Dragonflight Shadow Priest Sims

This repo includes scripts and sims for shadow priests.

Important Links

• Wiki
• SimulationCraft Wiki

Discussion

• Discord
• Website

Contributing

Contributions are always welcome, this project uses ruff for linting.

How to Run

All scripts are run built with python3, but should be able to be run with python2 (results may vary).

Prepare your environment

1. Run pip install -r requirements.txt in order for the scripts to work.
2. Edit and confirm config in config.yml, this controls how profiles are built, sims are ran, and aggregated. See the wiki for more information.

Full Suite

To run every folder of sims for composite and dungeons, you can simply run python suite.py. This will go through each and every sim folder and sim for each type based on config.yml for each sim. By default this currently excludes apl, talents, and talents-top (can be altered by passing a list to --exclude). This script keeps track of state via internal/suite.csv to know which sims it has run already for that suite run. This means if you start the suite and then stop/restart it later it will pick up where it left off. To start a new, clean run you just pash the --fresh argument. When using this command there is no reason to run the other two scripts below unless you are looking to do one off sims.

Create profiles

1. Run python profiles.py dir/ [--ptr, --dungeons, --talents [x, y]] for the directory you want to sim. Current talent options are indicated by the config keys under: builds. If you don't specify talents and that sim uses it based on config, all combinations will be automatically generated.

Simulate

Use raidbots

1. To run the sim in Raidbots create api_secrets.py inside the root directory. Set api_key = 'XXX'.
2. By default if a file already exists in output/ or if the weight in internal/weights.py is 0, sim.py will skip it.
3. To run the sims use python sim.py dir/ [--iterations 10000, --dungeons, --talents [x, y]] where dir/ is the sim directory you want to sim. If you don't specify talents and that sim uses it based on config, all combinations will be automatically generated.
4. Based on config keys in analyze markdown, csv, and json will be generated for the aggregated sims. These will output separate files for Composite, Single Target, or Dungeon sims. You can find all output files in the results/ folder in any sim folder.

Use local simc

1. To run with a local simc you have three options:

• use the simc you have on your Path ( nothing to setup here )
• use a simc located in a separated folder, create a local_secrets.py inside the root directory and set simc_path = '{"nightly": "path/to/executable", "rework": "another/path"}'
• add the --auto_download parameter to download the latest simc nightly build into the auto_download/ folder. This will override any choice in config.yml with the path to the newly downloaded version. (note: this is only supported on windows currently)

We use a dict here to support different simcVersions like raidbots, you can so define a different simc installation by every key you define. If you don't supply the local_secrets.py we will use the simc on the path for every different simcVersion defined in config.yml.

2. By default if a file already exists in output/ or if the weight in internal/weights.py is 0, sim.py will skip it.
3. To run the sims use python sim.py dir/ [--iterations 10000, --dungeons, --talents [am, hv] --local] where dir/ is the sim directory you want to sim. If you don't specify talents and that sim uses it based on config, all combinations will be automatically generated.
4. Based on config keys in analyze markdown, csv, and json will be generated for the aggregated sims. These will output separate files for Composite, Single Target, or Dungeon sims. You can find all output files in the results/ folder in any sim folder.

Output Formats

Based on config.yml sim results will output in up to 3 different formats: Markdown, CSV, and JSON. The files are all located in the results/ folder of each sim type. The following sections go over how the data is listed here. Each output file is made up into various sections based on how the sim is configured. There are several layers of results.

Layers

• Sim Type
  • Composite
  • Dungeons
  • Single
• Talents

These layers are used to create the file name. So if a sim is run with default setup without talent builds it will just be Results_SIMTYPE i.e. Results_Composite.md. If the sim is ran with talents this is augmented with that talent build (see config for current builds) i.e.Results_Composite_x.md.

Markdown

The markdown files are broken up and contain easy layer name at the top as a header. Each file is just a simple table broken up by Actor, DPS, and the Increase of that actor against the Base actor.

CSV

The CSV files are the simplest of the bunch, the only trick is that we parse the sim type profile as the start of the row. So a row is simply profile,actor,DPS,increase,.

JSON

JSON files contain the most rich information about the results. The basic structure is as follows:

{
    "name": "SIMTYPE - TALENT",
    "data": {
        "ACTOR": {
            "DPS": 1234
        },
    },
    "ids": {
        "ACTOR": null,
    },
    "simulated_steps": [
        "DPS"
    ],
    "sorted_data_keys": [
        "ACTOR",
    ],
    "last_updated": "YYYY-MM-DD"
}

• Name: simple name of the sim containing all relevant information about the data
• Data: this contains the sorted result list of each actor. The content of this will depend on the data itself as defined in config. For most sims this will just be a single entry for DPS as the step, but for things like trinkets this will be a step for each ilevel i.e. 420, 425.
• IDs: This will contain rich information for each actor to be used to generate tooltip data. There are several different ways to lookup this data, namely spell, item, or none. This lookupType is defined for each type of sim in config. Spell type will use this file that links actor names to spell IDs to populate this. The item type searches in .simc files of the sim for the corresponding actor name and adds the ID it finds on that line.
• Simulated Steps: This just enumerates the steps used in data
• Sorted Data Keys: This list is sorted by maximum dps in order.
• Last Updated: simply contains the timestamp that the chart was generated