Files
sglang/python/sglang/srt/debug_utils/comparator/entrypoint.py

367 lines
12 KiB
Python

from __future__ import annotations
import argparse
import re
import sys
from pathlib import Path
from typing import Any, Iterator, Optional, Union
import polars as pl
from sglang.srt.debug_utils.comparator.aligner.token_aligner.entrypoint import (
TokenAlignerResult,
compute_maybe_token_aligner_result,
)
from sglang.srt.debug_utils.comparator.aligner.token_aligner.smart.aux_loader import (
AUX_NAMES,
)
from sglang.srt.debug_utils.comparator.aligner.token_aligner.smart.types import (
TokenAlignerPlan,
)
from sglang.srt.debug_utils.comparator.bundle_comparator import compare_bundle_pair
from sglang.srt.debug_utils.comparator.bundle_matcher import (
TensorBundleInfo,
match_bundles,
)
from sglang.srt.debug_utils.comparator.display import emit_display_records
from sglang.srt.debug_utils.comparator.meta_overrider import MetaOverrider
from sglang.srt.debug_utils.comparator.output_types import (
ComparisonRecord,
ConfigRecord,
NonTensorRecord,
SkipRecord,
SummaryRecord,
report_sink,
)
from sglang.srt.debug_utils.comparator.per_token_visualizer import (
generate_per_token_heatmap,
)
from sglang.srt.debug_utils.comparator.per_token_visualizer import (
generate_per_token_heatmap,
)
from sglang.srt.debug_utils.comparator.utils import Pair
from sglang.srt.debug_utils.dump_loader import read_meta, read_tokenizer_path
def main() -> None:
args = _parse_args()
sys.exit(run(args))
def run(args: argparse.Namespace) -> int:
report_path: Optional[Path] = _resolve_report_path(args)
report_sink.configure(
output_format=args.output_format,
report_path=report_path,
)
try:
report_sink.add(ConfigRecord.from_args(args))
dfs: Pair[pl.DataFrame] = _read_df(args)
tokenizer: Any = _maybe_load_tokenizer(args)
for label, df, dump_dir in [
("baseline", dfs.x, Path(args.baseline_path)),
("target", dfs.y, Path(args.target_path)),
]:
emit_display_records(
df=df,
dump_dir=dump_dir,
label=label,
tokenizer=tokenizer,
)
ta_result: TokenAlignerResult = compute_maybe_token_aligner_result(args, dfs)
if ta_result.mode == "smart":
dfs = dfs.map(lambda df: df.filter(~pl.col("name").is_in(AUX_NAMES)))
bundle_info_pairs: list[Pair[TensorBundleInfo]] = match_bundles(
dfs=dfs,
skip_keys=_compute_skip_keys(
args, has_token_aligner=ta_result.mode is not None
),
)
viz_output_dir: Optional[Path] = (
Path(args.viz_output_dir) if args.viz_bundle_details else None
)
visualize_per_token: Optional[Path] = (
Path(args.visualize_per_token) if args.visualize_per_token else None
)
meta_overrider: MetaOverrider = MetaOverrider.from_args_and_config(
override_dims=args.override_dims,
override_baseline_dims=args.override_baseline_dims,
override_target_dims=args.override_target_dims,
override_config=(
Path(args.override_config) if args.override_config else None
),
)
comparison_records = _compare_bundle_pairs(
bundle_info_pairs=bundle_info_pairs,
baseline_path=Path(args.baseline_path),
target_path=Path(args.target_path),
token_aligner_mode=ta_result.mode,
token_aligner_plan=ta_result.plan,
diff_threshold=args.diff_threshold,
thd_seq_lens_by_step_pair=ta_result.thd_seq_lens_by_step_pair,
viz_output_dir=viz_output_dir,
compute_per_token=visualize_per_token is not None,
meta_overrider=meta_overrider,
)
summary, skipped_names = _consume_comparison_records(
comparison_records=comparison_records,
visualize_per_token=visualize_per_token,
)
return _compute_exit_code(
summary,
allow_skip_pattern=args.allow_skip_pattern,
skipped_names=skipped_names,
)
finally:
report_sink.close()
if report_path is not None:
print(f"Report: {report_path}", file=sys.stderr)
def _compute_exit_code(
summary: SummaryRecord,
*,
allow_skip_pattern: str,
skipped_names: list[str],
) -> int:
if summary.failed > 0:
return 1
pattern: re.Pattern[str] = re.compile(allow_skip_pattern)
forbidden: list[str] = [n for n in skipped_names if not pattern.fullmatch(n)]
if forbidden:
return 1
return 0
def _resolve_report_path(args: argparse.Namespace) -> Optional[Path]:
if args.report_path is not None:
return Path(args.report_path) if args.report_path else None
return Path(args.target_path) / "comparator_report.jsonl"
def _maybe_load_tokenizer(args: argparse.Namespace) -> Any:
tokenizer_path: Optional[str] = getattr(args, "tokenizer", None)
if tokenizer_path is None:
for directory in [Path(args.baseline_path), Path(args.target_path)]:
tokenizer_path = read_tokenizer_path(directory)
if tokenizer_path is not None:
break
if tokenizer_path is None:
return None
try:
from transformers import AutoTokenizer
return AutoTokenizer.from_pretrained(tokenizer_path)
except Exception:
return None
def _read_df(args: argparse.Namespace) -> Pair[pl.DataFrame]:
df_baseline = read_meta(args.baseline_path)
df_target = read_meta(args.target_path)
df_target = df_target.filter(
(pl.col("step") >= args.start_step) & (pl.col("step") <= args.end_step)
)
if args.filter:
df_target = df_target.filter(pl.col("filename").str.contains(args.filter))
assert all(c in df_target.columns for c in ["rank", "step", "dump_index", "name"])
return Pair(x=df_baseline, y=df_target)
