rbtr/pachinko

modular pluggable media sorter

pachinko's Issues

bad sort on apostrophes

e.g.
Tom Clancy's Jack Ryan -> s Jack Ryan

Generated config sometimes unusable

The config generated by pachinko config is not always usable. This is because the order of the processors list is non-deterministic and potentially reordered each time it is written.

The particular issue I ran into is that the video-id processor must be executed before the tv-id processor (because of the check in video.go#L42. If the processors are out of order, it will not work.

Example bad config created by the command:

dry-run: false
inputs:
- name: path
  src-dir: ./src
log-format: text
log-level: info
outputs:
- create-dirs: true
  dry-run: true
  name: move
  overwrite: false
- name: stdout
pipeline:
  buffer: 0
processors:
- api-key: ""
  name: tvdb
  request-limit: 10
- dest-dir: dest
  episode-names: false
  format: not-implemented
  name: tv-dest
  season-dirs: true
  tv-prefix: tv
- matchers:
  - (?i)(?:s+\d+)(?:\.|\s|-|_|x)*(e+\d+)
  - (?i)(?:season)+(?:(?:\.|\s|-|_|x|\/)\d+)
  - (?i)(?:[^[:alpha:]\d])\d{1,3}[x-]\d{1,3}
  name: tv-id
- matchers:
  - (?i)\b([\s\w.-]*)[\s.\/-]+(?:\((\d+)\))?[\s.\/-]?(\d{1,3})[x-](\d{1,3})
  - (?i)\b([\s\w.-]*?)?(?:[\s\(.\/-](\d{4})[\s\).\/-])?[\s\w.-]?(?:s+(\d+))(?:\.|\s|-|_|x)*(?:e+(\d+))
  - (?i)\b([\s\w.-]*)[\s.\/-]+(?:\((\d+)\))?[\s.\/-]?(?:season|series).?(\d+).?(?:episode)?[^\d(]?(\d+)
  name: tv-meta
  sanitize-name: true
- file-types:
  - .avi
  - .divx
  - .m4v
  - .mkv
  - .mov
  - .mp4
  - .xvid
  name: video-id

[plugin] add TMDb plugin

themoviedb identifier
themoviedb decorator

[core] enable pprof

provide a way to use pprof

extracting metadata from filename should start with the file and proceed up the tree, not the reverse

for example,
/src/My Cool Show Season 1-2-3-4-5-6-7/My Cool Show S01E01.mkv
will be incorrectly matched as "My Cool Show Season 1-2-3-4-5" Season 6 Episode 7, since the path parser extracts metadata from the entire path of the file, lazily, left to right.
it would be better to parse the leaf filename initially, then proceed up the dir tree if the full metadata is not present in just the filename. the extraction from the filename should be prioritized over subsequent extractions from the containing dirs, if in conflict.

Config command fails if ~/.pachinko.yaml exists

Found this one when trying to write up issue #6. The following error happens if the ~/.pachinko.yaml file exists and you then run pachinko config.

FATA[0000] 8 error(s) decoding:

* 'inputs[0]' expected type 'string', got unconvertible type 'map[interface {}]interface {}'
* 'outputs[0]' expected type 'string', got unconvertible type 'map[interface {}]interface {}'
* 'outputs[1]' expected type 'string', got unconvertible type 'map[interface {}]interface {}'
* 'processors[0]' expected type 'string', got unconvertible type 'map[interface {}]interface {}'
* 'processors[1]' expected type 'string', got unconvertible type 'map[interface {}]interface {}'
* 'processors[2]' expected type 'string', got unconvertible type 'map[interface {}]interface {}'
* 'processors[3]' expected type 'string', got unconvertible type 'map[interface {}]interface {}'
* 'processors[4]' expected type 'string', got unconvertible type 'map[interface {}]interface {}'

bad sort on colons

eg 3:10 to Yuma -> To Yuma

[pipeline] add prometheus

add prometheus support
allow plugins to register metrics

[plugin] add trakt plugin

plugin ideas:
trakt identifier - trakt version of the tvdb plugin for adding metadata
trakt collector - sink/output which "adds to collection" on trakt

[plugin] add file deletion plugin

something to clean up the extra trash - text files, images, subs - in the src directories
bonus delete the empty src directories when sorted