Create script to generate combined data blob

common/get-movie-data.js

@@ -0,0 +1,57 @@
+const { MovieDb } = require("moviedb-promise");
+const { dailyCache } = require("./cache");
+require("dotenv").config();
+
+const moviedb = new MovieDb(process.env.MOVIEDB_API_KEY);
+
+const searchMovieAndCacheResults = ({
+  slug,
+  year: yearValue,
+  normalizedTitle,
+}) =>
+  dailyCache(`moviedb-search-${yearValue || "no-year"}-${slug}`, async () => {
+    const getPayload = (additional = {}) => ({
+      query: normalizedTitle,
+      ...additional,
+    });
+
+    // If there's no year provided, just search for the title
+    if (!yearValue) {
+      return moviedb.searchMovie(getPayload());
+    }
+
+    const year = parseInt(yearValue, 10);
+
+    // Try to find a movie first released on that year
+    let search = await moviedb.searchMovie(
+      getPayload({ primary_release_year: year }),
+    );
+
+    // If there's no matches, then try to find a movie with some release related
+    // to that year
+    if (search.results.length === 0) {
+      search = await moviedb.searchMovie(getPayload({ year }));
+    }
+
+    // If there's no matches, sometimes the movie listing has the year off by 1,
+    // so try to find a movie with some release related to the next year
+    if (search.results.length === 0) {
+      return moviedb.searchMovie(getPayload({ year: year + 1 }));
+    }
+
+    return search;
+  });
+
+const getMovieInfoAndCacheResults = ({ id }) =>
+  dailyCache(`moviedb-info-${id}`, async () => {
+    const payload = {
+      id,
+      append_to_response: "credits,external_ids,keywords,release_dates,videos",
+    };
+    return moviedb.movieInfo(payload);
+  });
+
+module.exports = {
+  searchMovieAndCacheResults,
+  getMovieInfoAndCacheResults,
+};

common/hydrate.js

@@ -1,58 +1,10 @@
 const slugify = require("slugify");
-const { MovieDb } = require("moviedb-promise");
-const { dailyCache } = require("./cache");
 const { parseMinsToMs } = require("./utils");
 const normalizeTitle = require("./normalize-title");
-require("dotenv").config();
-
-const moviedb = new MovieDb(process.env.MOVIEDB_API_KEY);
-
-const searchMovieAndCacheResults = ({
-  slug,
-  year: yearValue,
-  normalizedTitle,
-}) =>
-  dailyCache(`moviedb-search-${yearValue || "no-year"}-${slug}`, async () => {
-    const getPayload = (additional = {}) => ({
-      query: normalizedTitle,
-      ...additional,
-    });
-
-    // If there's no year provided, just search for the title
-    if (!yearValue) {
-      return moviedb.searchMovie(getPayload());
-    }
-
-    const year = parseInt(yearValue, 10);
-
-    // Try to find a movie first released on that year
-    let search = await moviedb.searchMovie(
-      getPayload({ primary_release_year: year }),
-    );
-
-    // If there's no matches, then try to find a movie with some release related
-    // to that year
-    if (search.results.length === 0) {
-      search = await moviedb.searchMovie(getPayload({ year }));
-    }
-
-    // If there's no matches, sometimes the movie listing has the year off by 1,
-    // so try to find a movie with some release related to the next year
-    if (search.results.length === 0) {
-      return moviedb.searchMovie(getPayload({ year: year + 1 }));
-    }
-
-    return search;
-  });
-
-const getMovieInfoAndCacheResults = ({ id }) =>
-  dailyCache(`moviedb-info-${id}`, async () => {
-    const payload = {
-      id,
-      append_to_response: "credits,external_ids,keywords,videos",
-    };
-    return moviedb.movieInfo(payload);
-  });
+const {
+  searchMovieAndCacheResults,
+  getMovieInfoAndCacheResults,
+} = require("./get-movie-data");
 
 const getMovieTitleAndYearFrom = (title) => {
   const hasYear = title.trim().match(/^(.*?)\s*\((\d{4})\)$/);

package.json

@@ -11,18 +11,21 @@
     "output:highlight-hydration-misses-for-review": "node ./scripts/highlight-hydration-misses-for-review.js",
     "clear:cache": "rm -rf ./cache && git checkout ./cache",
     "clear:output": "rm -rf ./output && git checkout ./output",
-    "populate:output": "./scripts/get-latest-release-assets.sh"
+    "populate:output": "./scripts/get-latest-release-assets.sh",
+    "generate:combined-data": "node ./scripts/generate-combined-data.js"
   },
   "author": "Alistair Brown <github@alistairjcbrown.com>",
   "license": "MIT",
   "dependencies": {
     "ajv": "^8.17.1",
     "ajv-formats": "^3.0.1",
     "cheerio": "^1.0.0-rc.12",
+    "compress-json": "^3.1.0",
     "date-fns": "^3.6.0",
     "dotenv": "^16.4.5",
     "ics": "^3.7.6",
     "moviedb-promise": "^4.0.7",
+    "nanoid": "^3.3.7",
     "playwright": "^1.48.1",
     "replace-special-characters": "^1.2.7",
     "slugify": "^1.6.6"

