Blame for tasks/leetcode/tests/oracle-runtime/clone/leetcode_app/query.py

"""Problem-set search, filtering, sorting and paging.

The catalogue lives in ``leetcode_questions`` and every list the site renders
is produced by one SQL query against it, joined with the signed-in member's
``leetcode_progress`` rows when there is one. Filter state is expressed
entirely in the URL so a filtered list can be linked, refreshed and shared.
"""

from __future__ import annotations

import json
from contextlib import closing
from dataclasses import dataclass, field
from typing import Any, Iterable, Sequence
from urllib.parse import urlencode

from . import content, store


# The captured problem-set page rendered the pinned daily question plus
# questions 1-100 in one uninterrupted list, so one page is exactly that many
# rows. Paging then continues into the rest of the local catalogue.
PER_PAGE = 101
MAX_PAGE = 999

# Every key the query layer accepts. ``frequency`` and ``contest-point`` are
# deliberately absent from the visible sort menu (``content.SORT_OPTIONS``):
# the source prints both behind a Premium lock and captured no values, so the
# clone will not offer them as if they were source facts. They stay orderable
# here because the stored figures are deterministic and the JSON surface is
# expected to sort by any stored column.
SORT_COLUMNS: dict[str, str] = {
    "custom": "q.number",
    "difficulty": "CASE q.difficulty WHEN 'Easy' THEN 1 WHEN 'Medium' THEN 2"
    " ELSE 3 END",
    "acceptance": "q.acceptance",
    "question-id": "q.number",
    "title": "LOWER(q.title)",
    "last-submitted": "COALESCE(p.last_submitted_at, '')",
    "frequency": "q.frequency",
    "contest-point": "q.contest_point",
}

DEFAULT_ORDER: dict[str, str] = {
    "custom": "asc",
    "difficulty": "asc",
    "acceptance": "asc",
    "question-id": "asc",
    "title": "asc",
    "last-submitted": "desc",
    "frequency": "desc",
    "contest-point": "desc",
}


@dataclass
class ProblemQuery:
    category: str = "all"
    search: str = ""
    difficulty: str = ""
    status: str = ""
    topics: tuple[str, ...] = ()
    language: str = ""
    sort: str = "custom"
    order: str = ""
    page: int = 1

    @property
    def effective_order(self) -> str:
        if self.order in {"asc", "desc"}:
            return self.order
        return DEFAULT_ORDER.get(self.sort, "asc")

    @property
    def is_filtered(self) -> bool:
        return bool(
            self.search
            or self.difficulty
            or self.status
            or self.topics
            or self.language
            or self.sort != "custom"
            or self.order
        )

    def path(self) -> str:
        if self.category and self.category != "all":
            return f"/problemset/{self.category}"
        return "/problemset/"

    def as_params(self, **overrides: Any) -> list[tuple[str, str]]:
        values: list[tuple[str, str]] = []
        search = overrides.get("search", self.search)
        difficulty = overrides.get("difficulty", self.difficulty)
        status = overrides.get("status", self.status)
        topics = overrides.get("topics", self.topics)
        language = overrides.get("language", self.language)
        sort = overrides.get("sort", self.sort)
        order = overrides.get("order", self.order)
        page = overrides.get("page", self.page)
        if search:
            values.append(("search", search))
        if difficulty:
            values.append(("difficulty", difficulty))
        if status:
            values.append(("status", status))
        for topic in topics:
            values.append(("topic", topic))
        if language:
            values.append(("language", language))
        if sort and sort != "custom":
            values.append(("sort", sort))
        if order:
            values.append(("order", order))
        if page and int(page) > 1:
            values.append(("page", str(page)))
        return values

    def url(self, **overrides: Any) -> str:
        category = overrides.pop("category", self.category)
        base = (
            f"/problemset/{category}"
            if category and category != "all"
            else "/problemset/"
        )
        params = self.as_params(**overrides)
        return f"{base}?{urlencode(params)}" if params else base


def _first(values: Sequence[str]) -> str:
    for value in values:
        text = str(value).strip()
        if text:
            return text
    return ""


def parse_query(params: Any, category: str = "all") -> ProblemQuery:
    """Build the query state from the request's own URL parameters."""

