traxxx/src/store-releases.js

474 lines
16 KiB
JavaScript

'use strict';
const config = require('config');
const Promise = require('bluebird');
const argv = require('./argv');
const logger = require('./logger')(__filename);
const knex = require('./knex');
const slugify = require('./utils/slugify');
const bulkInsert = require('./utils/bulk-insert');
const resolvePlace = require('./utils/resolve-place');
const chunk = require('./utils/chunk');
const { formatDate } = require('./utils/qu');
const { associateActors, associateDirectors, scrapeActors, toBaseActors } = require('./actors');
const { associateReleaseTags } = require('./tags');
const { curateEntity } = require('./entities');
const { associateReleaseMedia } = require('./media');
const { notify } = require('./alerts');
async function curateReleaseEntry(release, batchId, existingRelease, type = 'scene') {
const slugBase = release.title
|| (release.actors?.length && `${release.entity.slug} ${release.actors.map((actor) => actor.name).join(' ')}`)
|| (release.date && `${release.entity.slug} ${formatDate(release.date, 'YYYY MM DD')}`)
|| null;
const slug = slugify(slugBase, '-', {
encode: true,
limit: config.titleSlugLength,
});
const curatedRelease = {
title: release.title,
entry_id: release.entryId || null,
entity_id: release.entity.id,
studio_id: release.studio?.id || null,
url: release.url,
date: Number(release.date) ? release.date : null,
date_precision: release.datePrecision,
slug,
description: release.description,
comment: release.comment,
// director: release.director,
// likes: release.rating && release.rating.likes,
// dislikes: release.rating && release.rating.dislikes,
// rating: release.rating && release.rating.stars && Math.floor(release.rating.stars),
deep: typeof release.deep === 'boolean' ? release.deep : false,
deep_url: release.deepUrl,
updated_batch_id: batchId,
};
if (type === 'scene') {
curatedRelease.shoot_id = release.shootId || null;
curatedRelease.production_date = Number(release.productionDate) ? release.productionDate : null;
curatedRelease.duration = release.duration;
}
if (release.productionLocation) {
curatedRelease.production_location = release.productionLocation;
if (argv.resolvePlace) {
const productionLocation = await resolvePlace(release.productionLocation);
if (productionLocation) {
curatedRelease.production_city = productionLocation.city;
curatedRelease.production_state = productionLocation.state;
curatedRelease.production_country_alpha2 = productionLocation.country;
}
}
}
if (!existingRelease && !release.id) {
curatedRelease.created_batch_id = batchId;
}
return curatedRelease;
}
async function attachChannelEntities(releases) {
const releasesWithoutEntity = releases.filter((release) => release.channel && (!release.entity || release.entity.type === 'network'));
const channelEntities = await knex('entities')
.select(knex.raw('entities.*, row_to_json(parents) as parent'))
.whereIn('entities.slug', releasesWithoutEntity.map((release) => release.channel))
.where('entities.type', 'channel')
.leftJoin('entities AS parents', 'parents.id', 'entities.parent_id');
const channelEntitiesBySlug = channelEntities.reduce((acc, entity) => ({ ...acc, [entity.slug]: entity }), {});
const releasesWithChannelEntity = await Promise.all(releases
.map(async (release) => {
if (release.channel && channelEntitiesBySlug[release.channel]) {
const curatedEntity = await curateEntity(channelEntitiesBySlug[release.channel]);
return {
...release,
entity: curatedEntity,
};
}
if (release.entity) {
return release;
}
logger.error(`Unable to match channel '${release.channel?.slug || release.channel}' from generic URL ${release.url}`);
return null;
}));
return releasesWithChannelEntity.filter(Boolean);
}
async function attachStudios(releases) {
const studioSlugs = releases.map((release) => release.studio).filter(Boolean);
const studios = await knex('entities')
.whereIn('slug', studioSlugs)
.where('type', 'studio');