scripts/generate-combined-data.js

@@ -0,0 +1,179 @@
+const { writeFileSync } = require("node:fs");
+const path = require("node:path");
+const { nanoid } = require("nanoid");
+const { compress, trimUndefinedRecursively } = require("compress-json");
+const getSites = require("../common/get-sites");
+const normalizeTitle = require("../common/normalize-title");
+const { getMovieInfoAndCacheResults } = require("../common/get-movie-data");
+const { parseMinsToMs } = require("../common/utils");
+
+const getId = () => nanoid(8);
+
+const getCertification = ({ release_dates: { results } }) => {
+  const result = results.find(({ iso_3166_1: locale }) => locale === "GB");
+  if (!result) return undefined;
+
+  const { release_dates: releaseDates } = result;
+  const releaseDateWithCertification = releaseDates.find(
+    ({ certification }) => !!certification,
+  );
+
+  if (!releaseDateWithCertification) return undefined;
+  return releaseDateWithCertification.certification;
+};
+
+const getDirectors = ({ credits: { crew } }) =>
+  crew
+    .filter(({ job }) => job.toLowerCase() === "director")
+    .map(({ id, name }) => ({ id, name }));
+
+const getActors = ({ credits: { cast } }) =>
+  cast
+    .slice(0, 10)
+    .filter(({ popularity }) => popularity >= 5)
+    .map(({ id, name }) => ({ id, name }));
+
+const getGenres = ({ genres }) => genres;
+
+const getYoutubeTrailer = ({ videos: { results } }) => {
+  const trailer = results.find(
+    ({ type, site }) =>
+      type.toLowerCase() === "trailer" && site.toLowerCase() === "youtube",
+  );
+  return trailer ? trailer.key : undefined;
+};
+
+const getImddId = ({ external_ids: externalIds = {} }) => externalIds.imdb_id;
+
+const data = getSites().reduce((mapping, site) => {
+  let attributes;
+  let shows;
+  try {
+    attributes = require(
+      path.join(__dirname, "..", "cinemas", `${site}`, "attributes.js"),
+    );
+    shows = require(path.join(__dirname, "..", "output", `${site}-shows.json`));
+  } catch (e) {
+    return mapping;
+  }
+  return { ...mapping, [site]: { attributes, shows } };
+}, {});
+
+const siteData = {
+  venues: {},
+  people: {},
+  genres: {},
+  movies: {},
+};
+
+(async function () {
+  for (cinema in data) {
+    console.log(`[🎞️  Cinema: ${cinema}]`);
+    const venueId = getId();
+    const {
+      attributes: { name, url, address, geo },
+      shows,
+    } = data[cinema];
+
+    siteData.venues[venueId] = {
+      id: venueId,
+      name,
+      url,
+      address,
+      geo,
+    };
+
+    for (show of shows) {
+      const { title, url, overview, performances, moviedb } = show;
+
+      let movieInfo;
+      if (moviedb) {
+        const outputTitle = title.slice(0, 35);
+        process.stdout.write(
+          ` - Retriving data for ${outputTitle} ... ${"".padEnd(35 - outputTitle.length, " ")}`,
+        );
+        try {
+          movieInfo = await getMovieInfoAndCacheResults(moviedb);
+          console.log(`\t✅ Retrieved`);
+        } catch (e) {
+          console.log(`\t❌ Error retriving`);
+          throw e;
+        }
+      }
+
+      const movieId = movieInfo ? movieInfo.id : getId();
+      if (!siteData.movies[movieId]) {
+        if (movieInfo) {
+          const directors = getDirectors(movieInfo);
+          const actors = getActors(movieInfo);
+          const genres = getGenres(movieInfo);
+
+          directors.forEach((crew) => (siteData.people[crew.id] = crew));
+          actors.forEach((cast) => (siteData.people[cast.id] = cast));
+          genres.forEach((genre) => (siteData.genres[genre.id] = genre));
+
+          siteData.movies[movieId] = {
+            id: movieId,
+            title: movieInfo.title,
+            normalizedTitla: normalizeTitle(movieInfo.title),
+            certification: getCertification(movieInfo),
+            overview: movieInfo.overview,
+            year: movieInfo.release_date.split("-")[0],
+            duration: parseMinsToMs(movieInfo.runtime),
+            directors: directors.map(({ id }) => id),
+            actors: actors.map(({ id }) => id),
+            genres: genres.map(({ id }) => id),
+            imdbId: getImddId(movieInfo),
+            youtubeTrailer: getYoutubeTrailer(movieInfo),
+            posterPath: movieInfo.poster_path,
+            showings: {},
+            performances: [],
+          };
+        } else {
+          siteData.movies[movieId] = {
+            id: movieId,
+            title: title,
+            isUnmatched: true,
+            showings: {},
+            performances: [],
+          };
+        }
+      }
+
+      const showingId = getId();
+      const movie = siteData.movies[movieId];
+
+      movie.showings[showingId] = {
+        id: showingId,
+        venueId,
+        title: title !== movie.title ? title : undefined,
+        url,
+        overview,
+      };
+
+      movie.performances = movie.performances.concat(
+        performances.map(({ time, notes, bookingUrl }) => ({
+          showingId,
+          time,
+          notes: notes !== "" ? notes : undefined,
+          bookingUrl,
+        })),
+      );
+    }
+    console.log(" ");
+  }
+
+  process.stdout.write(`Compressing data ...   `);
+  try {
+    const dataFile = `./output/combined-data.json`;
+    trimUndefinedRecursively(siteData);
+    const compressed = compress(siteData).toString();
+    console.log(`✅ Compressed`);
+
+    writeFileSync(dataFile, compressed);
+    console.log(`🗂️  Combined file created`);
+  } catch (e) {
+    console.log(`\t❌ Error compressing`);
+    throw e;
+  }
+})();