Disco-Courier

A data-mover-and-shaker for Disco Elysium.

The goal of Disco-Courier is to normalize and condense data for easy reading or transfer to the format of your choice. It drives data around in a SSSSOUPED UP MOTOR CARRIAGE!

It is the first of three repos aimed at a larger project.

Things you can use Disco-Courier for:

• quickly look up game elements to plan a walkthrough.
• generate a markdown-friendly table of some interesting dialog.
• quickly format your own custom views of the data for display.
• normalize data into a SQL db for use in your personal project.
• slice-and-dice results as pre-prep for putting into noSQL docs.
• isolate Garte-only dialog and train an AI model on it.*

*(please, please let me know if you do this).

(Note/Update: for Jamais Vu support (May 2023), see the 1.0-beta4RC1 branch)

Requirements

• You'll need a purchased copy of the game and access to its data.
• Assumes you have Node.js and npm installed.

Setup

• clone or download this repo
• run npm install in the directory created.
• place a copy of your exported data in /data/dialog.json
• note: make sure your extracted game data is in the JSON format!
• run a health check: npm run courier-health

If everything is setup correctly, you'll get a json response with summary details of the game's "FYS" attribute (...get it? ...Health check? ...GET IT?!? Yeah me either).

Usage

Here's a bunch of sample commands you can try to give you an idea of what's possibe (skip to next section for the full options/arguments reference):

npm run...

Command	Result
courier -- --export=read locations	prints a list of locations in json format to the screen.
courier -- --export=json items	writes a list of all in-game items to a json file in /data/json/items.json
courier -- --export=db actors	generates a Sequelize seed file for a table named "Actors" and populates it with all actors
courier -- --export=read --start=100 variables	displays all variables, starting at entry 100 to finish.
courier -- --export=md --start=4 --results=1 conversations	writes entry #4 in conversations as a table to /data/markdown/conversations.md
courier -- --export=read --results=4 actors.skill conversations.task	prints the first four results for both actors that are a skill, and conversations representing a task.
courier -- --export=md conversations.whitecheck conversations.redcheck	writes all white checks as a markdown-friendly table to /data/markdown/conversations.whitecheck.md and likewise all redchecks to data/markdown/conversations.redcheck.md
courier -- --export=db conversations.link	generates a seed file representing the entire dialog graph, normalized for a cross-reference table in a relational db
courier -- --export=json --start=188 conversations.dialog	writes all dialog entries starting at the 188th result to data/json/conversations.dialog.json
courier -- --export=db --start=1 conversations.link conversations.task conversations.subtask conversations.check	generates seed files for the dialog graph, a list of all tasks, a list of all subtasks, and all checks (white/red/passive) for their respective tables and populates with the first entry of each.
courier -- --export=read --results=5 --actor=3 --conversant=8 conversations.dialogtext	Creates a markdown-friendly table of the first five instances of Kim dialog where Jean Viqmarc is the conversant.
courier -- --export=json --actor=3 --OR=true --conversant=6 conversations.dialog	Creates a detailed json export where the speaker is Kim, OR the conversant is Garte.

Important: Make sure you include the empty double-dash (--) between the courier command and the arguments.

Note this is not raw output: each command passes through an extensive templating system that can be customized to taste.

###Arguments

Arguments take the form of options for paging and output, and a list of entities and sub-groups you want to export.

Options

flag	result
--start=<#>	begins output at specified number, starting with 1 (no zero index). Similar to an offset command.
--results=<#>	limits results to specified number. If used with start, will print the expected number of results from the start number. Similar to a limit command.
--export=<read|json|md|db>	The read option prints results to your terminal. The json option writes the results to an "entity.group.json" file. The md option attempts to format the result in a markdown-friendly table. The db option generates a sequelize seed file (see the Sequelize seed section for more on this option).
--actor=<id#>	filter results by actor ID (only applies to relevant conversation groups)
--conversant=<id#>	filter results by conversant ID (only applies to relevant conversation groups)
--OR=true	used to apply filters conditionally (actor OR conversant are true) default is AND (both conditions must be met)

Entities and groups

Use these as last argument to specify output. Can use multiple args. Omitting a sub-item exports all groups for the entity (e.g. "actors" on its own producs all rows in actors)

Entity	Group
locations	(none)
actors	actors,skill, actors.attribute
variables	(none)
items	(none - may re-introduce sub-types at a later date)
conversations	conversations.task, conversations.subtask, conversations.dialog, conversations.dialogtext, conversations.orb, conversations.check conversations.whitecheck, conversations.redcheck conversations.passivecheck

Customizing output

• All output is controlled by the template functions defined in /templates.
• Helper and utility methods are in the /search directory.
• Add a json-querying utility for extra flexibility if you like, or use the utilities that are in /search.

Adding new groups

• options are defined in /lib/args.ts
• routing to templates is done in /templates/index.ts
• if stuck, refer to /lib/migration.ts for details on how entities and groups are managed.
• if using the group with the --output=db flag, you'll need to generate a new schema using the sequelize-cli.

Live / Exploration Option

• use npm run dev -- --output=read <rest-of-options> if you want to use nodemon. Useful when making edits to the template system.

Sequelize Option

• If unfamiliar with SQL databases or the Sequelize ORM, this section probably isn't for you.
• The output of the db flag will place a seed in /data/seeders/<timestamp-entity-group.js>.
  • The resulting seed will reference Entity and Group names for the table name, with the first letter Capitalized, followed by an underscore and group name (if a group was specified):
    Example: --output=db actors.skill will generate File: /data/seeders/202107271226-add-actors.skill.js targeting Table: Actors_skill. See above section for more on Entities and groups.
    
• after you've generated some seed files, you can do npm run db:up to populate the database of your choice.
• SQLite should be ready out-of-the-box, you'll need to follow the standard Sequelize setup for Postgres, etc.

Generating Models

Sample models and migration scripts have been provided for nearly every sample template in the app. To import:

• Have a database that Sequelize supports, and make sure to install the proper Sequelize items so you can talk to it.
• Run at least one --output=db command on an entity you wish to transfer to a db.
• Run npm run db:up to create your tables and run the seeders.
• Run npm run db:down to roll back.

Notes

• Sample sequelize models and migrations can be found in /data/models and /data/migrations respectively.
• using the --results flag will greatly improve speed if limiting the number of results, but --start will need to go through the full dataset until it reaches the starting index. Further optimization and work on paging needs to be done.
• Dialogs within conversations are still a bit work-in-progress.
• USE CAUTION when working with a large amount of conversations.
• There are lengthy TODO comments in the source for contributors to check out. I don't have a lot of time, so contributions welcome.

Roadmap, barring time/interest/help

• Tidy up dialogueEntry formatting and make the generic model for it.
• Finish up checks and modifiers (see above) (Done).
• Add support for foreign key relationships for sequelize.
• While you can do a lot with the json export option for getting items into noSQL databases, it could use some better convenience options.
• CSV and spreadsheets would be nice (for easy dumps for things like checks).
• Markdown tables? If there's something easy, why not.
• Currently supports the version released 6th April 2021. Build on initial work to support additional versions.