Spaces:
Running on Zero
Running on Zero
File size: 8,616 Bytes
f1ef7e2 | 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 | import argparse
import csv
import json
import subprocess
import sys
from pathlib import Path
ML_SERVICES_ROOT = Path(__file__).resolve().parents[2]
PROJECT_ROOT = ML_SERVICES_ROOT.parent
DEFAULT_TRANSCRIPT_DIR = PROJECT_ROOT / "data" / "sentence_segments"
DEFAULT_METADATA_PATH = ML_SERVICES_ROOT / "data" / "processed" / "apptek_selected_domains" / "apptek_selected_domain_metadata_with_source_ids.csv"
CONVERTED_DIR = ML_SERVICES_ROOT / "data" / "processed" / "transcription_segments" / "converted_batch"
SIMPLE_OUTPUT_DIR = ML_SERVICES_ROOT / "outputs" / "apptek" / "simple_sentiment_results"
TIMESTAMPED_OUTPUT_DIR = ML_SERVICES_ROOT / "outputs" / "apptek" / "timestamped_sentiment_results"
BATCH_SUMMARY_PATH = ML_SERVICES_ROOT / "outputs" / "apptek" / "batch_transcript_sentiment_summary.csv"
def parse_args():
parser = argparse.ArgumentParser()
parser.add_argument("--transcript-dir", default=str(DEFAULT_TRANSCRIPT_DIR))
parser.add_argument("--metadata-path", default=str(DEFAULT_METADATA_PATH))
parser.add_argument("--limit", type=int, default=None)
return parser.parse_args()
def load_metadata(metadata_path):
mapping = {}
with open(metadata_path, newline="", encoding="utf-8") as file:
reader = csv.DictReader(file)
for row in reader:
source_id = row.get("source_apptek_id")
call_id = row.get("call_id")
if source_id and call_id:
mapping[source_id] = call_id
return mapping
def get_channel_audio_paths(call_id, metadata_mapping):
channel1_source_id = f"{call_id}_channel1"
channel2_source_id = f"{call_id}_channel2"
if channel1_source_id not in metadata_mapping:
raise ValueError(f"Missing channel1 mapping for {channel1_source_id}")
if channel2_source_id not in metadata_mapping:
raise ValueError(f"Missing channel2 mapping for {channel2_source_id}")
agent_call_id = metadata_mapping[channel1_source_id]
customer_call_id = metadata_mapping[channel2_source_id]
agent_audio_path = ML_SERVICES_ROOT / "data" / "processed" / "apptek_selected_domains" / "audio" / f"{agent_call_id}.wav"
customer_audio_path = ML_SERVICES_ROOT / "data" / "processed" / "apptek_selected_domains" / "audio" / f"{customer_call_id}.wav"
if not agent_audio_path.exists():
raise FileNotFoundError(f"Missing agent audio file: {agent_audio_path}")
if not customer_audio_path.exists():
raise FileNotFoundError(f"Missing customer audio file: {customer_audio_path}")
internal_call_id = f"{agent_call_id}_{customer_call_id.split('_')[-1]}"
return internal_call_id, agent_audio_path, customer_audio_path
def run_command(command):
print()
print("Running:")
print(" ".join(command))
subprocess.run(command, cwd=ML_SERVICES_ROOT, check=True)
def read_json(path):
with open(path, "r", encoding="utf-8") as file:
return json.load(file)
def main():
args = parse_args()
transcript_dir = Path(args.transcript_dir)
metadata_path = Path(args.metadata_path)
CONVERTED_DIR.mkdir(parents=True, exist_ok=True)
SIMPLE_OUTPUT_DIR.mkdir(parents=True, exist_ok=True)
TIMESTAMPED_OUTPUT_DIR.mkdir(parents=True, exist_ok=True)
BATCH_SUMMARY_PATH.parent.mkdir(parents=True, exist_ok=True)
metadata_mapping = load_metadata(metadata_path)
transcript_files = sorted(transcript_dir.rglob("*.json"))
if args.limit:
transcript_files = transcript_files[: args.limit]
print("Batch timestamped sentiment started")
print("-" * 80)
print(f"Transcript files found: {len(transcript_files)}")
print("-" * 80)
summary_rows = []
