File size: 11,712 Bytes
178b774 | 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 | """
X-Box Pipeline CLI β orchestrates the full analysis pipeline.
Usage:
python -m xbox.cli analyze --archive /path/to/tweets.xlsx --senator "Mike Lee"
python -m xbox.cli analyze --hf-dataset --all-senators
python -m xbox.cli fetch-handles
python -m xbox.cli list-models
"""
import json
import logging
import sys
from pathlib import Path
import click
from rich.console import Console
from rich.logging import RichHandler
from rich.progress import Progress, SpinnerColumn, TextColumn
from rich.table import Table
from .config import (
CLASSIFIER_MODELS,
EMBEDDING_MODEL,
OUTPUT_DIR,
TOXICITY_MODEL,
)
console = Console()
def setup_logging(verbose: bool = False):
level = logging.DEBUG if verbose else logging.INFO
logging.basicConfig(
level=level,
format="%(message)s",
handlers=[RichHandler(console=console, show_time=False, show_path=False)],
)
@click.group()
@click.option("--verbose", "-v", is_flag=True, help="Enable debug logging")
def cli(verbose):
"""X-Box Pipeline β Classifier-based tweet analysis for political accounts."""
setup_logging(verbose)
@cli.command()
def fetch_handles():
"""Fetch current US senator Twitter/X handles."""
from .data import fetch_senator_handles
df = fetch_senator_handles(cache=True)
table = Table(title=f"US Senators with Twitter Handles ({len(df)} found)")
table.add_column("Name", style="cyan")
table.add_column("Party", style="green")
table.add_column("State")
table.add_column("Handle", style="yellow")
for _, row in df.iterrows():
table.add_row(
f"{row['first_name']} {row['last_name']}",
row["party"],
row["state"],
f"@{row['twitter_handle']}",
)
console.print(table)
@cli.command()
def list_models():
"""List all models used in the pipeline."""
table = Table(title="Pipeline Models")
table.add_column("Component", style="cyan")
table.add_column("Model ID", style="yellow")
table.add_column("~Params", style="green")
table.add_row("Embeddings", EMBEDDING_MODEL, "600M")
for name, model_id in CLASSIFIER_MODELS.items():
table.add_row(f"Classifier ({name})", model_id, "~125M")
table.add_row("Toxicity", TOXICITY_MODEL, "~355M")
console.print(table)
@cli.command()
@click.option("--archive", "-a", type=click.Path(exists=True), help="Path to tweet archive (xlsx/csv/json)")
@click.option("--senator", "-s", type=str, default="", help="Senator name for labeling")
@click.option("--handle", "-h", type=str, default="", help="Twitter handle for labeling")
@click.option("--hf-dataset", is_flag=True, help="Also load HuggingFace senator-tweets dataset")
@click.option("--output-dir", "-o", type=click.Path(), default=None, help="Output directory")
@click.option("--skip-embeddings", is_flag=True, help="Skip embedding generation (faster)")
@click.option("--batch-size", "-b", type=int, default=32, help="Classification batch size")
@click.option("--party", type=str, default="", help="Party affiliation")
@click.option("--state", type=str, default="", help="State")
def analyze(archive, senator, handle, hf_dataset, output_dir, skip_embeddings, batch_size, party, state):
"""
Run the full analysis pipeline on a tweet dataset.
Example:
python -m xbox.cli analyze -a /mnt/c/x_box/BasedMikeLee_full_archive.xlsx -s "Mike Lee" -h "BasedMikeLee"
"""
from .behavioral import BehavioralAnalyzer
from .classifiers import MultiHeadClassifier
from .data import load_local_archive, load_hf_senator_tweets
from .embeddings import TweetEmbedder
from .fusion import ScoreFusion
from .report import generate_json_report, generate_markdown_report
out_dir = Path(output_dir) if output_dir else OUTPUT_DIR
out_dir.mkdir(parents=True, exist_ok=True)
# ββ Step 1: Load data βββββββββββββββββββββββββββββ
console.print("\n[bold cyan]Step 1: Loading data...[/]")
frames = []
if archive:
df = load_local_archive(archive, senator_name=senator)
frames.append(df)
if hf_dataset:
hf_df = load_hf_senator_tweets()
frames.append(hf_df)
if not frames:
console.print("[red]No data source specified. Use --archive or --hf-dataset[/]")
sys.exit(1)
import pandas as pd
data = pd.concat(frames, ignore_index=True)
console.print(f" Loaded [green]{len(data):,}[/] tweets")
# ββ Step 2: Behavioral analysis βββββββββββββββββββ
console.print("\n[bold cyan]Step 2: Behavioral analysis...[/]")
analyzer = BehavioralAnalyzer()
behavioral = analyzer.analyze(data, senator_name=senator, twitter_handle=handle)
console.print(f" Compulsion score: [yellow]{behavioral.compulsion_score}/100[/]")
# ββ Step 3: Text classification βββββββββββββββββββ
console.print("\n[bold cyan]Step 3: Multi-head classification...[/]")
classifier = MultiHeadClassifier()
if "text" not in data.columns:
console.print("[red]No 'text' column found in data[/]")
sys.exit(1)
classified = classifier.classify_tweets(data, batch_size=batch_size)
console.print(f" Classified [green]{len(classified):,}[/] tweets across {len(CLASSIFIER_MODELS) + 1} heads")
# ββ Step 4: Embeddings (optional) βββββββββββββββββ
if not skip_embeddings:
console.print("\n[bold cyan]Step 4: Generating embeddings...[/]")
embedder = TweetEmbedder()
emb_path = str(out_dir / f"{handle or senator or 'tweets'}_embeddings.npy")
embeddings = embedder.embed_dataframe(classified, save_path=emb_path)
console.print(f" Generated embeddings: shape {embeddings.shape}")
else:
console.print("\n[dim]Step 4: Skipping embeddings (--skip-embeddings)[/]")
# ββ Step 5: Score fusion ββββββββββββββββββββββββββ
console.print("\n[bold cyan]Step 5: Score fusion...[/]")
fusion = ScoreFusion()
classified = fusion.compute_tweet_virulence(classified)
profile = fusion.aggregate_senator_profile(
classified, behavioral,
senator_name=senator,
twitter_handle=handle,
party=party,
state=state,
)
console.print(f" Virulence score: [yellow]{profile.virulence_score}/100[/]")
console.print(f" Overall risk: [bold red]{profile.overall_risk_score}/100[/]")
# ββ Step 6: Report generation βββββββββββββββββββββ
console.print("\n[bold cyan]Step 6: Generating reports...[/]")
slug = handle or senator.replace(" ", "_") or "analysis"
json_path = str(out_dir / f"{slug}_report.json")
md_path = str(out_dir / f"{slug}_report.md")
generate_json_report(profile, output_path=json_path)
generate_markdown_report(profile, output_path=md_path)
# Save classified tweets β coerce problematic mixed-type columns to string
for col in classified.columns:
if classified[col].dtype == object:
classified[col] = classified[col].astype(str)
classified_path = str(out_dir / f"{slug}_classified_tweets.parquet")
classified.to_parquet(classified_path, index=False)
console.print(f"\n[bold green]Done![/]")
console.print(f" JSON report: {json_path}")
console.print(f" MD report: {md_path}")
console.print(f" Tweets data: {classified_path}")
@cli.command()
@click.option("--output-dir", "-o", type=click.Path(), default=None)
def batch_analyze(output_dir):
"""
Batch analyze all senators with available data.
Fetches handles, loads available datasets, runs pipeline for each.
"""
from .data import fetch_senator_handles, load_hf_senator_tweets
from .behavioral import BehavioralAnalyzer
from .classifiers import MultiHeadClassifier
from .fusion import ScoreFusion
from .report import generate_json_report, generate_markdown_report
out_dir = Path(output_dir) if output_dir else OUTPUT_DIR / "batch"
out_dir.mkdir(parents=True, exist_ok=True)
console.print("[bold cyan]Loading senator handles...[/]")
handles = fetch_senator_handles()
console.print("[bold cyan]Loading HuggingFace senator tweets...[/]")
try:
hf_data = load_hf_senator_tweets()
except Exception as e:
console.print(f"[yellow]Could not load HF data: {e}[/]")
hf_data = None
# Load models once
console.print("[bold cyan]Loading classification models...[/]")
classifier = MultiHeadClassifier()
classifier.load_all()
fusion = ScoreFusion()
analyzer = BehavioralAnalyzer()
results = []
import pandas as pd
for _, row in handles.iterrows():
name = f"{row['first_name']} {row['last_name']}"
handle = row["twitter_handle"]
console.print(f"\n[cyan]Analyzing {name} (@{handle})...[/]")
# Filter HF data for this senator
senator_tweets = pd.DataFrame()
if hf_data is not None and "username" in hf_data.columns:
senator_tweets = hf_data[
hf_data["username"].str.lower() == handle.lower()
].copy()
if senator_tweets.empty:
console.print(f" [dim]No tweets found, skipping[/]")
continue
console.print(f" Found {len(senator_tweets)} tweets")
# Run pipeline
behavioral = analyzer.analyze(senator_tweets, senator_name=name, twitter_handle=handle)
classified = classifier.classify_tweets(senator_tweets)
classified = fusion.compute_tweet_virulence(classified)
profile = fusion.aggregate_senator_profile(
classified, behavioral,
senator_name=name,
twitter_handle=handle,
party=row.get("party", ""),
state=row.get("state", ""),
)
# Save reports
slug = handle
generate_json_report(profile, output_path=str(out_dir / f"{slug}_report.json"))
generate_markdown_report(profile, output_path=str(out_dir / f"{slug}_report.md"))
results.append({
"senator": name,
"handle": handle,
"party": row.get("party", ""),
"state": row.get("state", ""),
"compulsion_score": profile.compulsion_score,
"virulence_score": profile.virulence_score,
"overall_risk": profile.overall_risk_score,
"n_tweets": profile.n_tweets_analyzed,
})
# Summary table
if results:
summary = pd.DataFrame(results)
summary = summary.sort_values("overall_risk", ascending=False)
summary.to_csv(str(out_dir / "batch_summary.csv"), index=False)
table = Table(title="Senator Analysis Summary")
table.add_column("Senator")
table.add_column("Party")
table.add_column("State")
table.add_column("Tweets")
table.add_column("Compulsion", justify="right")
table.add_column("Virulence", justify="right")
table.add_column("Overall Risk", justify="right")
for _, r in summary.head(20).iterrows():
table.add_row(
r["senator"],
r["party"],
r["state"],
str(r["n_tweets"]),
f"{r['compulsion_score']:.1f}",
f"{r['virulence_score']:.1f}",
f"{r['overall_risk']:.1f}",
)
console.print(table)
console.print(f"\nFull results saved to {out_dir}/")
if __name__ == "__main__":
cli()
|