Skip to main content

Manuscript

Project description

Installation

Для пользователей

pip install manuscript-ocr

Для разработчиков

pip install -r requirements-dev.txt

Примечание: requirements-dev.txt включает GPU версию PyTorch, инструменты тестирования, форматирования и сборки.

GPU поддержка

Если вы пользователь и хотите GPU поддержку:

pip install manuscript-ocr
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu118 --force-reinstall

Проверка GPU:

import torch
print(f"CUDA доступна: {torch.cuda.is_available()}")

Usage Example

from PIL import Image
from manuscript.detectors import EASTInfer

# Инициализация
det = EASTInfer(score_thresh=0.9)

# Инфер с визуализацией
page, vis_image = det.infer(r"example\ocr_example_image.jpg", vis=True)

print(page)

# Покажет картинку с наложенными боксами
Image.fromarray(vis_image).show()

# Или сохранить результат на диск:
Image.fromarray(vis_image).save(r"example\ocr_example_image_infer.png")

Результат

Текстовые блоки будут выведены в консоль, например:

Page(blocks=[Block(words=[Word(polygon=[(874.1005, 909.1005), (966.8995, 909.1005), (966.8995, 956.8995), (874.1005, 956.8995)]),
                          Word(polygon=[(849.1234, 810.5678), … ])])])

А визуализация сохранится в файл example/ocr_example_image_infer.png:

OCR Inference Result

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

manuscript_ocr-0.1.8.tar.gz (60.5 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

manuscript_ocr-0.1.8-py3-none-any.whl (66.6 kB view details)

Uploaded Python 3

File details

Details for the file manuscript_ocr-0.1.8.tar.gz.

File metadata

  • Download URL: manuscript_ocr-0.1.8.tar.gz
  • Upload date:
  • Size: 60.5 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.11.8

File hashes

Hashes for manuscript_ocr-0.1.8.tar.gz
Algorithm Hash digest
SHA256 8fb1859dff3814156c8a4f730a2f31ff360e315564065a9b0d3b62d98073049f
MD5 8f7c50d286f71dfe19dc11cec81f0d48
BLAKE2b-256 556bfb9707462afcd5ef149ec2638801d7af8eb804b5aeafdd80f29ec22fc693

See more details on using hashes here.

File details

Details for the file manuscript_ocr-0.1.8-py3-none-any.whl.

File metadata

  • Download URL: manuscript_ocr-0.1.8-py3-none-any.whl
  • Upload date:
  • Size: 66.6 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.1.0 CPython/3.11.8

File hashes

Hashes for manuscript_ocr-0.1.8-py3-none-any.whl
Algorithm Hash digest
SHA256 2b5d31aa89efdcf599f1aa3f49a4800cdfde0e6fd5e349e487874fe6fceb7c18
MD5 6311107c7fc5a3b049099558f9bc299a
BLAKE2b-256 892a30b1a909728683f5fbfe390d19d68ea4b0db3b052be4b85f7dcf6c48fd53

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page