legex/harvey.py

"""Ingest Harvey-AI extractions from a single xlsx into per-country JSONL.

A VAULT_REVIEW export (Sheet1) holds Harvey's answers grouped by Folder. Each
question occupies a fixed block of columns: the older export used 4 columns per
question (Value / Reasoning / Citations / Comments); the newer export uses 2
(Answer / Citations). We keep only the Answer column, clean it, match each row
to a Goldenset case_id via the PDF filename, and write JSONL that looks exactly
like `legex.inference` output — so evaluate/analysis treat Harvey as just
another model. The per-question block width is detected from the sheet, so both
export layouts work without per-file configuration.
"""

import argparse
import logging
import re
import sys
from collections import defaultdict
from pathlib import Path

from openpyxl import load_workbook

from legex.config import settings
from legex.inference import _output_columns
from legex.utils import (
    inference_path,
    write_inference_jsonl,
    goldenset_path,
    goldenset_sheet,
    norm_case_id,
)

log = logging.getLogger(__name__)


HARVEY_FOLDER_TO_CC: dict[str, str] = {
    "Albania": "al",
    "Armenia": "am",
    "Australia": "au",
    "Austria": "at",
    "Belgium": "be",
    "Brazil": "br",
    "China": "cn",
    "Dominican_Republic": "do",
    "France": "fr",
    "Georgia": "ge",
    "Germany": "de",
    "Hong_Kong": "hk",
    "India": "in",
    "Italy": "it",
    "Korea": "kr",
    "Kosovo": "xk",
    "Liechtenstein": "li",
    "Luxembourg": "lu",
    "Nepal": "np",
    "New_Zealand": "nz",
    "Philippines": "ph",
    "Russia": "ru",
    "Schweiz_final": "ch",
    "Switzerland": "ch",
    "Serbia": "rs",
    "Singapore": "sg",
    "South_Korea fixed": "kr",
    "Spain": "es",
    "Taiwan": "tw",
    "Ukraine": "ua",
    "United_Kingdom": "uk",
    "United_States": "us",
}

# The 12 review-table questions, in export order. Column 0 of each row is the
# PDF name and column 1 the Folder. Thhe first question (case_id) starts at
# column index 3 and each subsequent question sits one block-width further on.
# We match rows by PDF filename with the Goldensets.
HARVEY_FIELDS_ORDER: tuple[str, ...] = (
    "case_id",
    "legal_subject_judgement",
    "trial_start_date",
    "trial_end_date",
    "dispute_value_nominal",
    "plaintiff_loosing_share",
    "court_cost_awarded_nominal",
    "party_compensation_awarded_nominal",
    "plaintiffs_all_count",
    "defendants_all_count",
    "plaintiff_no1_ISIC1_industry_category",
    "defendant_no1_ISIC1_industry_category",
)
_FIRST_ANSWER_COL = 3  # After Name (0), Folder (1), Document Classification (2)

_CITATION_RE = re.compile(r"\s*(?:\[\d+\])+\s*$")
_EMPTY_LITERALS = {"", "—"}
_DATE_IN_NAME_RE = re.compile(r"(\d{4}-\d{2}-\d{2})")


def _infer_date_from_name(xlsx: Path) -> str | None:
    m = _DATE_IN_NAME_RE.search(xlsx.name)
    return m.group(1) if m else None


def _folder_to_cc(folder: str) -> str | None:
    """Map a VAULT export folder to a country code, tolerating the ``(new)``
    suffix and spaces (e.g. ``'Hong Kong (new)'`` -> ``hk``)."""
    base = re.sub(r"\s*\(new\)\s*$", "", str(folder)).strip()
    return HARVEY_FOLDER_TO_CC.get(base) or HARVEY_FOLDER_TO_CC.get(base.replace(" ", "_"))


def _value_columns(n_cols: int) -> dict[int, str]:
    """Answer-column index -> Goldenset field, for either export layout.

    Block width per question = (n_cols - 3) / 12 (4 for the old Value/Reasoning/
    Citations/Comments layout, 2 for the new Answer/Citations layout)."""
    width = (n_cols - _FIRST_ANSWER_COL) // len(HARVEY_FIELDS_ORDER)
    if width < 1:
        raise ValueError(f"unexpected Harvey sheet width: {n_cols} columns")
    return {
        _FIRST_ANSWER_COL + j * width: field
        for j, field in enumerate(HARVEY_FIELDS_ORDER)
        if field != "case_id"
    }


def _clean(value: object) -> str:
    if value is None:
        return ""
    s = str(value).strip()
    if s in _EMPTY_LITERALS:
        return ""
    s = _CITATION_RE.sub("", s).strip()
    if s.lower() == "nonpecuniary":
        return "nonpecuniary"
    return s


