mirror of
https://github.com/MODSetter/SurfSense.git
synced 2026-05-17 18:35:19 +02:00
9bcd50164d
Adds the full parser_compare experiment for the multimodal_doc suite:
six arms compared on 30 PDFs / 171 questions from MMLongBench-Doc with
anthropic/claude-sonnet-4.5 across the board.
Source code:
- core/parsers/{azure_di,llamacloud,pdf_pages}.py: direct parser SDK
callers (Azure Document Intelligence prebuilt-read/layout, LlamaParse
parse_page_with_llm/parse_page_with_agent) used by the LC arms,
bypassing the SurfSense backend so each (basic/premium) extraction
is a clean A/B independent of backend ETL routing.
- suites/multimodal_doc/parser_compare/{ingest,runner,prompt}.py:
six-arm benchmark (native_pdf, azure_basic_lc, azure_premium_lc,
llamacloud_basic_lc, llamacloud_premium_lc, surfsense_agentic) with
byte-identical prompts per question, deterministic grader, Wilson
CIs, and the per-page preprocessing tariff cost overlay.
Reproducibility:
- pyproject.toml + uv.lock pin pypdf, azure-ai-documentintelligence,
llama-cloud-services as new deps.
- .env.example documents the AZURE_DI_* and LLAMA_CLOUD_API_KEY env
vars now required for parser_compare.
- 12 analysis scripts under scripts/: retry pass with exponential
backoff, post-retry accuracy merge, McNemar / latency / per-PDF
stats, context-overflow hypothesis test, etc. Each produces one
number cited by the blog report.
Citation surface:
- reports/blog/multimodal_doc_parser_compare_n171_report.md: 1219-line
technical writeup (16 sections) covering headline accuracy, per-format
accuracy, McNemar pairwise significance, latency / token / per-PDF
distributions, error analysis, retry experiment, post-retry final
accuracy, cost amortization model with closed-form derivation, threats
to validity, and reproducibility appendix.
- data/multimodal_doc/runs/2026-05-14T00-53-19Z/parser_compare/{raw,
raw_retries,raw_post_retry}.jsonl + run_artifact.json + retry summary
whitelisted via data/.gitignore as the verifiable numbers source.
Gitignore:
- ignore logs_*.txt + retry_run.log; structured artifacts cover the
citation surface, debug logs are noise.
- data/.gitignore default-ignores everything, whitelists the n=171 run
artifacts only (parser manifest left ignored to avoid leaking local
Windows usernames in absolute paths; manifest is fully regenerable
via 'ingest multimodal_doc parser_compare').
- reports/.gitignore now whitelists hand-curated reports/blog/.
Also retires the abandoned CRAG Task 3 implementation (download script,
streaming Task 3 ingest, CragTask3Benchmark + tests) and trims the
runner / ingest module APIs to match.
Co-authored-by: Cursor <cursoragent@cursor.com>
77 lines
2.3 KiB
Python
77 lines
2.3 KiB
Python
"""Query SurfSense for the status of every MMLongBench PDF in scope.
|
|
|
|
Uses the existing SurfSense documents client to query
|
|
``/documents/status?document_ids=...`` for both the known-existing 5
|
|
PDFs (doc ids 5219-5223) and the recently-uploaded mmlongbench batch
|
|
(7577-7600 range).
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
import asyncio
|
|
import os
|
|
from pathlib import Path
|
|
|
|
import httpx
|
|
from dotenv import load_dotenv
|
|
|
|
|
|
REPO = Path(__file__).resolve().parents[1]
|
|
PDF_DIR = REPO / "data" / "multimodal_doc" / "mmlongbench" / "pdfs"
|
|
|
|
|
|
async def main() -> None:
|
|
load_dotenv(REPO / ".env")
|
|
base = os.environ.get("SURFSENSE_API_BASE", "http://localhost:8000").rstrip("/")
|
|
token = os.environ.get("SURFSENSE_JWT")
|
|
if not token:
|
|
raise SystemExit("SURFSENSE_JWT missing from .env")
|
|
|
|
pdf_names = sorted(p.name for p in PDF_DIR.glob("*.pdf"))
|
|
print(f"local cached PDFs: {len(pdf_names)}")
|
|
|
|
candidate_ids = list(range(5219, 5224)) + list(range(7577, 7625))
|
|
|
|
headers = {
|
|
"Authorization": f"Bearer {token}",
|
|
"Accept": "application/json",
|
|
}
|
|
async with httpx.AsyncClient(timeout=30.0) as http:
|
|
r = await http.get(
|
|
f"{base}/api/v1/documents/status",
|
|
params={
|
|
"search_space_id": 55,
|
|
"document_ids": ",".join(str(d) for d in candidate_ids),
|
|
},
|
|
headers=headers,
|
|
)
|
|
r.raise_for_status()
|
|
items = r.json().get("items", [])
|
|
|
|
by_title: dict[str, dict] = {}
|
|
for it in items:
|
|
by_title[it.get("title", "")] = {
|
|
"id": it.get("id"),
|
|
"state": (it.get("status") or {}).get("state"),
|
|
"reason": (it.get("status") or {}).get("reason"),
|
|
}
|
|
|
|
by_state: dict[str, int] = {}
|
|
print()
|
|
for name in pdf_names:
|
|
info = by_title.get(name)
|
|
if info is None:
|
|
print(f" [missing ] {name}")
|
|
by_state["missing"] = by_state.get("missing", 0) + 1
|
|
else:
|
|
tag = info["state"] or "?"
|
|
print(f" [{tag:13s}] doc_id={info['id']:>5} {name}")
|
|
by_state[tag] = by_state.get(tag, 0) + 1
|
|
print()
|
|
print("summary:")
|
|
for k, v in sorted(by_state.items()):
|
|
print(f" {k}: {v}")
|
|
|
|
|
|
if __name__ == "__main__":
|
|
asyncio.run(main())
|