File size: 10,395 Bytes
7e6c03a | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 | """
Conditional Duration task generator for temporal reasoning dataset.
Multi-hop task: First apply a temporal condition (before/after anchor),
then compare durations of events satisfying that condition.
Uses PreprocessedESC50Dataset for accurate effective durations.
"""
import random
from pathlib import Path
from typing import Dict, List, Optional, Tuple
from utils import (
AudioProcessor,
QuestionGenerator,
setup_logger,
set_random_seed,
generate_sample_durations_for_task,
concatenate_to_target_duration,
generate_controlled_gap_durations,
create_preprocessed_dataset,
)
from tasks.multihop_base import MultihopBaseGenerator
class ConditionalDurationTaskGenerator(MultihopBaseGenerator):
"""Generates conditional_duration task dataset samples."""
TASK_NAME = "conditional_duration"
def __init__(self, config: dict, logger=None):
super().__init__(config, logger)
# Override dataset with preprocessed version for effective durations
preprocessed_path = self.task_config.get(
"preprocessed_data_path",
config["tasks"].get("duration", {}).get("preprocessed_data_path", ""),
)
self.dataset = create_preprocessed_dataset(config, preprocessed_path=preprocessed_path)
def generate_sample(
self,
sample_id: int,
target_duration_seconds: float = None,
question_type: str = None,
) -> Optional[Dict]:
"""
Generate a single conditional_duration sample.
Pipeline:
1. Build sequential scene with 5-8 events, each with known effective duration
2. Pick an anchor sound
3. Among events before/after anchor, find longest/shortest
4. Generate question
"""
n_events = random.randint(
self.task_config.get("min_events", 5),
self.task_config.get("max_events", 8),
)
# Sample unique categories
n_unique = min(n_events, len(self.dataset.CATEGORIES))
categories = self.dataset.sample_categories(n_unique)
# Pad if needed
while len(categories) < n_events:
categories.append(random.choice(categories[:n_unique]))
random.shuffle(categories)
# Load audio with known durations
from pydub import AudioSegment as PydubSegment
source_files = []
audio_segments = []
effective_durations = []
num_gaps = n_events - 1
for cat in categories:
fname, fpath, eff_dur = self.dataset.sample_file_from_category_with_duration(cat)
audio = self.audio_processor.load_audio(fpath)
# Use effective duration to extend clip
per_event_s = max(self.source_clip_duration, eff_dur)
audio_segments.append(audio)
source_files.append(fname)
effective_durations.append(eff_dur)
# Build gaps
if num_gaps > 0:
gap_durations = generate_controlled_gap_durations(
num_gaps, min_gap_ms=300, max_gap_ms=1500, gap_multiplier=2.0
)
else:
gap_durations = []
# Assemble audio
result = audio_segments[0]
events_meta = [
{"index": 0, "category": categories[0],
"start_ms": 0, "end_ms": len(audio_segments[0]),
"effective_duration_s": effective_durations[0]}
]
current_ms = len(audio_segments[0])
for i in range(1, n_events):
gap_ms = gap_durations[i - 1]
result = result + PydubSegment.silent(duration=gap_ms)
current_ms += gap_ms
event_start = current_ms
result = result + audio_segments[i]
current_ms += len(audio_segments[i])
events_meta.append(
{"index": i, "category": categories[i],
"start_ms": event_start, "end_ms": current_ms,
"effective_duration_s": effective_durations[i],
"gap_before_ms": gap_ms}
)
# Save audio
output_path = self.audio_output / f"{sample_id}.wav"
result.export(str(output_path), format="wav")
# Select question type
if question_type is None:
question_type = random.choice(self.task_config["question_types"])
# Pick anchor (not first or last for before/after questions)
anchor_idx = random.randint(1, n_events - 2) if n_events > 2 else 0
anchor_sound = categories[anchor_idx]
# Compute answer based on question type
mcq_data, open_data, q_meta = self._generate_question(
question_type, categories, effective_durations, anchor_idx, anchor_sound
)
if mcq_data is None:
return None
metadata = {
"id": sample_id,
"audio_path": str(output_path.relative_to(self.output_base.parent)),
"n_events": n_events,
"categories": categories,
"source_files": source_files,
"effective_durations": effective_durations,
"question_type": question_type,
"anchor_sound": anchor_sound,
"anchor_index": anchor_idx,
"target_duration_s": target_duration_seconds,
"actual_duration_s": len(result) / 1000.0,
"mcq_question": mcq_data["question"],
"mcq_options": mcq_data["options"],
"mcq_correct_answer": mcq_data["correct_answer"],
"open_text_question": open_data["question"],
"open_text_answer": open_data["correct_answer"],
**q_meta,
}
self.logger.info(
f"Generated conditional_duration sample {sample_id}: "
f"{n_events} events, type={question_type}, anchor={anchor_sound}"
)
return metadata
def _generate_question(
self,
question_type: str,
categories: List[str],
effective_durations: List[float],
anchor_idx: int,
anchor_sound: str,
):
"""Generate conditional duration question."""
