joeboy / scenewriter Goto Github PK

1.0 1.0 0.0 3.5 MB

Fountain screenplay format parser in Rust / Nom

License: MIT License

Rust 100.00%

scenewriter's Introduction

Scenewriter - a Fountain Parser

Scenewriter is a command line program that parses files in Fountain markup language, and generates PDFs in the standard screenplay format. It doesn't implement the full Fountain spec yet, and probably has bugs. But it seems to kinda mostly work, to the limited extent I've tested it so far.

If you just want to download and use it, see the releases. Note that it's still an unfinished / preview release. But for basic scenes, it might already do everything you need. Issues that might be offputting are:

Full Fountain spec is not yet implemented
Only really tested with straightforward UTF8 / ASCII encoded documents, probably won't handle weirder encodings well
Title page is very basic - eg. no contact details

Here's direct links to the Windows, Mac and Linux versions.

Here's the reference version of the John August's Big Fish screenplay from fountain.io, and here's Scenewriter's effort at formatting the same fountain file. It's in the ballpark, but not 100% there yet. For the even less polished HTML output see here.

On my laptop, HTML generation is basically instant (~25ms for Big Fish). Generating the 119 page PDF from the same file takes about 0.4s.

If you'd especially like anything implemented or fixed please raise an issue. See the issues to get a vague idea of the current priorities.

Usage

./scenewriter --help

Usage: scenewriter [..options..] input_filename

Options:

             --pdf     Write PDF file (default)
            --html     Write HTML file
           --stats     Show screenplay stats

              --a4     A4 page size (default)
                -a

          --letter     US Letter page size
                -l

 --output filename     Choose output filename (default is the input
       -o filename     filename but with .pdf or .html extension)

            --help     Show this help


     Eg: scenewriter --a4 --pdf -o "My Screenplay final-v23.pdf" my_screenplay.fountain

or just: scenewriter my_screenplay.fountain

(if you're using the Windows version it'll be scenewriter.exe).

Python bindings

If you want to call scenewriter code from python, there are very rudimentary bindings here.

scenewriter's People

Contributors

Stargazers

Watchers

scenewriter's Issues

Centred text

Support "forced" element interpretations

eg. leading "!" should mean the text is interpreted as an Action, leading "." for forced Scene Heading etc.

Builds / releases

Should do automated builds / releases, at least for Windows and Mac. I guess with Rust's static binaries it should be easy for Linux too? Will have to check that out.

Improve title page

Support more stuff.

Specify page size

Maybe just provide US Letter and A4 options.

Switch to Courier Prime

Seems to be the most liked font for screenplays (there's also Courier Screenplay but Courier Prime seems to be the frontrunner?)

Check behaviour with weird encodings / unsupported chars

Support HTML output

Should be pretty easy to do, and it'll be easier to test HTML output than PDF.

Redo PDF generation

I think at some point I might want to redo PDF generation, cutting out genpdf and probably printpdf. I think, given the simple monospaced nature of the output, it should be a massive deal to work at a lower level (ie. lopdf). That'd mean I could cut out the genpdf / printpdf forks I introduced for underlining, could do smarter page breaks (widows / orphans). I guess not being based on forked packages might also make it more publishable as a crate.

Full(er) support for the fountain spec

Support text formatting within elements

Ie. *italics*, **bold**, probably other things if I bothered to check the spec.

I think this should just be a matter of passing the parsed element texts through a simple formatter.

Parentheticals

Stats

Smarter page breaks

At the moment it's entirely dumb about where it puts page breaks, it'll happily put the speaker name on page 12 and their dialogue on page 13. Should look into that.

Slim down PDF output

At present Farce produces PDFs that are >400k for a "Hello World" fountain file. I suspect almost all the bloat is the embedded font. Would be good if there was some way to exclude unused glyphs or something. Could also provide an option to use the default PDF Courier font.

Fix multiple bracket interpretation in Character lines

Eg.

Dave (V.O.) (CONT'D)

results in weird buggy output. Investigate and fix. Either allow multiple "dialogue extension"s, or maybe just make the whole thing a single dialogue extension?

Support underlining

It turns out underlining isn't exactly supported in the PDF format, you have to "manually" draw a line in the right place (or add an annotation, which doesn't look any easier). This seems annoyingly hard to do, probably needs to be supported upstream in genpdf or printpdf.

Tests

Given that there's a fair few fountain parsers in the wild, I think it'd be good to work on some "portable" tests. A simple way would be to have some hooks that do things like invoke fountain -> html conversion, and compare the results to a reference output. A complication would be that different tools will (legitimately) generate different looking html, so that'll either require some fiddling (more hooks?) or the specific html output will be part of the tests. Hmm.