diff options
Diffstat (limited to 'doc')
| -rw-r--r-- | doc/filmoteca.ms | 306 | ||||
| -rw-r--r-- | doc/filmoteca.pdf | bin | 0 -> 11310 bytes | |||
| -rw-r--r-- | doc/mkfile | 15 |
3 files changed, 321 insertions, 0 deletions
diff --git a/doc/filmoteca.ms b/doc/filmoteca.ms new file mode 100644 index 0000000..06b0d27 --- /dev/null +++ b/doc/filmoteca.ms @@ -0,0 +1,306 @@ +.nr PS 12 +.nr VS 14 +.TL +Filmoteca +.AU +Rodrigo G. López +.sp +rgl@antares-labs.eu +.AI +Antares Telecom Laboratories +Albatera, Alicante +.AB +.DA +This document is a work in progress and is subject to change by the +author at any time. It shows the ideas that are being implemented in +a 3TB multimedia dataset distributed among two hard drives, for the +author's personal consumption. +.AE +.SH +Introduction +.LP +The filmoteca is an attempt at organizing motion picture data like +movies, tv shows and documentaries in a formalised file hierarchy that +is easy to use and maintain. We take influence from the Plan 9 +operating system and its file servers to design an interface that +makes it simple to script any operation using standard Unix tools. +.NH 1 +The Tree +.PP +The file hierarchy is structured with the first-level directories +named after the titles they keep. Any collision is resolved by +appending a space and a roman numeral between parentheses to the +title, similar to how IMDb does it [IMDb], except we favor historical +correctness over time of insertion into the database. This means that +if we add Rush (2012) first and later we find about Rush (1991), we +will add it under the title +.I "Rush (I)" +and the latest one will become +.I "Rush (II)" . +The inner levels are thoroughly explained in Section 2. +.NH 2 +Movies +.P1 +.../filmoteca/ + title/ + title/release + title/synopsis + title/cover + title/video + title/history + title/sub/ + title/sub/en + title/sub/es + title/sub/... + title/dub/ + title/dub/es/ + title/dub/es/video + title/dub/... + title/extra/ + title/extra/Opening.mp4 + title/extra/Ending.mp4 + title/extra/Deleted Scenes.mp4 + title/extra/... + title/remake/ + title/remake/1984/ + title/remake/1984/video + title/remake/1984/... + title/remake/y... +.P2 +.NH 2 +Multipart Movies +.P1 + title/ + title/release + title/synopsis + title/cover + title/video1 + title/video2 + title/videon... + title/history + title/sub1/ + title/sub1/en + title/sub1/... + title/subn... + title/dub1/ + title/dub1/es/ + title/dub1/es/video + title/dub1/... + title/dubn... + title/extra/... + title/remake/y... +.P2 +.NH 2 +Series +.P1 + title/ + title/release + title/synopsis + title/cover + title/history + title/s/ + title/s/1/ + title/s/1/1/ + title/s/1/1/video + title/s/1/1/sub/ + title/s/1/1/sub/en + title/s/1/1/sub/... + title/s/1/1/dub/ + title/s/1/1/dub/es/ + title/s/1/1/dub/es/video + title/s/1/1/dub/... + title/s/1/n... + title/s/n... + title/extra/... + title/remake/y... +.P2 +.NH 1 +Walking Down the Tree +.NH 2 +The +.CW release +file +.PP +The +.I release +file stores the date on which the work was first released. In the +case of a movie (including multipart) there is only one line, whereas +in a series there is one line per season. If there is some season +lacking between two other seasons, an empty string is used instead. +The date string has two different formats: +.CW yyyy +and +.CW ddmmmyyyy , +e.g. +.CW 2016 , +.CW 1932 , +.CW 23mar1997 , +.CW 1jun1978 , +etc. +.sp +I am going to show you two examples, to clarify the structure of a hollowed +file. First we have +.I "Black Mirror" , +with two seasons, the first and the third: +.P1 +$ cat 'Black Mirror'/release +2011 + +2016 +$ xd -c 'Black Mirror'/release +0000000 2 0 1 1 \en \en 2 0 1 6 \en +000000b +.P2 +Then a extreme case, +.I "The X Files" , +where there is only one season, the tenth: +.P1 +$ cat 'The X Files'/release + + + + + + + + + +2016 +$ xd -c 'The X Files'/release +0000000 \en \en \en \en \en \en \en \en \en 2 0 1 6 \en +000000e +.P2 +You can then use the