diff --git a/openpype/hosts/tvpaint/api/lib.py b/openpype/hosts/tvpaint/api/lib.py index cbc86f7b032..539cebe6468 100644 --- a/openpype/hosts/tvpaint/api/lib.py +++ b/openpype/hosts/tvpaint/api/lib.py @@ -77,8 +77,9 @@ def set_context_settings(asset_doc=None): handle_start = handles handle_end = handles - frame_start -= int(handle_start) - frame_end += int(handle_end) + # Always start from 0 Mark In and set only Mark Out + mark_in = 0 + mark_out = mark_in + (frame_end - frame_start) + handle_start + handle_end - execute_george("tv_markin {} set".format(frame_start - 1)) - execute_george("tv_markout {} set".format(frame_end - 1)) + execute_george("tv_markin {} set".format(mark_in)) + execute_george("tv_markout {} set".format(mark_out)) diff --git a/openpype/hosts/tvpaint/plugins/publish/collect_instance_frames.py b/openpype/hosts/tvpaint/plugins/publish/collect_instance_frames.py new file mode 100644 index 00000000000..f291c363b8e --- /dev/null +++ b/openpype/hosts/tvpaint/plugins/publish/collect_instance_frames.py @@ -0,0 +1,37 @@ +import pyblish.api + + +class CollectOutputFrameRange(pyblish.api.ContextPlugin): + """Collect frame start/end from context. + + When instances are collected context does not contain `frameStart` and + `frameEnd` keys yet. They are collected in global plugin + `CollectAvalonEntities`. + """ + label = "Collect output frame range" + order = pyblish.api.CollectorOrder + hosts = ["tvpaint"] + + def process(self, context): + for instance in context: + frame_start = instance.data.get("frameStart") + frame_end = instance.data.get("frameEnd") + if frame_start is not None and frame_end is not None: + self.log.debug( + "Instance {} already has set frames {}-{}".format( + str(instance), frame_start, frame_end + ) + ) + return + + frame_start = context.data.get("frameStart") + frame_end = context.data.get("frameEnd") + + instance.data["frameStart"] = frame_start + instance.data["frameEnd"] = frame_end + + self.log.info( + "Set frames {}-{} on instance {} ".format( + frame_start, frame_end, str(instance) + ) + ) diff --git a/openpype/hosts/tvpaint/plugins/publish/collect_instances.py b/openpype/hosts/tvpaint/plugins/publish/collect_instances.py index cc236734e5f..27bd8e9edef 100644 --- a/openpype/hosts/tvpaint/plugins/publish/collect_instances.py +++ b/openpype/hosts/tvpaint/plugins/publish/collect_instances.py @@ -86,9 +86,6 @@ def process(self, context): instance.data["publish"] = any_visible - instance.data["frameStart"] = context.data["sceneMarkIn"] + 1 - instance.data["frameEnd"] = context.data["sceneMarkOut"] + 1 - self.log.debug("Created instance: {}\n{}".format( instance, json.dumps(instance.data, indent=4) )) diff --git a/openpype/hosts/tvpaint/plugins/publish/extract_sequence.py b/openpype/hosts/tvpaint/plugins/publish/extract_sequence.py index 0d125a1a50d..007b5c41f19 100644 --- a/openpype/hosts/tvpaint/plugins/publish/extract_sequence.py +++ b/openpype/hosts/tvpaint/plugins/publish/extract_sequence.py @@ -1,8 +1,6 @@ import os import shutil -import time import tempfile -import multiprocessing import pyblish.api from avalon.tvpaint import lib @@ -45,10 +43,64 @@ def process(self, instance): ) family_lowered = instance.data["family"].lower() - frame_start = instance.data["frameStart"] - frame_end = instance.data["frameEnd"] + mark_in = instance.context.data["sceneMarkIn"] + mark_out = instance.context.data["sceneMarkOut"] + # Frame start/end may be stored as float + frame_start = int(instance.data["frameStart"]) + frame_end = int(instance.data["frameEnd"]) + + # Handles are not stored per instance but on Context + handle_start = instance.context.data["handleStart"] + handle_end = instance.context.data["handleEnd"] + + # --- Fallbacks ---------------------------------------------------- + # This is required if validations