def _gold_case_id_index(cc: str) -> dict[str, str] | None:
    """{ norm_case_id(gold) → gold case_id } for one country, or None if no Goldenset."""
    gs = goldenset_path(cc)
    if not gs.exists():
        return None
    wb = load_workbook(gs, read_only=True, data_only=True)
    ws = goldenset_sheet(wb)
    rows = ws.iter_rows(values_only=True)
    header = [str(c) if c is not None else "" for c in next(rows)]
    try:
        cid_idx = header.index("case_id")
    except ValueError as e:
        raise ValueError(f"{gs}: GOLDENSET sheet has no case_id column") from e
    index: dict[str, str] = {}
    for row in rows:
        if not any(row):
            continue
        raw = row[cid_idx]
        if raw is None:
            continue
        gold = str(raw).strip()
        if not gold:
            continue
        index.setdefault(norm_case_id(gold), gold)
    return index


def ingest(
    xlsx: Path,
    prompt_version: str = "v3",
    source: str = "full_text",
    model: str = "harvey",
    inference_date: str | None = None,
) -> None:
    inference_date = inference_date or _infer_date_from_name(xlsx)
    if not inference_date:
        raise ValueError(f"cannot derive inference date from {xlsx.name}; pass --inference_date")

    columns = _output_columns()
    columns.insert(columns.index("model") + 1, "inference_date")
    wb = load_workbook(xlsx, read_only=True, data_only=True)
    if "Sheet1" not in wb.sheetnames:
        raise ValueError(f"{xlsx} missing Sheet1 (found {wb.sheetnames})")
    ws = wb["Sheet1"]
    col_to_gold = _value_columns(ws.max_column)
    rows_iter = ws.iter_rows(values_only=True)
    next(rows_iter)  # skip header

    by_folder: dict[str, list[tuple]] = defaultdict(list)
    for row in rows_iter:
        if not row or row[0] is None:
            continue
        folder = row[1]
        if folder is None:
            continue
        by_folder[str(folder)].append(row)

    for folder, rows in by_folder.items():
        cc = _folder_to_cc(folder)
        if cc is None:
            log.warning(f"unknown folder {folder!r}, skipping {len(rows)} row(s)")
            continue
        index = _gold_case_id_index(cc)
        if index is None:
            log.info(f"[{cc}] no Goldenset on disk, skipping {len(rows)} Harvey row(s)")
            continue

        out = inference_path(cc, prompt_version, source, model)
        out.parent.mkdir(parents=True, exist_ok=True)

        matched = 0
        unmatched = 0
        out_rows: list[dict] = []
        for row in rows:
            stem = Path(str(row[0])).stem
            key = norm_case_id(stem)
            gold = index.get(key)
            if gold is None:
                # Prefix pass: bundles sometimes shorten filenames
                # `NKP-11263.pdf` vs gold `NKP_11263_(078_NF_0004)`).
                cands = [k for k in index if k.startswith(key + "_")]
                if len(cands) == 1:
                    gold = index[cands[0]]
            if gold is None:
                unmatched += 1
                log.info(f"[{cc}] no Goldenset match for {row[0]!r}")
                continue
            out_row = {col: "" for col in columns}
            out_row["case_id"] = gold
            out_row["model"] = model
            out_row["inference_date"] = inference_date
            for harvey_idx, gold_col in col_to_gold.items():
                if harvey_idx < len(row):
                    out_row[gold_col] = _clean(row[harvey_idx])
            out_rows.append(out_row)
            matched += 1
        write_inference_jsonl(out, out_rows, columns)
        log.info(
            f"[{cc}] wrote {matched} Harvey row(s) → {out} "
            f"({unmatched} unmatched, {len(rows)} total)"
        )


def main() -> None:
    logging.basicConfig(
        level=logging.INFO,
        format="%(asctime)s [%(levelname)s] %(message)s",
        handlers=[logging.StreamHandler(sys.stderr)],
    )
    parser = argparse.ArgumentParser(
        prog="legex-harvey-ingest",
        description="Convert a Harvey VAULT export into per-country Goldenset_*_harvey.jsonl files.",
    )
    parser.add_argument(
        "--xlsx",
        type=Path,
        default=settings.raw_dir / "harvey_2026-06-30.xlsx",
        help="Path to the Harvey VAULT export xlsx (default: data/raw/harvey_2026-06-30.xlsx).",
    )
    parser.add_argument("--prompt_version", default="v3")
    parser.add_argument(
        "--source",
        choices=("full_text", "pdf"),
        default="full_text",
        help="Source bucket label used in the output filename (default: full_text).",
    )
    parser.add_argument("--model", default="harvey", help="Model slug for the output filename.")
    parser.add_argument(
        "--inference_date",
        default=None,
        help="ISO date stamped on every record (default: parsed from the xlsx filename).",
    )
    args = parser.parse_args()
    ingest(
        xlsx=args.xlsx,
        prompt_version=args.prompt_version,
        source=args.source,
        model=args.model,
        inference_date=args.inference_date,
    )


if __name__ == "__main__":
    main()