Asset-Frameworker/utils/prediction_utils.py

132 lines
5.3 KiB
Python

import logging
import re
from pathlib import Path
from typing import Optional, Dict, Any
from rule_structure import SourceRule, RuleSet, MapRule, AssetRule
from configuration import load_preset
from utils.structure_analyzer import analyze_archive_structure # Hypothetical utility
log = logging.getLogger(__name__)
# Regex to extract preset name (similar to monitor.py)
# Matches "[PresetName]_anything.zip/rar/7z"
PRESET_FILENAME_REGEX = re.compile(r"^\[?([a-zA-Z0-9_-]+)\]?_.*\.(zip|rar|7z)$", re.IGNORECASE)
class PredictionError(Exception):
"""Custom exception for prediction failures."""
pass
def generate_source_rule_from_archive(archive_path: Path, config: Dict[str, Any]) -> SourceRule:
"""
Generates a SourceRule hierarchy based on rules defined in a preset,
determined by the archive filename.
Args:
archive_path: Path to the input archive file.
config: The loaded application configuration dictionary, expected
to contain preset information or a way to load it.
Returns:
The generated SourceRule hierarchy.
Raises:
PredictionError: If the preset cannot be determined, loaded, or
if rule generation fails.
FileNotFoundError: If the archive_path does not exist.
"""
if not archive_path.is_file():
raise FileNotFoundError(f"Archive file not found: {archive_path}")
log.debug(f"Generating SourceRule for archive: {archive_path.name}")
match = PRESET_FILENAME_REGEX.match(archive_path.name)
if not match:
raise PredictionError(f"Filename '{archive_path.name}' does not match expected format '[preset]_filename.ext'. Cannot determine preset.")
preset_name = match.group(1)
log.info(f"Extracted preset name: '{preset_name}' from {archive_path.name}")
try:
# Assuming load_preset takes the name and maybe the base config/path
# Adjust based on the actual signature of load_preset
preset_config = load_preset(preset_name) # This might need config path or dict
if not preset_config:
raise PredictionError(f"Preset '{preset_name}' configuration is empty or invalid.")
# Assuming the preset config directly contains the RuleSet structure
# or needs parsing into RuleSet. Let's assume it needs parsing.
# This part is highly dependent on how presets are stored and loaded.
# For now, let's assume preset_config IS the RuleSet dictionary.
if not isinstance(preset_config.get('rules'), dict):
raise PredictionError(f"Preset '{preset_name}' does not contain a valid 'rules' dictionary.")
rule_set_dict = preset_config['rules']
# Assuming RuleSet has a class method or similar for this
rule_set = RuleSet.from_dict(rule_set_dict) # Placeholder for actual deserialization
except FileNotFoundError:
raise PredictionError(f"Preset file for '{preset_name}' not found.")
except Exception as e:
log.exception(f"Failed to load or parse preset '{preset_name}': {e}")
raise PredictionError(f"Failed to load or parse preset '{preset_name}': {e}")
if not rule_set:
raise PredictionError(f"Failed to obtain RuleSet for preset '{preset_name}'.")
log.debug(f"Successfully loaded RuleSet for preset: {preset_name}")
# This simulates what a RuleBasedPredictionHandler might do, but without
# needing the actual extracted files for *this* step. The rules themselves
# define the expected structure. The ProcessingEngine will later use this
# rule against the actual extracted files.
# The actual structure (AssetRules, MapRules) comes directly from the RuleSet.
# We might need to adapt the archive name slightly (e.g., remove preset prefix)
# for the root node name, depending on desired output structure.
root_name = archive_path.stem
source_rule = SourceRule(name=root_name, rule_set=rule_set)
# Potentially add logic here if basic archive structure analysis *is* needed
# for rule generation (e.g., using utils.structure_analyzer if it exists)
log.info(f"Generated initial SourceRule for '{archive_path.name}' based on preset '{preset_name}'.")
# No temporary workspace needed/created in this function based on current plan.
# Cleanup is not required here.
return source_rule
# Example Usage (Conceptual - requires actual config/presets)
if __name__ == '__main__':
logging.basicConfig(level=logging.DEBUG)
log.info("Testing prediction_utils...")
dummy_archive = Path("./[TestPreset]_MyAsset.zip")
dummy_archive.touch()
preset_dir = Path(__file__).parent.parent / "Presets"
preset_dir.mkdir(exist_ok=True)
dummy_preset_path = preset_dir / "TestPreset.json"
dummy_preset_content = """
{
"name": "TestPreset",
"description": "A dummy preset for testing",
"rules": {
"map_rules": [
{"pattern": ".*albedo.*", "map_type": "Albedo", "color_space": "sRGB"},
{"pattern": ".*normal.*", "map_type": "Normal", "color_space": "Non-Color"}
],
"asset_rules": [
{"pattern": ".*", "material_name": "{asset_name}"}
]
},
"settings": {}
}
"""
log.warning("Note: Main execution block is commented out as it requires specific implementations of load_preset and RuleSet.from_dict.")