIT
@@ -89,6 +89,49 @@ A powerful CLI tool for extracting text from documents using DeepSeek OCR and ge
 - 🌐 **Flexible Modes**: API or self-hosted for both stages
 - 📈 **Progress Tracking**: Real-time progress bars
 
+## Installation
+
+### From PyPI (Recommended)
+
+```bash
+pip install bookdatamaker
+```
+
+### From Source
+
+```bash
+git clone https://github.com/yourusername/bookdatamaker.git
+cd bookdatamaker
+pip install -r requirements.txt
+pip install -e .
+```
+
+### Optional: Local Inference Support
+
+```bash
+# For self-hosted OCR and LLM generation
+pip install bookdatamaker[local]  # From PyPI
+# OR
+pip install -e ".[local]"  # From source - installs transformers==4.46.3, torch, flash-attn, etc.
+```
+
+**Note**: The project requires `transformers==4.46.3` for optimal compatibility with DeepSeek-OCR. A warning will be displayed if a different version is detected.
+
+### System Requirements
+
+**For API Mode:**
+- Python 3.10+
+- API keys (OpenAI, DeepSeek, etc.)
+
+**For Local Mode:**
+- Python 3.10-3.12 (3.13 not supported due to vLLM compatibility)
+- NVIDIA GPU with CUDA support (or CPU, though slower)
+- 16GB+ VRAM recommended for GPU
+- transformers==4.46.3
+- Linux or WSL2 (recommended)
+
+---
+
 ## Quick Start
 
 ### Prerequisites
@@ -103,7 +146,7 @@ export DEEPSEEK_API_KEY=your_deepseek_key    # For API OCR mode
 
 ```bash
 # 1. Install
-pip install -r requirements.txt && pip install -e .
+pip install bookdatamaker
 
 # 2. Extract → Generate → Export
 bookdatamaker extract book.pdf -o ./extracted
@@ -115,7 +158,7 @@ bookdatamaker export-dataset dataset.db -o output.parquet
 
 ```bash
 # 1. Install with local dependencies
-pip install -r requirements.txt && pip install -e ".[local]"
+pip install bookdatamaker[local]
 
 # 2. Extract with local OCR
 bookdatamaker extract book.pdf --mode local --batch-size 8 -o ./extracted
@@ -131,27 +174,9 @@ bookdatamaker generate ./extracted \
 bookdatamaker export-dataset dataset.db -o output.parquet
 ```
 
-## Installation
-
-### Basic Installation
-
-```bash
-git clone https://github.com/yourusername/bookdatamaker.git
-cd bookdatamaker
-pip install -r requirements.txt
-pip install -e .
-```
-
-### Optional: Local Inference Support
-
-```bash
-# For self-hosted OCR and LLM generation
-pip install -e ".[local]"  # Installs transformers==4.46.3, torch, flash-attn, etc.
-```
-
-**Note**: The project requires `transformers==4.46.3` for optimal compatibility with DeepSeek-OCR. A warning will be displayed if a different version is detected.
+---
 
-### System Requirements
+## System Requirements
 
 **For API Mode:**
 - Python 3.10+