facefusion/processors/modules/age_modifier.py

from argparse import ArgumentParser
from typing import Any, List

import cv2
import numpy
from cv2.typing import Size
from numpy.typing import NDArray

import facefusion.jobs.job_manager
import facefusion.jobs.job_store
import facefusion.processors.core as processors
from facefusion import config, content_analyser, face_classifier, face_detector, face_landmarker, face_masker, face_recognizer, inference_manager, logger, process_manager, state_manager, wording
from facefusion.common_helper import create_int_metavar, map_float
from facefusion.download import conditional_download_hashes, conditional_download_sources
from facefusion.face_analyser import get_many_faces, get_one_face
from facefusion.face_helper import merge_matrix, paste_back, warp_face_by_face_landmark_5
from facefusion.face_masker import create_occlusion_mask, create_static_box_mask
from facefusion.face_selector import find_similar_faces, sort_and_filter_faces
from facefusion.face_store import get_reference_faces
from facefusion.filesystem import in_directory, is_image, is_video, resolve_relative_path, same_file_extension
from facefusion.processors import choices as processors_choices
from facefusion.processors.typing import AgeModifierInputs
from facefusion.program_helper import find_argument_group
from facefusion.thread_helper import thread_semaphore
from facefusion.typing import Args, Face, InferencePool, Mask, ModelOptions, ModelSet, ProcessMode, QueuePayload, UpdateProgress, VisionFrame
from facefusion.vision import read_image, read_static_image, write_image

MODEL_SET : ModelSet =\
{
	'styleganex_age':
	{
		'hashes':
		{
			'age_modifier':
			{
				'url': 'https://github.com/facefusion/facefusion-assets/releases/download/models-3.0.0/styleganex_age.hash',
				'path': resolve_relative_path('../.assets/models/styleganex_age.hash')
			}
		},
		'sources':
		{
			'age_modifier':
			{
				'url': 'https://github.com/facefusion/facefusion-assets/releases/download/models-3.0.0/styleganex_age.onnx',
				'path': resolve_relative_path('../.assets/models/styleganex_age.onnx')
			}
		},
		'template': 'ffhq_512',
		'size': (512, 512)
	}
}


def get_inference_pool() -> InferencePool:
	model_sources = get_model_options().get('sources')
	return inference_manager.get_inference_pool(__name__, model_sources)


def clear_inference_pool() -> None:
	inference_manager.clear_inference_pool(__name__)


def get_model_options() -> ModelOptions:
	return MODEL_SET[state_manager.get_item('age_modifier_model')]


def register_args(program : ArgumentParser) -> None:
	group_processors = find_argument_group(program, 'processors')
	if group_processors:
		group_processors.add_argument('--age-modifier-model', help = wording.get('help.age_modifier_model'), default = config.get_str_value('processors.age_modifier_model', 'styleganex_age'), choices = processors_choices.age_modifier_models)
		group_processors.add_argument('--age-modifier-direction', help = wording.get('help.age_modifier_direction'), type = int, default = config.get_int_value('processors.age_modifier_direction', '0'), choices = processors_choices.age_modifier_direction_range, metavar = create_int_metavar(processors_choices.age_modifier_direction_range))
		facefusion.jobs.job_store.register_step_keys([ 'age_modifier_model', 'age_modifier_direction' ])


def apply_args(args : Args) -> None:
	state_manager.init_item('age_modifier_model', args.get('age_modifier_model'))
	state_manager.init_item('age_modifier_direction', args.get('age_modifier_direction'))


def pre_check() -> bool:
	download_directory_path = resolve_relative_path('../.assets/models')
	model_hashes = get_model_options().get('hashes')
	model_sources = get_model_options().get('sources')

	return conditional_download_hashes(download_directory_path, model_hashes) and conditional_download_sources(download_directory_path, model_sources)


def pre_process(mode : ProcessMode) -> bool:
	if mode in [ 'output', 'preview' ] and not is_image(state_manager.get_item('target_path')) and not is_video(state_manager.get_item('target_path')):
		logger.error(wording.get('choose_image_or_video_target') + wording.get('exclamation_mark'), __name__.upper())
		return False
	if mode == 'output' and not in_directory(state_manager.get_item('output_path')):
		logger.error(wording.get('specify_image_or_video_output') + wording.get('exclamation_mark'), __name__.upper())
		return False
	if mode == 'output' and not same_file_extension([ state_manager.get_item('target_path'), state_manager.get_item('output_path') ]):
		logger.error(wording.get('match_target_and_output_extension') + wording.get('exclamation_mark'), __name__.upper())
		return False
	return True