of ranges are ignored. + # - all of this code won't change processing if range to render + # match to range of expected output + + # Prepare output frames + output_frame_start = frame_start - handle_start + output_frame_end = frame_end + handle_end + + # Change output frame start to 0 if handles cause it's negative number + if output_frame_start < 0: + self.log.warning(( + "Frame start with handles has negative value." + " Changed to \"0\". Frames start: {}, Handle Start: {}" + ).format(frame_start, handle_start)) + output_frame_start = 0 + + # Check Marks range and output range + output_range = output_frame_end - output_frame_start + marks_range = mark_out - mark_in + + # Lower Mark Out if mark range is bigger than output + # - do not rendered not used frames + if output_range < marks_range: + new_mark_out = mark_out - (marks_range - output_range) + self.log.warning(( + "Lowering render range to {} frames. Changed Mark Out {} -> {}" + ).format(marks_range + 1, mark_out, new_mark_out)) + # Assign new mark out to variable + mark_out = new_mark_out + + # Lower output frame end so representation has right `frameEnd` value + elif output_range > marks_range: + new_output_frame_end = ( + output_frame_end - (output_range - marks_range) + ) + self.log.warning(( + "Lowering representation range to {} frames." + " Changed frame end {} -> {}" + ).format(output_range + 1, mark_out, new_mark_out)) + output_frame_end = new_output_frame_end + + # ------------------------------------------------------------------- - filename_template = self._get_filename_template(frame_end) + filename_template = self._get_filename_template( + # Use the biggest number + max(mark_out, frame_end) + ) ext = os.path.splitext(filename_template)[1].replace(".", "") self.log.debug("Using file template \"{}\"".format(filename_template)) @@ -57,7 +109,9 @@ def process(self, instance): output_dir = instance.data.get("stagingDir") if not output_dir: # Create temp folder if staging dir is not set - output_dir = tempfile.mkdtemp().replace("\\", "/") + output_dir = ( + tempfile.mkdtemp(prefix="tvpaint_render_") + ).replace("\\", "/") instance.data["stagingDir"] = output_dir self.log.debug( @@ -65,23 +119,36 @@ def process(self, instance): ) if instance.data["family"] == "review": - repre_files, thumbnail_fullpath = self.render_review( - filename_template, output_dir, frame_start, frame_end + output_filenames, thumbnail_fullpath = self.render_review( + filename_template, output_dir, mark_in, mark_out ) else: # Render output - repre_files, thumbnail_fullpath = self.render( - filename_template, output_dir, frame_start, frame_end, + output_filenames, thumbnail_fullpath = self.render( + filename_template, output_dir, + mark_in, mark_out, filtered_layers ) + # Sequence of one frame + if not output_filenames: + self.log.warning("Extractor did not create any output.") + return + + repre_files = self._rename_output_files( + filename_template, output_dir, + mark_in, mark_out, + output_frame_start, output_frame_end + ) + # Fill tags and new families tags = [] if family_lowered in ("review", "renderlayer"): tags.append("review") # Sequence of one frame - if len(repre_files) == 1: + single_file = len(repre_files) == 1 + if single_file: repre_files = repre_files[0] new_repre = { @@ -89,10 +156,13 @@ def process(self, instance): "ext": ext, "files": repre_files, "stagingDir": output_dir, - "frameStart": frame_start, - "frameEnd": frame_end, "tags": tags } + + if not single_file: + new_repre["frameStart"] = output_frame_start + new_repre["frameEnd"] = output_frame_end + self.log.debug("Creating new representation: {}".format(new_repre)) instance.data["representations"].append(new_repre) @@ -133,9 +203,45 @@ def _get_filename_template(self, frame_end): return "{{frame:0>{}}}".format(frame_padding) + ".png" - def