# Tag Normalization Configuration
#
# This file controls how tags are processed, normalized, and filtered.
# The tag normalizer helps standardize tags by:
# 1. Converting aliases to canonical forms
# 2. Filtering out implied or redundant tags
# 3. Removing unwanted tags
# 4. Standardizing formatting
#
# NOTE: All spaces in tag names are automatically converted to underscores internally.
# When `use_underscores` is false, spaces are reintroduced before saving the result.
# When writing patterns, aliases, or tag names in this file, you can use either:
#   - Spaces: "rating explicit"  
#   - Underscores: "rating_explicit"
# Both will work the same way, as spaces are converted to underscores during processing.
# This applies to all tag names in blacklists, aliases, regexp patterns, etc.

# === FILE SELECTION ===

# Only process files matching this pattern
# This is a regular expression that must match the entire filename
# Default: matches .txt, .cap, .caption, etc. files
include_filename_regexp = ".*?\\.(txt|cap.*|wd|jtp)$"

# Skip files matching this pattern (even if they match the include pattern)
# Default: skips files with "sample" or "prompt" in their names
exclude_filename_regexp = ".*samples?-prompts?.*"


# === FILE STRUCTURE ===

# Regular expression pattern for separating tags in input files
# This pattern is used to split the content of tag files into individual tags
# Default: "[,\n]" (split on commas and newlines)
tag_separator_regexp = "[,\\n]"

# Regular expression pattern with named capture groups for extracting the tag section
# The optional prefix and suffix are concatenated unchanged to the processed middle tag section.
# Format: Use (?P<prefix>...), (?P<tags>...), and (?P<suffix>...) groups
# Default: "(?P<tags>.*?)", process the entire file as tags
#
# The prefix and suffix can be optional - only a match the tags group is required
# Example configuration: Process text between ||| and ]]] markers:
#    tag_section_pattern = '''(?P<prefix>.*?\|\|\|)?(?P<tags>.*?)(?P<suffix>\]\]\].*)?'''
tag_section_pattern = '''(?P<tags>.*?)'''


# === TAG FORMATTING ===

# Whether to use underscores or spaces in output tags
# true = use underscores (tag_name)
# false = use spaces (tag name)
# Default: false (use spaces)
use_underscores = false

# List of specific tags where underscores should always be preserved
# Even if use_underscores=false, these tags will keep their underscores
keep_underscores = [
    "rating_explicit",
    "rating_questionable", 
    "rating_safe",
    "source_pony",
    "source_furry",
    "source_anime",
    "source_cartoon",
]

# === TAG IMPLICATIONS ===

# Whether to keep or remove implied tags
# For example, if "dog" implies "mammal", and keep_implied=false,
# then "mammal" would be removed when "dog" is present
# Options:
#   false = remove implied tags
#   true = keep all implied tags
#   [list of tags] = keep only these specific implied tags
# Default: false (remove implied tags)
keep_implied = false

# Frequency threshold for reliable implications
# Tags that appear in fewer than this many posts on e621
# will not have their implications removed.
# Instead, the implied tags will be kept alongside the rare tag.
# Higher number = stricter filtering
# Default: 0 (disabled)
min_antecedent_freq = 200

# Frequency threshold for replacing rare tags with their implications
# Tags with fewer than this many posts will be replaced by the tags they imply
# This must be less than or equal to min_antecedent_freq
# For example, if "rare_species" implies "mammal" and has fewer posts than this number,
# "rare_species" would be removed and "mammal" kept
# Default: 0 (disabled)
drop_antecedent_freq = 50

# === ARTIST TAGS ===

# Whether to add "by_" prefix to artist tags
# For example, "artist_name" becomes "by_artist_name"
# Default: true
artist_by_prefix = true

# Whether to remove category suffixes like _(species) or _(character)
# when they match the tag's actual category
# This only happens if removing the suffix doesn't create a conflict
# Default: true
remove_parens = true

# === BLACKLISTING (REMOVING UNWANTED TAGS) ===

# Specific tags to completely remove during normalization
# These tags will never appear in the output
blacklist = [
    "invalid tag",
    "by conditional dnp",
    "hi res",
    "absurd res",
    "superabsurd res",
    "4k",
    "uncensored",
    "ambiguous gender",
    "translation edit",
    "story in description",
    "non- balls",
    "non- nipples",
    "non- breasts",
    "feet out of frame",
    "funny_post_number",
    "tagme",
    "edit_request",
]

# Pattern-based tag removal
# Tags matching these patterns will be removed
# These are full-match patterns (must match the entire tag)
# Note: When writing patterns, remember that spaces in tags are converted to underscores
# internally, so spaces are also replaced by underscores in these regexp.
blacklist_regexp = [
    "(\\d+s?|\\d+:\\d+)", # Numbers, years and aspect ratio
    ".*?_at_source",
]

# Whether to also remove tags that are implied by blacklisted tags
# For example, if "bad_tag" implies "general_tag" and blacklist_implied=true,
# then blacklisting "bad_tag" would also remove "general_tag"
# Default: true
blacklist_implied = true

# Entire categories of tags to remove
# Categories in e621: "artist", "character", "copyright", "general", 
# "meta", "species", "pool"
blacklist_categories = ["pool"]

# === ALIAS HANDLING ===

# Tags to remove from the alias system
# This prevents specific tags from being normalized to something else
# Useful for keeping tags separate when they would otherwise be merged
remove_aliases = []

# How to handle conflicts when creating aliases for `[alias]` and 
# This happens when the alias already maps to a normalized form different
# than the new one.
# Options:
#   "ignore" = Don't create the new alias if there's a conflict (keep existing mapping)
#   "overwrite" = Replace the existing mapping with the new one
#   "overwrite_rarest" = Keep the mapping to the more frequent tag (lower ID number)
#   "warn" = Log a warning then ignore
#   "raise" = Stop with an error
# Default: "warn"
on_alias_conflict = "warn"

# === TAG MAPPINGS ===

# Define custom aliases (alternate names for the same tag)
# Format: alias = "canonical_form"
# For example, "explicit = rating_explicit" means the tag "explicit"
# will be converted to "rating_explicit" during processing
# Note: You can use spaces or underscores interchangeably in both the alias name and value
# For example: "score explicit" = "rating explicit" works the same as "score_explicit" = "rating_explicit"
[aliases]
explicit = "rating_explicit"
score_explicit = "rating_explicit"
score_safe = "rating_safe"
score_questionable = "rating_questionable"

# Aliases that override existing mappings
# Use this carefully as it can change the meaning of existing tags
# These take precedence over regular aliases
[aliases_overrides]
safe = "rating_safe"
questionable = "rating_questionable"

# Specify tags to rename in the output
# This changes the canonical form of a tag
# The tag will still be recognized in all its forms, but will be output
# using the new name specified here
# Format: original_name = "new_name"
[renames]
domestic_cat = "cat"
domestic_dog = "dog"