def post_process() -> None:
	read_static_image.cache_clear()
	if state_manager.get_item('video_memory_strategy') in [ 'strict', 'moderate' ]:
		clear_inference_pool()
	if state_manager.get_item('video_memory_strategy') == 'strict':
		content_analyser.clear_inference_pool()
		face_classifier.clear_inference_pool()
		face_detector.clear_inference_pool()
		face_landmarker.clear_inference_pool()
		face_masker.clear_inference_pool()
		face_recognizer.clear_inference_pool()


def modify_age(target_face : Face, temp_vision_frame : VisionFrame) -> VisionFrame:
	model_template = get_model_options().get('template')
	model_size = get_model_options().get('size')
	face_landmark_5 = target_face.landmark_set.get('5/68').copy()
	extend_face_landmark_5 = (face_landmark_5 - face_landmark_5[2]) * 2 + face_landmark_5[2]
	crop_vision_frame, affine_matrix = warp_face_by_face_landmark_5(temp_vision_frame, face_landmark_5, model_template, (256, 256))
	extend_vision_frame, extend_affine_matrix = warp_face_by_face_landmark_5(temp_vision_frame, extend_face_landmark_5, model_template, model_size)
	extend_vision_frame_raw = extend_vision_frame.copy()
	box_mask = create_static_box_mask(model_size, state_manager.get_item('face_mask_blur'), (0, 0, 0, 0))
	crop_masks =\
	[
		box_mask
	]

	if 'occlusion' in state_manager.get_item('face_mask_types'):
		occlusion_mask = create_occlusion_mask(crop_vision_frame)
		combined_matrix = merge_matrix([ extend_affine_matrix, cv2.invertAffineTransform(affine_matrix) ])
		occlusion_mask = cv2.warpAffine(occlusion_mask, combined_matrix, model_size)
		crop_masks.append(occlusion_mask)
	crop_vision_frame = prepare_vision_frame(crop_vision_frame)
	extend_vision_frame = prepare_vision_frame(extend_vision_frame)
	extend_vision_frame = apply_modify(crop_vision_frame, extend_vision_frame)
	extend_vision_frame = normalize_extend_frame(extend_vision_frame)
	extend_vision_frame = fix_color(extend_vision_frame_raw, extend_vision_frame)
	extend_crop_mask = cv2.pyrUp(numpy.minimum.reduce(crop_masks).clip(0, 1))
	extend_affine_matrix *= extend_vision_frame.shape[0] / 512
	paste_vision_frame = paste_back(temp_vision_frame, extend_vision_frame, extend_crop_mask, extend_affine_matrix)
	return paste_vision_frame


def apply_modify(crop_vision_frame : VisionFrame, crop_vision_frame_extended : VisionFrame) -> VisionFrame:
	age_modifier = get_inference_pool().get('age_modifier')
	age_modifier_inputs = {}

	for age_modifier_input in age_modifier.get_inputs():
		if age_modifier_input.name == 'target':
			age_modifier_inputs[age_modifier_input.name] = crop_vision_frame
		if age_modifier_input.name == 'target_with_background':
			age_modifier_inputs[age_modifier_input.name] = crop_vision_frame_extended
		if age_modifier_input.name == 'direction':
			age_modifier_inputs[age_modifier_input.name] = prepare_direction(state_manager.get_item('age_modifier_direction'))

	with thread_semaphore():
		crop_vision_frame = age_modifier.run(None, age_modifier_inputs)[0][0]

	return crop_vision_frame


def fix_color(extend_vision_frame_raw : VisionFrame, extend_vision_frame : VisionFrame) -> VisionFrame:
	color_difference = compute_color_difference(extend_vision_frame_raw, extend_vision_frame, (48, 48))
	color_difference_mask = create_static_box_mask(extend_vision_frame.shape[:2][::-1], 1.0, (0, 0, 0, 0))
	color_difference_mask = numpy.stack((color_difference_mask, ) * 3, axis = -1)
	extend_vision_frame = normalize_color_difference(color_difference, color_difference_mask, extend_vision_frame)
	return extend_vision_frame


def compute_color_difference(extend_vision_frame_raw : VisionFrame, extend_vision_frame : VisionFrame, size : Size) -> VisionFrame:
	extend_vision_frame_raw = extend_vision_frame_raw.astype(numpy.float32) / 255
	extend_vision_frame_raw = cv2.resize(extend_vision_frame_raw, size, interpolation = cv2.INTER_AREA)
	extend_vision_frame = extend_vision_frame.astype(numpy.float32) / 255
	extend_vision_frame = cv2.resize(extend_vision_frame, size, interpolation = cv2.INTER_AREA)
	color_difference = extend_vision_frame_raw - extend_vision_frame
	return color_difference


