Movies

A list of well-rated movies built from IMDb's public datasets, with the country each movie comes from. It is updated automatically every day and can be browsed two ways:

• On the web: https://dimap443.github.io/Movies/. Works in any browser, including on a phone.
• On Windows: the desktop app Movies.exe. It can also run the update itself.

Which movies are listed

A title is on the list when, in IMDb's data:

• its type is "movie" (feature films; TV movies, shorts and series are left out), and
• it is rated 6.8 or higher by at least 999 voters.

Each movie shows its title, year, genre, rating, number of votes, country and the date it was added to the list. Click a title on the web page, or double-click a row in the app, to open its IMDb page.

How the update works

The update runs on GitHub every day, and also when you press Update in the desktop app:

1. Download. It fetches title.basics.tsv.gz and title.ratings.tsv.gz from IMDb's datasets. A file is downloaded only if IMDb has changed it since the last update. If neither file changed and the last build was complete, the update stops here.
2. Filter. It keeps the titles with a rating of 6.8+ and 999+ votes from the ratings file, then the movies among them from the basics file.
3. Look up countries. Countries already known from the previous list are reused. The others are looked up on Wikidata by IMDb ID first, then on TMDb for titles Wikidata doesn't know. The TMDb step needs an API token (see Configuration) and is skipped without one; once a token is added, the next update runs even if IMDb hasn't changed. If a lookup fails, those countries stay blank and the next update tries again.
4. Save. It writes merged.tsv. A movie keeps its original Date added. A movie new to the list gets today's date. Files are written in full before they replace the old ones, so an interrupted update never leaves a half-written file.

On GitHub

The workflow .github/workflows/update.yml runs twice a day. The second run catches a missed first one and changes nothing if the first one ran. Each run:

1. builds the program and runs the update without a window (--update),
2. commits merged.tsv if it changed, and
3. publishes web/index.html together with the new merged.tsv to GitHub Pages.

It can also be started by hand from the repository's Actions tab ("Run workflow"). Pushing a change to web/, merged.tsv or the workflow republishes the site without an update.

Using the list

The web page and the desktop app behave the same way:

• Filter shows only rows containing the text you type, in any column.
• Hide country hides every movie from that country, including co-productions that involve it, and movies with no known country, since they may be from it too. India is listed first; the other countries follow alphabetically.
• Sorting: click a column header to sort, and click again to reverse. When sorting by year, movies from the same year are ordered by Date added, newest first, so a year's latest additions lead it.
• New movies: movies added today are highlighted in green for the whole day and shown at the top until you sort by a column.

The desktop app

Movies.exe is a Windows app with its own Java runtime built in, so Java doesn't need to be installed. Keep the whole Movies folder together: the exe needs its app and runtime folders.

• On startup, it shows the list it has. A merged.tsv placed next to Movies.exe is loaded if it is newer than the app's own copy. Each build of the exe includes the latest merged.tsv.
• It then checks the web site in the background: if the site's merged.tsv was published after the app's copy was written, the app downloads it and shows it. Update is disabled until this check finishes.
• Update runs the update described above on your PC. That takes a few minutes the first time, since IMDb's files are large.
• The app keeps its data in %LOCALAPPDATA%\Movies: merged.tsv, the downloaded IMDb files, and downloads.properties, a record of the downloaded versions.

Building

The program is a single Java class, src/main/java/org/example/FilterAndJoinTsv.java, built with Maven and JDK 25.

mvn clean package

This produces target\dist\Movies\Movies.exe with the current merged.tsv copied next to it. Run git pull first to include the latest daily update. Packaging the exe uses jpackage and Windows commands, so it must be built on Windows. Add -Dexec.skip=true to build only the jar (in target/libs), as the GitHub workflow does.

Run from the IDE, the program keeps its data in the working directory instead of %LOCALAPPDATA%\Movies.

To run the update without a window:

java --enable-preview -cp "target/libs/*" org.example.FilterAndJoinTsv --update

Configuration

• TMDb token (optional): the program reads it from the TMDB_API_TOKEN environment variable, or from a tmdb.txt file in the data directory. tmdb.txt is git-ignored, so the token is never committed. On GitHub, it is the repository secret TMDB_API_TOKEN.
• Rating and vote thresholds: SECOND_COL2_THRESHOLD (6.8) and SECOND_COL3_THRESHOLD (999) in FilterAndJoinTsv.java.

Files

File	What it is
src/main/java/org/example/FilterAndJoinTsv.java	The desktop app and the update logic
web/index.html	The web page; it loads merged.tsv from the same site
merged.tsv	The current list, updated daily by the workflow
.github/workflows/update.yml	The daily update and the GitHub Pages publish
pom.xml	The Maven build, including packaging Movies.exe

merged.tsv has no header row. Its columns are IMDb's title.basics columns (ID, type, title, original title, adult, year, end year, runtime, genres), then rating and votes, then country and Date added. \N marks a missing value.