uranusseven / qing_bureau_of_construction Goto Github PK

This project employs Optical Character Recognition (OCR) to digitize historical records from the Qing manufacturing office.

License: Apache License 2.0

Python 100.00%

qing_bureau_of_construction's Introduction

qing_bureau_of_construction's People

Contributors

Stargazers

Watchers

qing_bureau_of_construction's Issues

ENH: optimize advanced searching

Instead of using whoosh's syntax, use Google-like syntax. By doing that, we can easily highlight the keywords on the image.

ENH: format the outputs using line ID

ENH: handle request failure

When request failed, instead of exiting, we should keep trying.

ENH: color the text according to the probability

ENH: a better UI using streamlit

Pdf view bug in Windows & Chrome

Thanks for the sharing. I have added plugin Enable local file links to Chrome (Windows), but no pdf hyperlink appears. I want to identify the year information of the source. BTW, what is the version of your archives?

Two missing requirements in Windows. And an UTF-8 code change.

In Windows, you should add plotly to requirements.txt.
You have to specify altair to version 4 or 4.x.x

Line 88 in build_index.py have to be changed to
with open(f"{OCR_RESULTS_DIR}/{f}", 'r', encoding='utf-8') as fd:
to avoid encoding problem.

DOC: update doc

how to install streamlit
how to install enable local links chrome extension
how to start app

ENH: open PDF file instead of an image

/Applications/Google\ Chrome.app/Contents/MacOS/Google\ Chrome "file://$(path_to_pdf_file)#page=100"