render_review( - self, filename_template, output_dir, frame_start, frame_end + def _rename_output_files( + self, filename_template, output_dir, + mark_in, mark_out, output_frame_start, output_frame_end ): + # Use differnet ranges based on Mark In and output Frame Start values + # - this is to make sure that filename renaming won't affect files that + # are not renamed yet + mark_start_is_less = bool(mark_in < output_frame_start) + if mark_start_is_less: + marks_range = range(mark_out, mark_in - 1, -1) + frames_range = range(output_frame_end, output_frame_start - 1, -1) + else: + # This is less possible situation as frame start will be in most + # cases higher than Mark In. + marks_range = range(mark_in, mark_out + 1) + frames_range = range(output_frame_start, output_frame_end + 1) + + repre_filepaths = [] + for mark, frame in zip(marks_range, frames_range): + new_filename = filename_template.format(frame=frame) + new_filepath = os.path.join(output_dir, new_filename) + + repre_filepaths.append(new_filepath) + + if mark != frame: + old_filename = filename_template.format(frame=mark) + old_filepath = os.path.join(output_dir, old_filename) + os.rename(old_filepath, new_filepath) + + # Reverse repre files order if output + if mark_start_is_less: + repre_filepaths = list(reversed(repre_filepaths)) + + return [ + os.path.basename(path) + for path in repre_filepaths + ] + + def render_review(self, filename_template, output_dir, mark_in, mark_out): """ Export images from TVPaint using `tv_savesequence` command. Args: @@ -144,8 +250,8 @@ def render_review( keyword argument `{frame}` or index argument (for same value). Extension in template must match `save_mode`. output_dir (str): Directory where files will be stored. - first_frame (int): Starting frame from which export will begin. - last_frame (int): On which frame export will end. + mark_in (int): Starting frame index from which export will begin. + mark_out (int): On which frame index export will end. Retruns: tuple: With 2 items first is list of filenames second is path to @@ -154,10 +260,8 @@ def render_review( self.log.debug("Preparing data for rendering.") first_frame_filepath = os.path.join( output_dir, - filename_template.format(frame=frame_start) + filename_template.format(frame=mark_in) ) - mark_in = frame_start - 1 - mark_out = frame_end - 1 george_script_lines = [ "tv_SaveMode \"PNG\"", @@ -170,13 +274,22 @@ def render_review( ] lib.execute_george_through_file("\n".join(george_script_lines)) - output = [] first_frame_filepath = None - for frame in range(frame_start, frame_end + 1): + output_filenames = [] + for frame in range(mark_in, mark_out + 1): filename = filename_template.format(frame=frame) - output.append(filename) + output_filenames.append(filename) + + filepath = os.path.join(output_dir, filename) + if not os.path.exists(filepath): + raise AssertionError( + "Output was not rendered. File was not found {}".format( + filepath + ) + ) + if first_frame_filepath is None: - first_frame_filepath = os.path.join(output_dir, filename) + first_frame_filepath = filepath thumbnail_filepath = os.path.join(output_dir, "thumbnail.jpg") if first_frame_filepath and os.path.exists(first_frame_filepath): @@ -184,11 +297,10 @@ def render_review( thumbnail_obj = Image.new("RGB", source_img.size, (255, 255, 255)) thumbnail_obj.paste(source_img) thumbnail_obj.save(thumbnail_filepath) - return output, thumbnail_filepath - def render( - self, filename_template, output_dir, frame_start, frame_end, layers - ): + return output_filenames, thumbnail_filepath + + def render(self, filename_template, output_dir, mark_in, mark_out, layers): """ Export images from TVPaint. Args: @@ -197,8 +309,8 @@ def render( keyword argument `{frame}` or index argument (for same value). Extension in template must