File size: 12,023 Bytes
73b0303 | 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 | """CLI entrypoints for the project reboot."""
from __future__ import annotations
import json
from pathlib import Path
import typer
from feedback_intelligence.app.gradio_app import create_demo
from feedback_intelligence.benchmarks.tfidf_logreg import run_tfidf_logreg_baseline
from feedback_intelligence.config import (
AmazonTransferEvaluationConfig,
BaselineExperimentConfig,
LocalEvaluationConfig,
ReviewAnalysisConfig,
TransformerTrainingConfig,
)
from feedback_intelligence.data.amazon_reviews import (
load_amazon_polarity_reviews,
summarize_reviews as summarize_amazon_reviews,
)
from feedback_intelligence.data.imdb import load_local_imdb_reviews, summarize_reviews
from feedback_intelligence.data.local_reviews import load_local_labeled_reviews
from feedback_intelligence.inference.sentiment import load_sentiment_predictor
from feedback_intelligence.pipeline.review_analysis import analyze_reviews_with_predictor
from feedback_intelligence.pipeline.transfer_evaluation import evaluate_reviews_with_predictor
from feedback_intelligence.training.transformer import train_transformer_model
from feedback_intelligence.utils.io import write_json
app = typer.Typer(
help="Feedback Intelligence project commands.",
no_args_is_help=True,
)
BASE_PATH_OPTION = typer.Option(
Path("aclImdb"),
exists=True,
file_okay=False,
dir_okay=True,
help="Path to the local IMDb dataset root.",
)
SAMPLE_SIZE_OPTION = typer.Option(
2_000,
min=2,
help="Balanced number of train rows to sample for inspection.",
)
SEED_OPTION = typer.Option(42, help="Deterministic sampling seed.")
OUTPUT_PATH_OPTION = typer.Option(
Path("artifacts/benchmarks/tfidf_logreg_imdb.json"),
help="Where to write the benchmark artifact.",
)
CONFIG_PATH_OPTION = typer.Option(
None,
exists=True,
file_okay=True,
dir_okay=False,
help="Optional JSON config for the baseline experiment.",
)
ANALYSIS_OUTPUT_OPTION = typer.Option(
Path("artifacts/analysis/review_analysis_imdb.json"),
help="Where to write the review analysis artifact.",
)
ANALYSIS_CONFIG_PATH_OPTION = typer.Option(
None,
exists=True,
file_okay=True,
dir_okay=False,
help="Optional JSON config for the review analysis workflow.",
)
HOST_OPTION = typer.Option("127.0.0.1", help="Host interface for the demo server.")
PORT_OPTION = typer.Option(7860, min=1, max=65535, help="Port for the demo server.")
SHARE_OPTION = typer.Option(False, help="Create a public Gradio share link.")
TRAINER_CONFIG_PATH_OPTION = typer.Option(
None,
exists=True,
file_okay=True,
dir_okay=False,
help="Optional JSON config for transformer training.",
)
TRANSFER_CONFIG_PATH_OPTION = typer.Option(
None,
exists=True,
file_okay=True,
dir_okay=False,
help="Optional JSON config for Amazon transfer evaluation.",
)
TRANSFER_OUTPUT_OPTION = typer.Option(
Path("artifacts/evaluations/amazon_transfer_tfidf_imdb.json"),
help="Where to write the Amazon transfer evaluation artifact.",
)
LOCAL_EVAL_CONFIG_PATH_OPTION = typer.Option(
None,
exists=True,
file_okay=True,
dir_okay=False,
help="Optional JSON config for evaluating a labeled local feedback CSV.",
)
LOCAL_EVAL_OUTPUT_OPTION = typer.Option(
Path("artifacts/evaluations/customer_feedback_eval_200.json"),
help="Where to write the local customer-feedback evaluation artifact.",
)
@app.callback()
def main() -> None:
"""Top-level CLI group."""
