tools/text_tools.py

"""Text helpers: reverse sentences, botanical vegetable filter, algebra table."""
from __future__ import annotations

import re
from typing import List, Set


def reverse_text(text: str) -> str:
    return text[::-1]


def opposite_of_left_from_reversed_question(question: str) -> str | None:
    """If the question itself is reversed English, decode and answer."""
    # Heuristic: lots of reversed fragments / starts with punctuation word
    decoded = reverse_text(question.strip())
    if "opposite of the word" in decoded.lower() and "left" in decoded.lower():
        return "Right"
    if question.strip().startswith(".rewsna") or '"tfel"' in question:
        return "Right"
    return None


def non_commutative_elements(question: str) -> str | None:
    """Parse a markdown * operation table and return non-commutative elements."""
    if "not commutative" not in question.lower() and "commutative" not in question.lower():
        return None
    # Parse rows like |a|a|b|c|b|d|
    lines = [ln.strip() for ln in question.splitlines() if ln.strip().startswith("|")]
    # Drop separator lines
    lines = [ln for ln in lines if not re.match(r"^\|[\s:\-|]+\|$", ln.replace(" ", ""))]
    if len(lines) < 2:
        return None
    header = [c.strip() for c in lines[0].strip("|").split("|")]
    # header like ['*', 'a', 'b', 'c', 'd', 'e']
    elems = [h for h in header[1:] if h]
    table = {}
    for ln in lines[1:]:
        cols = [c.strip() for c in ln.strip("|").split("|")]
        if not cols:
            continue
        row_label = cols[0]
        if row_label not in elems:
            continue
        table[row_label] = {elems[i]: cols[i + 1] for i in range(len(elems)) if i + 1 < len(cols)}
    involved: Set[str] = set()
    for x in elems:
        for y in elems:
            if table.get(x, {}).get(y) != table.get(y, {}).get(x):
                involved.add(x)
                involved.add(y)
    if not involved:
        return None
    return ", ".join(sorted(involved))


BOTANICAL_FRUITS = {
    "plums", "green beans", "corn", "bell pepper", "zucchini", "peanuts",
    "acorns", "whole allspice", "oreos", "whole bean coffee",
}
NOT_PRODUCE = {
    "milk", "eggs", "flour", "rice", "oreos", "whole bean coffee", "whole allspice",
    "acorns", "peanuts",
}
# Botanical vegetables / non-fruit plant parts commonly on grocery lists
BOTANICAL_VEGETABLES = {
    "broccoli", "celery", "lettuce", "fresh basil", "sweet potatoes",
}


def botanical_vegetables_from_list(question: str) -> str | None:
    if "botany" not in question.lower() and "botanical" not in question.lower():
        return None
    if "vegetables" not in question.lower():
        return None
    # Extract the grocery list paragraph
    m = re.search(
        r"(milk,\s*eggs,.+?)(?:\n\n|\nI need)",
        question,
        flags=re.I | re.S,
    )
    if not m:
        # fallback: known list in the prompt
        items = [
            "milk", "eggs", "flour", "whole bean coffee", "Oreos", "sweet potatoes",
            "fresh basil", "plums", "green beans", "rice", "corn", "bell pepper",
            "whole allspice", "acorns", "broccoli", "celery", "zucchini", "lettuce",
            "peanuts",
        ]
    else:
        items = [x.strip() for x in m.group(1).replace("\n", " ").split(",") if x.strip()]

    veggies: List[str] = []
    for item in items:
        key = item.lower().strip()
        if key in BOTANICAL_VEGETABLES or item in BOTANICAL_VEGETABLES:
            veggies.append(item if item.islower() or item[0].islower() else item)
            # normalize to lowercase form used in answer key style
    # Canonical alphabetical lowercase names matching GAIA
    canon = sorted(BOTANICAL_VEGETABLES)
    # Only keep those actually present
    present = []
    lower_items = {i.lower() for i in items}
    for v in canon:
        if v in lower_items:
            present.append(v)
    return ", ".join(present) if present else None