n = len(categories)
if question_type in ("longest_after", "shortest_after"):
subset_indices = list(range(anchor_idx + 1, n))
elif question_type in ("longest_before", "shortest_before"):
subset_indices = list(range(0, anchor_idx))
elif question_type == "repeated_compare":
# Find a category that appears both before and after anchor
before_cats = set(categories[:anchor_idx])
after_cats = set(categories[anchor_idx + 1:])
common = before_cats & after_cats
if not common:
return None, None, {}
target_sound = random.choice(list(common))
# Find indices
before_idx = [i for i in range(anchor_idx) if categories[i] == target_sound]
after_idx = [i for i in range(anchor_idx + 1, n) if categories[i] == target_sound]
before_dur = effective_durations[before_idx[0]]
after_dur = effective_durations[after_idx[0]]
if before_dur >= after_dur:
correct = f"before {anchor_sound}"
else:
correct = f"after {anchor_sound}"
mcq_text = self.task_config["mcq_questions"]["repeated_compare"].format(target_sound=target_sound, anchor_sound=anchor_sound)
open_text = self.task_config["open_text_questions"]["repeated_compare"].format(target_sound=target_sound, anchor_sound=anchor_sound)
options = [f"before {anchor_sound}", f"after {anchor_sound}"]
# Add distractors
other = [c for c in self.dataset.CATEGORIES if c != target_sound and c != anchor_sound]
random.shuffle(other)
options.extend(other[:2])
random.shuffle(options)
option_labels = ["A", "B", "C", "D"]
correct_label = option_labels[options.index(correct)]
option_map = {l: v for l, v in zip(option_labels, options)}
mcq_data = {"question": mcq_text, "options": option_map,
"correct_answer": correct_label, "correct_value": correct}
open_data = {"question": open_text, "correct_answer": correct}
q_meta = {"target_sound": target_sound, "correct_value": correct}
return mcq_data, open_data, q_meta
else:
return None, None, {}
if not subset_indices:
return None, None, {}
# Find longest or shortest in subset
subset_durations = [(i, effective_durations[i]) for i in subset_indices]
if "longest" in question_type:
best_idx, best_dur = max(subset_durations, key=lambda x: x[1])
else:
best_idx, best_dur = min(subset_durations, key=lambda x: x[1])
correct_category = categories[best_idx]
present_cats = [categories[i] for i in subset_indices]
mcq_text = self.task_config["mcq_questions"][question_type].format(anchor_sound=anchor_sound)
open_text = self.task_config["open_text_questions"][question_type].format(anchor_sound=anchor_sound)
mcq_data = self.question_generator.generate_category_mcq(
mcq_text, correct_category, present_cats, self.dataset.CATEGORIES
)
open_data = self.question_generator.generate_category_open_text(
open_text, correct_category
)
q_meta = {
"correct_category": correct_category,
"correct_duration_s": best_dur,
"subset_size": len(subset_indices),
}
return mcq_data, open_data, q_meta
def main(config_path: str = None):
"""Main entry point for conditional_duration task generation."""
import yaml
if config_path is None:
config_path = Path(__file__).parent.parent / "config.yaml"
with open(config_path, "r") as f:
config = yaml.safe_load(f)
set_random_seed(config["random_seed"])
logger = setup_logger(
"conditional_duration_task",
log_file=str(Path(config["output"]["base_path"]) / config["logging"]["log_file"]),
level=config["logging"]["level"],
console_output=config["logging"]["console_output"],
)
generator = ConditionalDurationTaskGenerator(config, logger)
generator.generate_dataset()
if __name__ == "__main__":
main()
|