@app.command("status")
def status(message: str | None = None) -> None:
"""Show the current reboot status."""
if message:
typer.echo(message)
return
typer.echo("Feedback Intelligence environment is configured.")
@app.command("describe-dataset")
def describe_dataset(
base_path: Path = BASE_PATH_OPTION,
sample_size: int = SAMPLE_SIZE_OPTION,
seed: int = SEED_OPTION,
) -> None:
"""Print a compact summary of the local IMDb dataset."""
train_records = load_local_imdb_reviews(
base_path=base_path,
split="train",
sample_size=sample_size,
seed=seed,
)
test_records = load_local_imdb_reviews(
base_path=base_path,
split="test",
sample_size=max(sample_size // 2, 2),
seed=seed,
)
payload = {
"train": summarize_reviews(train_records),
"test": summarize_reviews(test_records),
}
typer.echo(json.dumps(payload, indent=2))
@app.command("describe-amazon-dataset")
def describe_amazon_dataset(
sample_size: int = SAMPLE_SIZE_OPTION,
seed: int = SEED_OPTION,
) -> None:
"""Print a compact summary of a sampled Amazon polarity dataset slice."""
train_records = load_amazon_polarity_reviews(
split="train",
sample_size=sample_size,
seed=seed,
)
test_records = load_amazon_polarity_reviews(
split="test",
sample_size=max(sample_size // 2, 2),
seed=seed,
)
payload = {
"train": summarize_amazon_reviews(train_records),
"test": summarize_amazon_reviews(test_records),
}
typer.echo(json.dumps(payload, indent=2))
@app.command("run-baseline")
def run_baseline(
base_path: Path = BASE_PATH_OPTION,
output_path: Path = OUTPUT_PATH_OPTION,
config_path: Path = CONFIG_PATH_OPTION,
) -> None:
"""Run the first reproducible benchmark on the local IMDb dataset."""
config = (
BaselineExperimentConfig.from_json(config_path)
if config_path is not None
else BaselineExperimentConfig()
)
train_records = load_local_imdb_reviews(
base_path=base_path,
split="train",
sample_size=config.sample_size,
seed=config.seed,
)
test_records = load_local_imdb_reviews(
base_path=base_path,
split="test",
sample_size=max(config.sample_size // 2, 2),
seed=config.seed,
)
result = run_tfidf_logreg_baseline(
train_records=train_records,
test_records=test_records,
config=config,
)
write_json(output_path, result.to_dict())
typer.echo(f"Wrote benchmark artifact to {output_path}")
typer.echo(f"Saved baseline model to {config.model_output_path}")
@app.command("train-transformer")
def train_transformer(
base_path: Path = BASE_PATH_OPTION,
config_path: Path = TRAINER_CONFIG_PATH_OPTION,
) -> None:
"""Fine-tune a transformer sentiment model and save it for inference."""
config = (
TransformerTrainingConfig.from_json(config_path)
if config_path is not None
else TransformerTrainingConfig()
)
train_records = load_local_imdb_reviews(
base_path=base_path,
split="train",
sample_size=config.train_sample_size,
seed=config.seed,
)
test_records = load_local_imdb_reviews(
base_path=base_path,
split="test",
sample_size=config.test_sample_size,
seed=config.seed,
)
result = train_transformer_model(
train_records=train_records,
test_records=test_records,
config=config,
)
typer.echo(f"Saved transformer model to {result.output_dir}")
typer.echo(f"Best validation checkpoint came from epoch {result.best_epoch}")
typer.echo(f"Wrote transformer metrics to {config.metrics_output_path}")
@app.command("analyze-reviews")
def analyze_reviews_command(
base_path: Path = BASE_PATH_OPTION,
output_path: Path = ANALYSIS_OUTPUT_OPTION,
config_path: Path = ANALYSIS_CONFIG_PATH_OPTION,
) -> None:
"""Generate clustered review insights and review priorities."""
