#cache #hash #video #storing #companion #path #vid-dup-finder

video_hash_filesystem_cache

A companion cache for storing the video hashes used by the vid_dup_finder_lib crate

1 unstable release

0.1.0 Oct 30, 2021

#22 in #companion

25 downloads per month
Used in vid_dup_finder

MIT/Apache

27KB
334 lines

video_hash_filesystem_cache

A cache for storing the video hashes used by the vid_dup_finder crate


lib.rs:

A utility crate with a cache for vid_dup_finder_lib::VideoHash. This crate defines struct [VideoHashFilesystemCache], which caches hashes of videos to disk.

The cache stores the last modified time of the underlying files and will automatically update when this changes.

Example (with individual files)

use video_hash_filesystem_cache::*;
use vid_dup_finder_lib::*;

// Create a cache on disk which will save itself to disk after every 100 changes
let cache = VideoHashFilesystemCache::new(100, cache_file_path).expect("failed to create cache");

// Now create a video hash by calling get_update on the cache.
let video_hash : VideoHash = match cache.fetch_update(&vid_file_path) {
   Ok(Some(Ok(hash)))  => hash,     // A hash was successfully created/fetched
   Ok(None)            => panic!(), // None is returned when vid_file_path is removed from the filesystem
   Ok(Some(Err(_e)))   => panic!(), // Ok(Some(Err())) is returned when an error occurs while creating a VideoHash
   Err(cache_error)    => panic!(), //"All other Io errors")
};

// Subsequent calls will fetch the hash from the cache instead of creating it from the filesystem.

// The cache must be saved to disk at the end of execution,
// otherwise changes since the last save will be lost.
cache.save().unwrap()

Caching many videos at once

Struct crate::FileProjection is used for updating many files at once. when created with a set of starting paths, it can be passed to the cache to update all child files of those starting paths.

Example (caching an entire directory)

use video_hash_filesystem_cache::*;
use vid_dup_finder_lib::*;

// Create a cache on disk which will save itself to disk after every 100 changes
let cache = VideoHashFilesystemCache::new(100, cache_file_path).expect("failed to create cache");

// Create the projection representing two directories of video files.
// the second argument is a list of directories/paths to be ignored
let mut projection = FileProjection::new(&video_dirs, &excl_dirs, &excl_exts).unwrap();
let project_errs = projection.project_using_fs().unwrap();

// Update the cache using the projection. a list of individual loading errors will be returned.
let cache_update_errs = cache.update_using_fs(&projection).unwrap();

// Now all videos under videos_dir_1 and videos_dir_2 will be cached.
// They can be retrieved from the cache without touching the filesystem using
// VideoHashFilesystemCache::fetch
let video_hash : VideoHash = cache.fetch(&vid_file_path).unwrap();

// The cache must be saved to disk at the end of execution,
// otherwise changes since the last save will be lost.
cache.save().unwrap()

Dependencies

~23MB
~203K SLoC