following script to figure out the folders in +.I s +(explained in Section 2.8): +.P1 +$ <'Black Mirror'/release awk ' + BEGIN{ s = 0; } + /^$/ { s++; } + /^[0-9]+$/ { print ++s } + ' +1 +3 +.P2 +.NH 2 +The +.CW synopsis +file +.PP +Contains a brief summary of the plot. +.NH 2 +The +.CW cover +file* +.PP +The +.I cover +file is the poster of the movie or tv show encoded in well-known image +formats such as JPEG, PNG, TIFF or GIF. +.FS +* This file will be renamed to +.I poster +in the future. +.FE +.NH 2 +The +.CW video[n] +file +.PP +The +.I video +file is the fruit of the tree, it stores the movie or a series episode +under some multimedia container format, the most common being MP4, +Matroska (MKV) and AVI. +.NH 2 +The +.CW history +file +.PP +In an attempt to educate the user, a +.I history +file is provided containing an explanation, facts and references for +movies that claim to be based on/inspired by true events. You might +be surprised. +.NH 2 +The +.CW sub[n] +folder +.PP +Files within +.I sub +are SubRip (SRT) subtitle files, each named using a two-letter +country code* for the language they provide. +.FS +* This will be changed to use a reasonable subset of the +two/three-letter IANA language subtag registry [IANALang]. +.FE +.NH 2 +The +.CW dub[n] +folder +.PP +The +.I dub +folder stores revoicings and provides one folder per language, using +the same conventions as the subtitles. Inside each of these, there is +a video file (see Section 2.4)†. +.FS +† Although under best conditions it should be just an audio file one +could swap into the stream. This is an idea that is going to affect +the video file as well, since one could compose their own media +streams into a single experience, for example grabbing the original +video, adding japanese audio and english subtitles; a very common +setup for watching anime. +.FE +.NH 2 +The +.CW s +folder‡ +.PP +.I S +is a three-level directory containing the episodes of a series. The +first level are seasons by number, if there is some season that is +lacking, we completely avoid it (signaled by the +.I release +file, Section 2.1), there are no empty folders. The second level are +episodes following the same naming rules. The structure of an episode +folder is similar to that of a movie, except it only has a +.I video , +and the +.I sub +and +.I dub +folders. +.FS +‡ It really is a leftover from a previous design, since it means +``season'', but there is no +.I e +directory containing ``episodes''. +.FE +.NH 2 +The +.CW extra +folder +.PP +The +.I extra +folder holds a bunch of random files that make up the featurettes and +other behind-the-scenes content, such as the making-of, auditions, the +director's commentary, additional posters, OSTs, etc. +.NH 2 +The +.CW remake +folder +.PP +.I Remake +is a two-level directory, with the first level holding directories +named after the year the remake was released on, and inside each of these, the +exact same file structure you would find in a movie (see Section 1.1). +.bp +.SH +References +.LP +[IMDb] https://help.imdb.com/article/imdb/discover-watch/what-do-the-roman-numerals-like-i-and-ii-after-people-s-names-mean/GA827M8GK5KVH8TC?ref_=helpart_nav_33#, last checked March 19, 2019 +.br +[IANALang] https://www.iana.org/assignments/language-subtag-registry/language-subtag-registry diff --git a/doc/filmoteca.pdf b/doc/filmoteca.pdf Binary files differnew file mode 100644 index 0000000..37dc4bd --- /dev/null +++ b/doc/filmoteca.pdf diff --git a/doc/mkfile b/doc/mkfile new file mode 100644 index 0000000..3a61b6c --- /dev/null +++ b/doc/mkfile @@ -0,0 +1,15 @@ +FONTS='.FP times' +DOCNAME=filmoteca + +all:VQ: $DOCNAME.ps $DOCNAME.pdf + +clean:VQ: + rm -f $DOCNAME.ps $DOCNAME.pdf + +$DOCNAME.ps:V: $DOCNAME.ms + {echo $FONTS; cat $prereq}> _$prereq + eval `{doctype _$prereq} | lp -dstdout > $target && rm -f _$prereq + +$DOCNAME.pdf:V: $DOCNAME.ps + cat /sys/doc/docfonts $prereq > _$prereq + ps2pdf _$prereq $target && rm -f _$prereq |