cache_file.py

import os, time, shutil, heapq

MAX_CACHE_MB = 200
_CACHED_FILES = []
_CACHED_SIZE = None  # in bytes

def cache_file(filename):
    '''
    Automatically maintain a cached version of the given filename.

    It assumes that there is a RAM-disk cache at ~/ramdisk, and puts the
    cached files as the same path as in the original file system.  The
    cache size is limited (configurable) and it automatically removes
    oldest one if the size limit is exceeded like LRU.
    '''
    global _CACHED_FILES, _CACHED_SIZE
    assert os.path.isfile(filename)
    homedir = os.environ['HOME']
    ramdir = os.path.join(homedir, 'ramdisk')
    if not os.path.isdir(ramdir):
        return filename
    cachedir = os.path.join(ramdir, 'myscriptname')

    if _CACHED_SIZE is None:
        # update the current cache size when first called.
        _CACHED_SIZE = 0
        for root, dirs, files in os.walk(cachedir):
            for name in files:
                filepath = os.path.join(root, name)
                _CACHED_SIZE += os.path.getsize(filepath)
                heapq.heappush(_CACHED_FILES, (os.path.getmtime(filepath), filepath))
        print('cache_file: Initial status: {0:.3f} MB'.format(_CACHED_SIZE / 1024 / 1024))

    # NOTE: we need different path split mechanism for Windows.
    filename = os.path.abspath(filename)
    origdir, origname = os.path.split(filename)
    targetdir = os.path.join(cachedir, origdir[1:])
    cached_filename = os.path.join(targetdir, origname)
    os.makedirs(targetdir, exist_ok=True)
    # if the cached file does not exist or outdated, update it.
    if not (os.path.exists(cached_filename) and
            os.path.getmtime(cached_filename) >= os.path.getmtime(filename)):
        add_size = os.path.getsize(filename)
        timestamp = time.clock_gettime(time.CLOCK_REALTIME)
        while _CACHED_SIZE + add_size > MAX_CACHE_MB * 1024 * 1024:
            try:
                _, old_filename = heapq.heappop(_CACHED_FILES)
                _CACHED_SIZE -= os.path.getsize(old_filename)
                os.unlink(old_filename)
            except IndexError:
                break
        heapq.heappush(_CACHED_FILES, (timestamp, cached_filename))
        shutil.copyfile(filename, cached_filename)
        _CACHED_SIZE += add_size
    return cached_filename
    
# Example usage
filename = cache_file(filename)