const studioBySlug = studios.reduce((acc, studio) => ({ ...acc, [studio.slug]: studio }), {});
const releasesWithStudio = releases.map((release) => {
if (release.studio && studioBySlug[release.studio]) {
return {
...release,
studio: studioBySlug[release.studio],
};
}
if (release.studio) {
logger.warn(`Unable to match studio '${release.studio}' for ${release.url}`);
}
return release;
});
return releasesWithStudio;
}
function attachReleaseIds(releases, storedReleases, batchId) {
const storedReleaseIdsByEntityIdAndEntryId = storedReleases.reduce((acc, release) => {
if (!acc[release.entity_id]) acc[release.entity_id] = {};
acc[release.entity_id][release.entry_id] = release.id;
return acc;
}, {});
const releasesWithId = releases.map((release) => {
if (!release.entity) {
logger.error(`No entitity available for ${release.url}`);
return null;
}
const id = storedReleaseIdsByEntityIdAndEntryId[release.entity.id]?.[release.entryId]
|| storedReleaseIdsByEntityIdAndEntryId[release.entity.parent?.id]?.[release.entryId];
if (id) {
return {
...release,
id,
batchId,
};
}
return null;
}).filter(Boolean);
return releasesWithId;
}
function filterInternalDuplicateReleases(releases) {
const releasesByEntityIdAndEntryId = releases.reduce((acc, release) => {
if (!release.entity) {
return acc;
}
if (!release.entryId) {
logger.warn(`No entry ID supplied for "${release.title}" from '${release.entity.name}'`);
return acc;
}
if (!acc[release.entity.id]) {
acc[release.entity.id] = {};
}
acc[release.entity.id][release.entryId] = release;
return acc;
}, {});
return Object.values(releasesByEntityIdAndEntryId)
.map((entityReleases) => Object.values(entityReleases))
.flat();
}
async function filterDuplicateReleases(releases) {
const internalUniqueReleases = filterInternalDuplicateReleases(releases);
const internalUniqueReleaseChunks = chunk(internalUniqueReleases);
const duplicateReleaseEntryChunks = await Promise.map(internalUniqueReleaseChunks, async (internalUniqueReleasesChunk) => knex('releases')
.whereIn(['entry_id', 'entity_id'], internalUniqueReleasesChunk.map((release) => [release.entryId, release.entity.id]))
.orWhereIn(['entry_id', 'entity_id'], internalUniqueReleasesChunk
// scene IDs shared across network, mark as duplicate so scene can be updated with channel if only available on release day (i.e. Perv City)
.filter((release) => release.entity.parent?.parameters?.networkEntryIds)
.map((release) => [release.entryId, release.entity.parent.id])), { concurrency: 10 });
const duplicateReleaseEntries = duplicateReleaseEntryChunks.flat();
const duplicateReleasesByEntityIdAndEntryId = duplicateReleaseEntries.reduce((acc, release) => {
if (!acc[release.entity_id]) acc[release.entity_id] = {};
acc[release.entity_id][release.entry_id] = true;
return acc;
}, {});
const duplicateReleases = internalUniqueReleases.filter((release) => duplicateReleasesByEntityIdAndEntryId[release.entity.id]?.[release.entryId]
|| duplicateReleasesByEntityIdAndEntryId[release.entity.parent?.id]?.[release.entryId]);
const uniqueReleases = internalUniqueReleases.filter((release) => !duplicateReleasesByEntityIdAndEntryId[release.entity.id]?.[release.entryId]
&& !duplicateReleasesByEntityIdAndEntryId[release.entity.parent?.id]?.[release.entryId]);
return {
uniqueReleases,
duplicateReleases,
duplicateReleaseEntries,
};
}
async function updateSceneSearch(releaseIds) {
logger.info(`Updating search documents for ${releaseIds ? releaseIds.length : 'all' } releases`);
const documents = await knex.raw(`
SELECT
releases.id AS release_id,
TO_TSVECTOR(
'english',
COALESCE(releases.title, '') || ' ' ||
releases.entry_id || ' ' ||
entities.name || ' ' ||
entities.slug || ' ' ||
COALESCE(array_to_string(entities.alias, ' '), '') || ' ' ||
COALESCE(parents.name, '') || ' ' ||
COALESCE(parents.slug, '') || ' ' ||
COALESCE(array_to_string(parents.alias, ' '), '') || ' ' ||
COALESCE(releases.shoot_id, '') || ' ' ||
COALESCE(TO_CHAR(releases.date, 'YYYY YY MM FMMM FMMonth mon DD FMDD'), '') || ' ' ||