def _compute_skip_keys(args, *, has_token_aligner: bool) -> set[str]:
skip_keys: set[str] = {"dump_index", "filename"}
if args.grouping == "logical":
skip_keys |= {"rank", "recompute_status"}
if has_token_aligner:
skip_keys |= {"step"}
return skip_keys
def _compare_bundle_pairs(
*,
bundle_info_pairs: list[Pair[TensorBundleInfo]],
baseline_path: Path,
target_path: Path,
token_aligner_mode: Optional[str],
token_aligner_plan: Optional[TokenAlignerPlan],
diff_threshold: float,
thd_seq_lens_by_step_pair: Pair[Optional[dict[int, list[int]]]],
viz_output_dir: Optional[Path] = None,
compute_per_token: bool = False,
meta_overrider: Optional[MetaOverrider] = None,
) -> Iterator[Union[ComparisonRecord, SkipRecord, NonTensorRecord]]:
for bundle_info_pair in bundle_info_pairs:
if not bundle_info_pair.y:
continue
name: str = bundle_info_pair.y[0].name
filenames_pair: Pair[list[str]] = bundle_info_pair.map(
lambda infos: [info.filename for info in infos]
)
yield compare_bundle_pair(
name=name,
filenames_pair=filenames_pair,
baseline_path=baseline_path,
target_path=target_path,
token_aligner_mode=token_aligner_mode,
token_aligner_plan=token_aligner_plan,
diff_threshold=diff_threshold,
thd_seq_lens_by_step_pair=thd_seq_lens_by_step_pair,
viz_output_dir=viz_output_dir,
compute_per_token=compute_per_token,
meta_overrider=meta_overrider,
)
def _consume_comparison_records(
*,
comparison_records: Iterator[Union[ComparisonRecord, SkipRecord, NonTensorRecord]],
visualize_per_token: Optional[Path] = None,
) -> tuple[SummaryRecord, list[str]]:
counts: dict[str, int] = {"passed": 0, "failed": 0, "skipped": 0}
collected_comparisons: list[ComparisonRecord] = []
skipped_names: list[str] = []
for record in comparison_records:
counts[record.category] += 1
report_sink.add(record)
if isinstance(record, SkipRecord) and record.category == "skipped":
skipped_names.append(record.name)
if visualize_per_token is not None and isinstance(record, ComparisonRecord):
collected_comparisons.append(record)
summary: SummaryRecord = SummaryRecord(total=sum(counts.values()), **counts)
report_sink.add(summary)
if visualize_per_token is not None and collected_comparisons:
generate_per_token_heatmap(
records=collected_comparisons,
output_path=visualize_per_token,
)
return summary, skipped_names
if visualize_per_token is not None and collected_comparisons:
generate_per_token_heatmap(
records=collected_comparisons,
output_path=visualize_per_token,
)
def _parse_args() -> argparse.Namespace:
parser = argparse.ArgumentParser()
parser.add_argument("--baseline-path", type=str)
parser.add_argument("--target-path", type=str)
parser.add_argument("--start-step", type=int, default=0)
parser.add_argument("--end-step", type=int, default=1000000)
parser.add_argument("--diff-threshold", type=float, default=1e-3)
parser.add_argument(
"--filter", type=str, default=None, help="Regex to filter filenames"
)
parser.add_argument(
"--output-format",
type=str,
choices=["text", "json"],
default="text",
help="Output format: text (default) or json (JSONL, one JSON object per line)",
)
parser.add_argument(
"--grouping",
type=str,
choices=["logical", "raw"],
default="logical",
help="Grouping mode: logical (cross-rank unshard) or raw (rank-by-rank)",
)
parser.add_argument(
"--token-aligner",
type=str,
choices=["smart", "concat_steps"],
default="concat_steps",
help="Token aligner mode: concat_steps (BS=1, no aux needed) or smart (BS>1, sequence matching)",
)
parser.add_argument(
"--tokenizer",
type=str,
default=None,
help="Tokenizer path for decoding input_ids (auto-discovered from dump metadata if not set)",
)
parser.add_argument(
"--viz-bundle-details",
action="store_true",
default=False,
help="Generate comparison heatmap/histogram PNG for each compared tensor",
)
parser.add_argument(
"--viz-output-dir",
type=str,
default="/tmp/comparator_viz/",
help="Output directory for visualization PNGs (default: /tmp/comparator_viz/)",
)
parser.add_argument(
"--visualize-per-token",
type=str,
default=None,
help="Output path for per-token relative difference heatmap PNG",
)
# Dims override
parser.add_argument(
"--override-dims",
action="append",
default=[],
help="Override dims for both sides: 'name:dims_string' (repeatable)",
)
parser.add_argument(
"--override-baseline-dims",
action="append",
default=[],
help="Override dims for baseline only: 'name:dims_string' (repeatable)",
)
parser.add_argument(
"--override-target-dims",
action="append",
default=[],
help="Override dims for target only: 'name:dims_string' (repeatable)",
)
parser.add_argument(
"--override-config",
type=str,
default=None,
help="Path to YAML override config file (dims overrides, etc.)",
)
parser.add_argument(
"--allow-skip-pattern",
type=str,
default=".*",
help="Regex pattern for tensor names allowed to be skipped. "
"Default '.*' allows all skips. Use '^$' to forbid all skips.",
)
# Report output
parser.add_argument(
"--report-path",
type=str,
default=None,
help="Path for JSONL report (default: <target-path>/comparator_report.jsonl). "
"Pass empty string '' to disable.",
)
return parser.parse_args()