def normalize_color_difference(color_difference : VisionFrame, color_difference_mask : Mask, extend_vision_frame : VisionFrame) -> VisionFrame:
	color_difference = cv2.resize(color_difference, extend_vision_frame.shape[:2][::-1], interpolation = cv2.INTER_CUBIC)
	color_difference_mask = 1 - color_difference_mask.clip(0, 0.75)
	extend_vision_frame = extend_vision_frame.astype(numpy.float32) / 255
	extend_vision_frame += color_difference * color_difference_mask
	extend_vision_frame = extend_vision_frame.clip(0, 1)
	extend_vision_frame = numpy.multiply(extend_vision_frame, 255).astype(numpy.uint8)
	return extend_vision_frame


def prepare_direction(direction : int) -> NDArray[Any]:
	direction = map_float(float(direction), -100, 100, 2.5, -2.5) #type:ignore[assignment]
	return numpy.array(direction).astype(numpy.float32)


def prepare_vision_frame(vision_frame : VisionFrame) -> VisionFrame:
	vision_frame = vision_frame[:, :, ::-1] / 255.0
	vision_frame = (vision_frame - 0.5) / 0.5
	vision_frame = numpy.expand_dims(vision_frame.transpose(2, 0, 1), axis = 0).astype(numpy.float32)
	return vision_frame


def normalize_extend_frame(extend_vision_frame : VisionFrame) -> VisionFrame:
	extend_vision_frame = numpy.clip(extend_vision_frame, -1, 1)
	extend_vision_frame = (extend_vision_frame + 1) / 2
	extend_vision_frame = extend_vision_frame.transpose(1, 2, 0).clip(0, 255)
	extend_vision_frame = (extend_vision_frame * 255.0)
	extend_vision_frame = extend_vision_frame.astype(numpy.uint8)[:, :, ::-1]
	extend_vision_frame = cv2.pyrDown(extend_vision_frame)
	return extend_vision_frame


def get_reference_frame(source_face : Face, target_face : Face, temp_vision_frame : VisionFrame) -> VisionFrame:
	return modify_age(target_face, temp_vision_frame)


def process_frame(inputs : AgeModifierInputs) -> VisionFrame:
	reference_faces = inputs.get('reference_faces')
	target_vision_frame = inputs.get('target_vision_frame')
	many_faces = sort_and_filter_faces(get_many_faces([ target_vision_frame ]))

	if state_manager.get_item('face_selector_mode') == 'many':
		if many_faces:
			for target_face in many_faces:
				target_vision_frame = modify_age(target_face, target_vision_frame)
	if state_manager.get_item('face_selector_mode') == 'one':
		target_face = get_one_face(many_faces)
		if target_face:
			target_vision_frame = modify_age(target_face, target_vision_frame)
	if state_manager.get_item('face_selector_mode') == 'reference':
		similar_faces = find_similar_faces(many_faces, reference_faces, state_manager.get_item('reference_face_distance'))
		if similar_faces:
			for similar_face in similar_faces:
				target_vision_frame = modify_age(similar_face, target_vision_frame)
	return target_vision_frame


def process_frames(source_path : List[str], queue_payloads : List[QueuePayload], update_progress : UpdateProgress) -> None:
	reference_faces = get_reference_faces() if 'reference' in state_manager.get_item('face_selector_mode') else None

	for queue_payload in process_manager.manage(queue_payloads):
		target_vision_path = queue_payload['frame_path']
		target_vision_frame = read_image(target_vision_path)
		output_vision_frame = process_frame(
		{
			'reference_faces': reference_faces,
			'target_vision_frame': target_vision_frame
		})
		write_image(target_vision_path, output_vision_frame)
		update_progress(1)


def process_image(source_path : str, target_path : str, output_path : str) -> None:
	reference_faces = get_reference_faces() if 'reference' in state_manager.get_item('face_selector_mode') else None
	target_vision_frame = read_static_image(target_path)
	output_vision_frame = process_frame(
	{
		'reference_faces': reference_faces,
		'target_vision_frame': target_vision_frame
	})
	write_image(output_path, output_vision_frame)


def process_video(source_paths : List[str], temp_frame_paths : List[str]) -> None:
	processors.multi_process_frames(None, temp_frame_paths, process_frames)