A dossier web parser for http://oldcetatenie.just.ro.
Based on the common history of Romania and the Republic of Moldova, the citizens of the Republic of Moldova can apply for, and obtain Romanian citizenship. Annually, circa 80.000-100.000 people apply for Romanian citizenship.
Documents examination can take from half a year to 2-3 years and sometimes even more. People can check the status of their application, but the process is really tedious and error-prone.
The orders are published on http://oldcetatenie.just.ro/index.php/ro/ordine/articol-11 as PDF files. A person must access this web address, and check every order for his dossier number.
This tool tends to make the process easier, by automating the whole process.
The project makes use of several libraries, some of which require some manual setup.
The pdf2image library requires poppler to be installed, according to its README file.
You can install it using homebrew: brew install poppler.
The pytesseract library requires tesseract to be installed, according to its README file.
You can install it using homebrew: brew install tesseract.
To install the project dependencies, you can use pip and the requirements.txt file:
pip install -r requirements.txt
In the main.py search for the lookups. Set your dossier number according to the example and run the project.
For tesseract to properly read the PDF images, I've configured it's page segmentation mode to 4 (single column of text of variable sizes).
You can read more here: https://ai-facets.org/tesseract-ocr-best-practices.
The default storage path for the tools is set to $HOME/dev/romanian-citizenship-status-files.
You can change it by changing the FILES_STORAGE_PATH constant.
Sometimes the PDF conversion to TXT results in badly formatted data. Unfortunately, I don't have much experience with tesseract and couldn't configure it better.
Solutions: Manually convert the malformatted files using https://pdftotext.com/ or other tools.
I've also included all the PDF and TXT files until November 2020.
The tool can download and convert them by itself, but it might take a very long time, as there are more than 2000 files until now. This will help a lot.