STRING_AGG(COALESCE(actors.name, ''), ' ') || ' ' ||
STRING_AGG(COALESCE(directors.name, ''), ' ') || ' ' ||
STRING_AGG(COALESCE(tags.name, ''), ' ') || ' ' ||
STRING_AGG(COALESCE(tags_aliases.name, ''), ' ')
) as document
FROM releases
LEFT JOIN entities ON releases.entity_id = entities.id
LEFT JOIN entities AS parents ON parents.id = entities.parent_id
LEFT JOIN releases_actors AS local_actors ON local_actors.release_id = releases.id
LEFT JOIN releases_directors AS local_directors ON local_directors.release_id = releases.id
LEFT JOIN releases_tags AS local_tags ON local_tags.release_id = releases.id
LEFT JOIN actors ON local_actors.actor_id = actors.id
LEFT JOIN actors AS directors ON local_directors.director_id = directors.id
LEFT JOIN tags ON local_tags.tag_id = tags.id AND tags.priority >= 6
LEFT JOIN tags as tags_aliases ON local_tags.tag_id = tags_aliases.alias_for AND tags_aliases.secondary = true
${releaseIds ? 'WHERE releases.id = ANY(?)' : ''}
GROUP BY releases.id, entities.name, entities.slug, entities.alias, parents.name, parents.slug, parents.alias;
`, releaseIds && [releaseIds]);
if (documents.rows?.length > 0) {
await bulkInsert('releases_search', documents.rows, ['release_id']);
}
}
async function storeChapters(releases) {
const chapters = releases
.map((release) => release.chapters?.map((chapter, index) => ({
releaseId: release.id,
index: index + 1,
time: chapter.time,
duration: chapter.duration,
title: chapter.title,
description: chapter.description,
poster: chapter.poster,
photos: chapter.photos,
tags: chapter.tags,
})))
.flat()
.filter(Boolean)
.sort((chapterA, chapterB) => chapterA.time - chapterB.time);
const curatedChapterEntries = chapters.map((chapter) => ({
index: chapter.index,
time: chapter.time,
duration: chapter.duration,
title: chapter.title,
description: chapter.description,
release_id: chapter.releaseId,
}));
const storedChapters = await bulkInsert('chapters', curatedChapterEntries, ['release_id', 'index']);
const chapterIdsByReleaseIdAndChapter = storedChapters.reduce((acc, chapter) => ({
...acc,
[chapter.release_id]: {
...acc[chapter.release_id],
[chapter.index]: chapter.id,
},
}), {});
const chaptersWithId = chapters.map((chapter) => ({
...chapter,
id: chapterIdsByReleaseIdAndChapter[chapter.releaseId][chapter.index],
}));
await associateReleaseTags(chaptersWithId, 'chapter');
// media is more error-prone, associate separately
await associateReleaseMedia(chaptersWithId, 'chapter');
}
async function storeScenes(releases, useBatchId) {
if (!releases || releases.length === 0) {
return [];
}
const [batchId] = useBatchId ? [useBatchId] : await knex('batches').insert({ comment: null }).returning('id');
const releasesWithChannels = await attachChannelEntities(releases);
const releasesWithBaseActors = releasesWithChannels.map((release) => ({ ...release, actors: toBaseActors(release.actors) }));
const releasesWithStudios = await attachStudios(releasesWithBaseActors);
// uniqueness is entity ID + entry ID, filter uniques after adding entities
const { uniqueReleases, duplicateReleases, duplicateReleaseEntries } = await filterDuplicateReleases(releasesWithStudios);
const curatedNewReleaseEntries = await Promise.all(uniqueReleases.map((release) => curateReleaseEntry(release, batchId)));
const storedReleases = await bulkInsert('releases', curatedNewReleaseEntries);
const storedReleaseEntries = Array.isArray(storedReleases) ? storedReleases : [];
const uniqueReleasesWithId = attachReleaseIds(uniqueReleases, storedReleaseEntries, batchId);
const duplicateReleasesWithId = attachReleaseIds(duplicateReleases, duplicateReleaseEntries, batchId);
const releasesWithId = uniqueReleasesWithId.concat(duplicateReleasesWithId);
const updated = await knex.raw(`
UPDATE releases
SET url = COALESCE(new.url, releases.url),
date = COALESCE(new.date, releases.date),
entity_id = COALESCE((new.entity->>'id')::integer, releases.entity_id),
title = COALESCE(new.title, releases.title),