analysis_config = (
ReviewAnalysisConfig.from_json(config_path)
if config_path is not None
else ReviewAnalysisConfig()
)
analysis_records = load_local_imdb_reviews(
base_path=base_path,
split="test",
sample_size=analysis_config.analysis_sample_size,
seed=analysis_config.seed,
)
predictor = load_sentiment_predictor(
model_path=Path(analysis_config.sentiment_model_path).resolve(),
backend=analysis_config.sentiment_backend,
max_length=analysis_config.sentiment_max_length,
)
artifact = analyze_reviews_with_predictor(
review_records=analysis_records,
predictor=predictor,
analysis_config=analysis_config,
sentiment_model_info=predictor.describe(),
)
write_json(output_path, artifact.to_dict())
typer.echo(f"Wrote review analysis artifact to {output_path}")
@app.command("evaluate-amazon-transfer")
def evaluate_amazon_transfer(
output_path: Path = TRANSFER_OUTPUT_OPTION,
config_path: Path = TRANSFER_CONFIG_PATH_OPTION,
) -> None:
"""Evaluate a saved sentiment model on Amazon polarity reviews."""
config = (
AmazonTransferEvaluationConfig.from_json(config_path)
if config_path is not None
else AmazonTransferEvaluationConfig()
)
amazon_records = load_amazon_polarity_reviews(
split=str(config.dataset_split),
sample_size=config.dataset_sample_size,
seed=config.seed,
include_title=config.include_title,
dataset_name=config.dataset_name,
)
predictor = load_sentiment_predictor(
model_path=Path(config.sentiment_model_path).resolve(),
backend=config.sentiment_backend,
max_length=config.sentiment_max_length,
)
artifact = evaluate_reviews_with_predictor(
review_records=amazon_records,
predictor=predictor,
dataset_info={
"dataset_name": config.dataset_name,
"split": config.dataset_split,
"sample_size": config.dataset_sample_size,
"include_title": config.include_title,
"seed": config.seed,
},
max_error_examples=config.max_error_examples,
)
write_json(output_path, artifact.to_dict())
typer.echo(f"Wrote Amazon transfer evaluation artifact to {output_path}")
@app.command("evaluate-local-feedback")
def evaluate_local_feedback(
output_path: Path = LOCAL_EVAL_OUTPUT_OPTION,
config_path: Path = LOCAL_EVAL_CONFIG_PATH_OPTION,
) -> None:
"""Evaluate a saved model on a fixed local labeled customer-feedback CSV."""
config = (
LocalEvaluationConfig.from_json(config_path)
if config_path is not None
else LocalEvaluationConfig()
)
local_records = load_local_labeled_reviews(
dataset_path=Path(config.dataset_path),
text_column=config.text_column,
title_column=config.title_column,
label_column=config.label_column,
review_id_column=config.review_id_column,
split_name=config.split_name,
source_name=config.source_name,
)
predictor = load_sentiment_predictor(
model_path=Path(config.sentiment_model_path).resolve(),
backend=config.sentiment_backend,
max_length=config.sentiment_max_length,
)
artifact = evaluate_reviews_with_predictor(
review_records=local_records,
predictor=predictor,
dataset_info={
"dataset_name": config.source_name,
"split": config.split_name,
"dataset_path": config.dataset_path,
},
max_error_examples=config.max_error_examples,
)
write_json(output_path, artifact.to_dict())
typer.echo(f"Wrote local feedback evaluation artifact to {output_path}")
@app.command("launch-demo")
def launch_demo(
base_path: Path = BASE_PATH_OPTION,
config_path: Path = ANALYSIS_CONFIG_PATH_OPTION,
host: str = HOST_OPTION,
port: int = PORT_OPTION,
share: bool = SHARE_OPTION,
) -> None:
"""Launch the Gradio feedback-intelligence demo."""
analysis_config = (
ReviewAnalysisConfig.from_json(config_path)
if config_path is not None
else ReviewAnalysisConfig()
)
demo = create_demo(base_path=base_path, analysis_config=analysis_config)
demo.launch(server_name=host, server_port=port, share=share)
|