for index, transcript_path in enumerate(transcript_files, start=1):
print()
print("=" * 80)
print(f"[{index}/{len(transcript_files)}] Processing {transcript_path}")
print("=" * 80)
transcript_data = read_json(transcript_path)
source_call_id = transcript_data.get("call_id") or transcript_data.get("call")
domain = transcript_data.get("domain", transcript_path.parent.name)
try:
internal_call_id, agent_audio_path, customer_audio_path = get_channel_audio_paths(
source_call_id,
metadata_mapping,
)
converted_path = CONVERTED_DIR / domain / f"{source_call_id}_segments.json"
converted_path.parent.mkdir(parents=True, exist_ok=True)
simple_output_path = SIMPLE_OUTPUT_DIR / domain / f"{source_call_id}_simple_sentiment.json"
simple_output_path.parent.mkdir(parents=True, exist_ok=True)
timestamped_output_path = TIMESTAMPED_OUTPUT_DIR / f"{internal_call_id}_timestamped_sentiment.json"
run_command([
sys.executable,
"-m",
"src.data.convert_stereo_transcription_segments",
"--input-path",
str(transcript_path),
"--output-path",
str(converted_path),
"--call-id",
internal_call_id,
"--source-apptek-id",
source_call_id,
"--agent-audio-path",
str(agent_audio_path),
"--customer-audio-path",
str(customer_audio_path),
"--include-text",
"--min-duration-seconds",
"0",
])
run_command([
sys.executable,
"-m",
"src.inference.run_timestamped_sentiment",
"--segments-path",
str(converted_path),
])
run_command([
sys.executable,
"-m",
"src.inference.export_simple_sentiment_schema",
"--input-path",
str(timestamped_output_path),
"--output-path",
str(simple_output_path),
"--call-id",
source_call_id,
"--include-skipped",
])
timestamped_data = read_json(timestamped_output_path)
simple_data = read_json(simple_output_path)
summary_rows.append({
"source_call_id": source_call_id,
"domain": domain,
"status": "success",
"total_segments": timestamped_data.get("total_segments"),
"successful_segments": timestamped_data.get("successful_segments"),
"returned_simple_segments": len(simple_data.get("segments", [])),
"overall_audio_sentiment": timestamped_data.get("overall_audio_sentiment"),
"dominant_emotion": timestamped_data.get("dominant_emotion"),
"audio_escalation_score": timestamped_data.get("audio_escalation_score"),
"risk_level": timestamped_data.get("risk_level"),
"simple_output_path": str(simple_output_path.relative_to(ML_SERVICES_ROOT)),
})
except Exception as error:
print(f"FAILED for {source_call_id}: {error}")
summary_rows.append({
"source_call_id": source_call_id,
"domain": domain,
"status": f"failed: {error}",
"total_segments": "",
"successful_segments": "",
"returned_simple_segments": "",
"overall_audio_sentiment": "",
"dominant_emotion": "",
"audio_escalation_score": "",
"risk_level": "",
"simple_output_path": "",
})
fieldnames = [
"source_call_id",
"domain",
"status",
"total_segments",
"successful_segments",
"returned_simple_segments",
"overall_audio_sentiment",
"dominant_emotion",
"audio_escalation_score",
"risk_level",
"simple_output_path",
]
with open(BATCH_SUMMARY_PATH, "w", newline="", encoding="utf-8") as file:
writer = csv.DictWriter(file, fieldnames=fieldnames)
writer.writeheader()
writer.writerows(summary_rows)
successful = sum(1 for row in summary_rows if row["status"] == "success")
failed = len(summary_rows) - successful
print()
print("Batch timestamped sentiment completed")
print("-" * 80)
print(f"Total files: {len(summary_rows)}")
print(f"Successful: {successful}")
print(f"Failed: {failed}")
print(f"Summary CSV: {BATCH_SUMMARY_PATH}")
print("-" * 80)
if __name__ == "__main__":
main()
|