@spec normalize_strings(Ecto.Changeset.t(), atom() | [atom()], keyword()) ::
  Ecto.Changeset.t()

Trims string changes and converts blank strings to nil by default.

This is useful for optional form fields where whitespace-only input should be stored as nil. Non-string fields and fields without a change are left alone.

Set :empty_to_nil to false to retain an empty string after trimming.

Examples

changeset
|> normalize_strings([:first_name, :last_name])

changeset
|> normalize_strings(:reference, empty_to_nil: false)

@spec trim_whitespace(Ecto.Changeset.t(), atom() | [atom()], keyword()) ::
  Ecto.Changeset.t()

Trims leading and trailing whitespace from one or more string fields in the changeset. Ensures consistency and helps maintain uniqueness constraints (e.g., on citext fields).

Options

• :max (integer): Maximum allowed length after trimming. If exceeded, a validation error is added. Default is 255.
• :enforce_unique (boolean): When set to true, adds a unique_constraint/3 to the field. Default is false.

Examples

changeset
|> trim_whitespace(:username, max: 50, enforce_unique: true)

changeset
|> trim_whitespace(:username, enforce_unique: true)

Parameters

• changeset (Ecto.Changeset.t()): The changeset containing the field(s) to be processed.
• field (atom() or [atom()]): The field(s) to trim.
• opts (keyword()): Options for trimming and validation.

Returns

• An updated Ecto.Changeset.t() with trimmed values and optional validations.

@spec validate_any_required(Ecto.Changeset.t(), [atom()], keyword()) ::
  Ecto.Changeset.t()

Validates that at least one of the given fields has a non-blank value.

By default, the error is attached to the first field. Use :error_field and :message to customize the resulting changeset error.

Example

changeset
|> validate_any_required([:email, :phone],
  error_field: :email,
  message: "email or phone is required"
)