notes.txt

Twisted docs: http://twistedmatrix.com/trac/

Concept: consume a directory (or directories?) and return JSON at a few URL patterns
    root = titles I know about
    /name = issues of a given title
    /name/issue id = json dump of {page & image urls}
    /name/issue id/page = image url

TODO
    document functions and clean up pathing logic
    trap case where you restart server and temp directory has gone away but you're looking at issue file list again

Issues
    file name formatting is incredibly inconsistent
        ignore all the shit in parentheses
        if all else fails, lose ALL of the alpha characters and just get the numbers, fall back to parent directory name
    issue numbering is something of a mess
        once you figure out file name concept, strip out leading 0s to prevent 010 being in front of 02
    can I at least trust directory names to be meaningful and correct?
    be nice to process folders, files and contents via a generator, so there isn't a huge start up issue
    bonus points: watch directory for changes, update
    serve files
        https://github.com/yosinski/simple-twisted-image-server/blob/master/server.py
        http://stackoverflow.com/questions/1538617/http-download-very-big-file
        http://codereview.stackexchange.com/questions/2315/file-serving-twisted-resource
        http://stackoverflow.com/questions/4110579/twisted-web-how-do-i-properly-images-from-an-images-directory

RAR Files
    rarfile docs: http://rarfile.berlios.de/doc/
    http://stackoverflow.com/questions/1185959/read-content-of-rar-file-into-memory-in-python
    having trouble figuring out how to extract the files locally

Configuration file
    port number
    root directory

Server
    big Twisted tutorial at http://krondo.com/blog/?page_id=1327
    learn how to build a proper server and break out code
    threads - http://stackoverflow.com/questions/5904734/twisted-deferred-vs-blocking-in-web-services
    add ability to shut the thing down
    also make sure to close connections properly
        need to read http://blackjml.livejournal.com/23029.html
    some more info at http://zenmachine.wordpress.com/web-services-and-twisted/
        would be good to be caching results after startup

Misc
    Why does server find hundreds fewer files on Mac looking at the same directory?