discourse/script/bulk_import/uploads_importer.yml

source_db_path: "/path/to/your/db.sqlite3"
output_db_path: "/path/to/your/uploads.sqlite3"

root_paths:
  - "/path/to/your/files"
  - "/path/to/more/files"

# Files that are downloaded from URLs are cached in this directory.
download_cache_path: "/path/to/downloaded/files"

# The number of threads to use for processing uploads is calculated as:
#   thread_count = [number of cores] * [thread_count_factor]
# The thread count will be doubled if uploads are stored on S3 because there's a higher latency.
thread_count_factor: 1.5

# Delete uploads from the output database that are not found in the source database.
delete_surplus_uploads: false

# Delete uploads from the output database that do not have a Discourse upload record.
delete_missing_uploads: false

# Check if files are missing in the upload store and update the database accordingly.
# Set to false and re-run the script afterwards if you want to create new uploads for missing files.
fix_missing: false

# Create optimized images for post uploads and avatars.
create_optimized_images: false

site_settings:
  authorized_extensions: "*"
  max_attachment_size_kb: 102_400
  max_image_size_kb: 102_400

  enable_s3_uploads: true
  s3_upload_bucket: "your-bucket-name"
  s3_region: "your-region"
  s3_access_key_id: "your-access-key-id"
  s3_secret_access_key: "your-secret-access-key"
  s3_cdn_url: "https://your-cdn-url.com"

  # Set this to true if the site is a multisite and configure the `multisite_db_name` accordingly
  multisite: false
  multisite_db_name: "default"

# Sometimes a file can be found at one of many locations. Here's a list of transformations that can
# be applied to the path to try and find the file. The first transformation that results in a file
# being found will be used.
path_replacements:
#  - ["/foo/", "/bar"]
#  - ["/foo/", "/bar/baz/"]