description = COALESCE(new.description, releases.description),
duration = COALESCE(new.duration, releases.duration),
deep = new.url IS NOT NULL,
updated_at = NOW()
FROM json_to_recordset(:scenes)
AS new(id int, url text, date timestamptz, entity json, title text, description text, duration integer, deep boolean)
WHERE releases.id = new.id
`, {
scenes: JSON.stringify(duplicateReleasesWithId),
});
const [actors] = await Promise.all([
associateActors(releasesWithId, batchId),
associateReleaseTags(releasesWithId),
storeChapters(releasesWithId),
]);
await associateDirectors(releasesWithId, batchId); // some directors may also be actors, don't associate at the same time
await updateSceneSearch(releasesWithId.map((release) => release.id));
// media is more error-prone, associate separately
await associateReleaseMedia(releasesWithId);
if (argv.sceneActors && actors) {
await scrapeActors(actors.map((actor) => actor.name));
}
logger.info(`Stored ${storedReleaseEntries.length}, updated ${updated.rowCount} releases`);
await notify(releasesWithId);
return releasesWithId;
}
async function associateMovieScenes(movies, movieScenes) {
const moviesByEntityIdAndEntryId = movies.reduce((acc, movie) => ({
...acc,
[movie.entity.id]: {
...acc[movie.entity.id],
[movie.entryId]: movie,
},
}), {});
const associations = movieScenes.map((scene) => {
if (!scene.movie) {
return null;
}
const sceneMovie = moviesByEntityIdAndEntryId[scene.entity.id]?.[scene.movie.entryId];
if (sceneMovie?.id) {
return {
movie_id: sceneMovie.id,
scene_id: scene.id,
};
}
return null;
}).filter(Boolean);
await bulkInsert('movies_scenes', associations, false);
}
async function updateMovieSearch(movieIds) {
logger.info(`Updating search documents for ${movieIds ? movieIds.length : 'all' } movies`);
const documents = await knex.raw(`
SELECT
movies.id AS movie_id,
TO_TSVECTOR(
'english',
COALESCE(movies.title, '') || ' ' ||
entities.name || ' ' ||
entities.slug || ' ' ||
COALESCE(array_to_string(entities.alias, ' '), '') || ' ' ||
COALESCE(parents.name, '') || ' ' ||
COALESCE(parents.slug, '') || ' ' ||
COALESCE(array_to_string(parents.alias, ' '), '') || ' ' ||
COALESCE(TO_CHAR(movies.date, 'YYYY YY MM FMMM FMMonth mon DD FMDD'), '') || ' ' ||
STRING_AGG(COALESCE(releases.title, ''), ' ') || ' ' ||
STRING_AGG(COALESCE(actors.name, ''), ' ') || ' ' ||
STRING_AGG(COALESCE(tags.name, ''), ' ')
) as document
FROM movies
LEFT JOIN entities ON movies.entity_id = entities.id
LEFT JOIN entities AS parents ON parents.id = entities.parent_id
LEFT JOIN movies_scenes ON movies_scenes.movie_id = movies.id
LEFT JOIN releases ON releases.id = movies_scenes.scene_id
LEFT JOIN releases_actors ON releases_actors.release_id = movies_scenes.scene_id
LEFT JOIN releases_tags ON releases_tags.release_id = releases.id
LEFT JOIN actors ON actors.id = releases_actors.actor_id
LEFT JOIN tags ON tags.id = releases_tags.tag_id
${movieIds ? 'WHERE movies.id = ANY(?)' : ''}
GROUP BY movies.id, entities.name, entities.slug, entities.alias, parents.name, parents.slug, parents.alias;
`, movieIds && [movieIds]);
if (documents.rows?.length > 0) {
await bulkInsert('movies_search', documents.rows, ['movie_id']);
}
}
async function storeMovies(movies, useBatchId) {
if (!movies || movies.length === 0) {
return [];
}
const { uniqueReleases } = await filterDuplicateReleases(movies);
const [batchId] = useBatchId ? [useBatchId] : await knex('batches').insert({ comment: null }).returning('id');
const curatedMovieEntries = await Promise.all(uniqueReleases.map((release) => curateReleaseEntry(release, batchId, null, 'movie')));
const storedMovies = await bulkInsert('movies', curatedMovieEntries, ['entity_id', 'entry_id'], true);
const moviesWithId = attachReleaseIds(movies, storedMovies);
await updateMovieSearch(moviesWithId.map((movie) => movie.id));
await associateReleaseMedia(moviesWithId, 'movie');
return moviesWithId;
}
module.exports = {
associateMovieScenes,
storeScenes,
storeMovies,
updateSceneSearch,
updateMovieSearch,
};