Metadata-Version: 2.1
Name: DAnki
Version: 1.0.1
Summary: Automate deck creation for Anki to learn german
Home-page: https://github.com/dileivas/DAnki
Author: Diego Leivas
Author-email: dileivas@hotmail.com
License: MIT
Description: # DAnki: Automate deck creation for Anki to learn german
        
        `DAnki` automate the work of creating decks for Anki to learn german. From the export of your Kindle's notes
        or a csv file, DAnki can generate cards with the translation of the word for 8 different languages including
        the german pronunciation sound.
        
        *This library and its author(s) are not affiliated/associated with the main Anki project in any way.*
        
        It was created for my german students to develop their vocabular. Using Kindle you can just highlight the
        german words or expressions you want to practice, export these notes to your email and use the csv file
        with this program. It also create tags for Anki with the page number.
        
        *Note:* Your Kindle's language must be in *portuguese* or *english* .
        
        ## How to use
        
        Here is an exemple using a csv file from kindle.
        
        ```python
        language = 'en'
        csv_filepath = r'csv_exemple_files/sherlock_highlights_from_kindle_english.csv'
        deck_name = 'sherlock_en'
        from_kindle = True
        
        myDanki = DAnki(language, csv_filepath, deck_name, from_kindle)
        myDanki.create_translated_deck()
        ```
        
        Here is an exemple using a csv file with only 2 columns: German Word and Tag.
        
        ```python
        language = 'ch'
        csv_filepath = r'csv_exemple_files/not_from_kindle.csv'
        deck_name = 'not_from_kindle'
        from_kindle = False
        
        myDanki = DAnki(language, csv_filepath, deck_name, from_kindle)
        myDanki.create_translated_deck()
        ```
        
        *You can find the csv exemple files [HERE](https://github.com/dileivas/DAnki/tree/master/csv_exemple_files).*
        
        ## Available languages
        
        DAnki uses [leo.org](leo.org) to find a translation from german, so the available languages are:
        
        ('pt' - portuguese) ('en' - english) ('fr' - french) ('es' - spanich) 
        ('it' - italian) ('ch' - chinese) ('ru' - russian) ('pl' - polish)
        
        ## Must
        
        You must add 'de_DE_frami.dic' and 'de_DE_frami.aff' files in 
        `C:\Users\user\AppData\Local\Programs\Python\Python38\lib\site-packages\enchant\data\mingw64\share\enchant\hunspell`
        
        Dictionary link:  https://extensions.openoffice.org/en/project/german-de-de-frami-dictionaries
        
        For more information: https://pyenchant.github.io/pyenchant/tutorial.html
        
        ## Credits
        
        My thanks to:
        
        [genanki](https://github.com/kerrickstaley/genanki)
        
        [HanTa](https://github.com/wartaal/HanTa) - I don't use this library, but allowed me to understand lemmatization.
        
        [Vorverarbeitung von Texten mit Python und NLTK](http://textmining.wp.hs-hannover.de/Preprocessing.html)
        
        [treetaggerwrapper](https://treetaggerwrapper.readthedocs.io/en/)
        
        [gTTS](https://github.com/pndurette/gTTS)
        
        [LEO GmbH](leo.org)
        
Keywords: anki,flashcards,memorization,german,deutsch
Platform: UNKNOWN
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.8
Description-Content-Type: text/markdown
