# Constance Ring by Amalie Skram

This is a digital, free version of
[Constance Ring (English Wikipedia)](https://en.wikipedia.org/wiki/Constance_Ring),
[Constance Ring (Norwegian Wikipedia)](https://no.wikipedia.org/wiki/Constance_Ring)
by [Amalie Skram (English Wikipedia)](https://en.wikipedia.org/wiki/Amalie_Skram),
[Amalie Skram (Norwegian Wikipedia)](https://no.wikipedia.org/wiki/Amalie_Skram).
It is based of a scan of the book published in 1943 publication by Gyldendal,
[available at the Norwegian National Library](https://urn.nb.no/URN:NBN:no-nb_digibok_2013112808046).

The book is available in the [public domain license](https://www.nb.no/tilgang/lisens/#public-domain).
This license allows republication of the material, even for commercial purposed.

However, *this* edition of the book is available under the
[CC BY-SA 4.0 license](https://creativecommons.org/licenses/by-sa/4.0/) allowing
for commercial reproduction of the book as long as it is attributed back to this
book. See [LICENSE.md](./LICENSE.md) for details.

# Conversion

This book is created by downloading the PDF with all the pages from nb.no,
extracting the images and then performing OCR on them. The default OCR from the
national library wasn't very good.

The text files are then slightly massaged by removing headers and footers,
making chapters headers into Markdown headers and de-hyphentating the text.

The Epub is then generated by Pandoc.

# Proofreading

Run `make spellcheck-words` to run hunspell over the document. It will read
[`dict`](./dict) as the personal dictionary and output any misspelled words into
[`spellcheck-words`](./spellcheck-words).

The book is written in a much older Norwegian that what the dictionary contains
so there are many misidentifications.

Process:

For each word in `spellcheck-words`:

* Check if the word is in the original PDF. If so, add it to the dictionary.
  This happends when the word is just too old.
* Check with manual spellcheck with `make spellcheck`. This is more tedious as
  it runs through the entire document from the beginning.

# Hunspell

Hunspell has affix (.aff) files and dictionary (.dic) files. The documentation
claims that it should be possible to load single `.dic` files as "special
dictionaries", however I can't get that to work.

However, the base Norwegian affix file can be used as the affix file for the
spellchecker by creating a symlink, problem solved.