match `save_mode`. output_dir (str): Directory where files will be stored. - first_frame (int): Starting frame from which export will begin. - last_frame (int): On which frame export will end. + mark_in (int): Starting frame index from which export will begin. + mark_out (int): On which frame index export will end. layers (list): List of layers to be exported. Retruns: @@ -219,14 +331,11 @@ def render( # Sort layer positions in reverse order sorted_positions = list(reversed(sorted(layers_by_position.keys()))) if not sorted_positions: - return + return [], None self.log.debug("Collecting pre/post behavior of individual layers.") behavior_by_layer_id = lib.get_layers_pre_post_behavior(layer_ids) - mark_in_index = frame_start - 1 - mark_out_index = frame_end - 1 - tmp_filename_template = "pos_{pos}." + filename_template files_by_position = {} @@ -239,25 +348,47 @@ def render( tmp_filename_template, output_dir, behavior, - mark_in_index, - mark_out_index + mark_in, + mark_out ) - files_by_position[position] = files_by_frames + if files_by_frames: + files_by_position[position] = files_by_frames + else: + self.log.warning(( + "Skipped layer \"{}\". Probably out of Mark In/Out range." + ).format(layer["name"])) + + if not files_by_position: + layer_names = set(layer["name"] for layer in layers) + joined_names = ", ".join( + ["\"{}\"".format(name) for name in layer_names] + ) + self.log.warning( + "Layers {} do not have content in range {} - {}".format( + joined_names, mark_in, mark_out + ) + ) + return [], None output_filepaths = self._composite_files( files_by_position, - mark_in_index, - mark_out_index, + mark_in, + mark_out, filename_template, output_dir ) self._cleanup_tmp_files(files_by_position) + output_filenames = [ + os.path.basename(filepath) + for filepath in output_filepaths + ] + thumbnail_src_filepath = None - thumbnail_filepath = None if output_filepaths: - thumbnail_src_filepath = tuple(sorted(output_filepaths))[0] + thumbnail_src_filepath = output_filepaths[0] + thumbnail_filepath = None if thumbnail_src_filepath and os.path.exists(thumbnail_src_filepath): source_img = Image.open(thumbnail_src_filepath) thumbnail_filepath = os.path.join(output_dir, "thumbnail.jpg") @@ -265,11 +396,7 @@ def render( thumbnail_obj.paste(source_img) thumbnail_obj.save(thumbnail_filepath) - repre_files = [ - os.path.basename(path) - for path in output_filepaths - ] - return repre_files, thumbnail_filepath + return output_filenames, thumbnail_filepath def _render_layer( self, @@ -283,6 +410,22 @@ def _render_layer( layer_id = layer["layer_id"] frame_start_index = layer["frame_start"] frame_end_index = layer["frame_end"] + + pre_behavior = behavior["pre"] + post_behavior = behavior["post"] + + # Check if layer is before mark in + if frame_end_index < mark_in_index: + # Skip layer if post behavior is "none" + if post_behavior == "none": + return {} + + # Check if layer is after mark out + elif frame_start_index > mark_out_index: + # Skip layer if pre behavior is "none" + if pre_behavior == "none": + return {} + exposure_frames = lib.get_exposure_frames( layer_id, frame_start_index, frame_end_index ) @@ -341,8 +484,6 @@ def _render_layer( self.log.debug("Filled frames {}".format(str(_debug_filled_frames))) # Fill frames by pre/post behavior of layer - pre_behavior = behavior["pre"] - post_behavior = behavior["post"] self.log.debug(( "Completing image sequence of layer by pre/post behavior." " PRE: {} | POST: {}" @@ -530,17 +671,12 @@ def _composite_files( filepath = position_data[frame_idx] images_by_frame[frame_idx].append(filepath) - process_count = os.cpu_count() - if process_count > 1: - process_count -= 1 - - processes = {} output_filepaths = [] missing_frame_paths = [] random_frame_path = None for frame_idx in sorted(images_by_frame.keys()): image_filepaths = images_by_frame[frame_idx] - output_filename = filename_template.format(frame=frame_idx + 1) + output_filename = filename_template.format(frame=frame_idx) output_filepath = os.path.join(output_dir, output_filename) output_filepaths.append(output_filepath) @@ -553,45 +689,15 @@ def _composite_files( if len(image_filepaths) == 1: os.rename(image_filepaths[0], output_filepath) - # Prepare process for compositing of images + # Composite images else: - processes[frame_idx] = multiprocessing.Process( - target=composite_images, - args=(image_filepaths, output_filepath) - ) + composite_images(image_filepaths, output_filepath) # Store path of random output image that will 100% exist after all # multiprocessing as mockup for missing frames if random_frame_path is None: random_frame_path = output_filepath - self.log.info( - "Running {} compositing processes - this mey take a while.".format( - len(processes) - ) - ) - # Wait until all compositing processes are done - running_processes = {} - while True: - for idx in tuple(running_processes.keys()): - process = running_processes[idx] - if not process.is_alive(): - running_processes.pop(idx).join() - - if processes and len(running_processes) != process_count: - indexes = list(processes.keys()) - for _ in range(process_count - len(running_processes)): - if not indexes: - break - idx = indexes.pop(0) - running_processes[idx] = processes.pop(idx) - running_processes[idx].start() - - if not running_processes and not processes: - break - - time.sleep(0.01) - self.log.debug( "Creating transparent images for frames without render {}.".format( str(missing_frame_paths) diff --git a/openpype/hosts/tvpaint/plugins/publish/validate_marks.py b/openpype/hosts/tvpaint/plugins/publish/validate_marks.py index 73486d10059..e2ef81e4a44 100644 --- a/openpype/hosts/tvpaint/plugins/publish/validate_marks.py +++ b/openpype/hosts/tvpaint/plugins/publish/validate_marks.py @@ -14,37 +14,54 @@ class ValidateMarksRepair(pyblish.api.Action): def process(self, context, plugin): expected_data = ValidateMarks.get_expected_data(context) - expected_data["markIn"] -= 1 - expected_data["markOut"] -= 1 - - lib.execute_george("tv_markin {} set".format(expected_data["markIn"])) + lib.execute_george( + "tv_markin {} set".format(expected_data["markIn"]) + ) lib.execute_george( "tv_markout {} set".format(expected_data["markOut"]) ) class ValidateMarks(pyblish.api.ContextPlugin): - """Validate mark in and out are enabled.""" + """Validate mark in and out are enabled and it's duration. + + Mark In/Out does not have to match frameStart and frameEnd but duration is + important. + """ - label = "Validate Marks" + label = "Validate Mark In/Out" order = pyblish.api.ValidatorOrder optional = True actions = [ValidateMarksRepair] @staticmethod def get_expected_data(context): + scene_mark_in = context.data["sceneMarkIn"] + + # Data collected in `CollectAvalonEntities` + frame_end = context.data["frameEnd"] + frame_start = context.data["frameStart"] + handle_start = context.data["handleStart"] + handle_end = context.data["handleEnd"] + + # Calculate expeted Mark out (Mark In + duration - 1) + expected_mark_out = ( + scene_mark_in + + (frame_end - frame_start) + + handle_start + handle_end + ) return { - "markIn": int(context.data["frameStart"]), + "markIn": scene_mark_in, "markInState": True, - "markOut": int(context.data["frameEnd"]), + "markOut": expected_mark_out, "markOutState": True } def process(self, context): current_data = { - "markIn": context.data["sceneMarkIn"] + 1, + "markIn": context.data["sceneMarkIn"], "markInState": context.data["sceneMarkInState"], - "markOut": context.data["sceneMarkOut"] + 1, + "markOut": context.data["sceneMarkOut"], "markOutState": context.data["sceneMarkOutState"] } expected_data = self.get_expected_data(context)