    getlist = getattr(params, "getlist", None)
    def values(name: str) -> list[str]:
        if getlist is not None:
            return [str(item) for item in getlist(name)]
        raw = params.get(name)
        return [] if raw is None else [str(raw)]

    difficulty = _first(values("difficulty"))
    if difficulty not in {"", "Easy", "Medium", "Hard"}:
        difficulty = ""
    status = _first(values("status")).casefold()
    if status not in {"", "todo", "solved", "attempted"}:
        status = ""
    topics = tuple(
        topic
        for topic in dict.fromkeys(
            item.strip() for item in values("topic") if item.strip()
        )
        if topic in content.ALL_TOPICS
    )
    language = _first(values("language"))
    if language not in content.LANGUAGE_LABELS:
        language = ""
    sort = _first(values("sort")) or "custom"
    if sort not in SORT_COLUMNS:
        sort = "custom"
    order = _first(values("order")).casefold()
    if order not in {"asc", "desc"}:
        order = ""
    try:
        page = int(_first(values("page")) or "1")
    except ValueError:
        page = 1
    page = max(1, min(page, MAX_PAGE))
    valid_categories = {slug for slug, _ in content.CATEGORIES}
    return ProblemQuery(
        category=category if category in valid_categories else "all",
        search=_first(values("search"))[:120],
        difficulty=difficulty,
        status=status,
        topics=topics,
        language=language,
        sort=sort,
        order=order,
        page=page,
    )


@dataclass
class ProblemPage:
    rows: list[dict[str, Any]] = field(default_factory=list)
    total: int = 0
    page: int = 1
    per_page: int = PER_PAGE
    solved: int = 0
    catalogue_total: int = 0

    @property
    def pages(self) -> int:
        return max(1, -(-self.total // self.per_page))

    @property
    def start_index(self) -> int:
        return 0 if not self.total else (self.page - 1) * self.per_page + 1

    @property
    def end_index(self) -> int:
        return min(self.total, self.page * self.per_page)

    @property
    def is_empty(self) -> bool:
        return self.total == 0


def _where(query: ProblemQuery, subject_id: str | None) -> tuple[str, list[Any]]:
    clauses: list[str] = []
    parameters: list[Any] = []
    if query.category != "all":
        clauses.append("q.category = ?")
        parameters.append(query.category)
    if query.search:
        clauses.append("(LOWER(q.title) LIKE ? OR CAST(q.number AS TEXT) LIKE ?)")
        needle = f"%{query.search.casefold()}%"
        parameters.extend([needle, f"%{query.search.strip()}%"])
    if query.difficulty:
        clauses.append("q.difficulty = ?")
        parameters.append(query.difficulty)
    for topic in query.topics:
        clauses.append("q.topics_json LIKE ?")
        parameters.append(f'%"{topic}"%')
    if query.language:
        clauses.append("q.languages_json LIKE ?")
        parameters.append(f'%"{query.language}"%')
    if query.status and subject_id:
        if query.status == "todo":
            clauses.append("p.status IS NULL")
        else:
            clauses.append("p.status = ?")
            parameters.append(query.status)
    elif query.status == "solved" and not subject_id:
        clauses.append("1 = 0")
    return (" AND ".join(clauses) if clauses else "1 = 1"), parameters


def run_query(query: ProblemQuery, subject_id: str | None) -> ProblemPage:
    where, parameters = _where(query, subject_id)
    column = SORT_COLUMNS.get(query.sort, "q.number")
    direction = "ASC" if query.effective_order == "asc" else "DESC"
    join_subject = subject_id or ""
    # The daily question sits above the list, as the captured page rendered it,
    # but only while the visitor has not chosen their own ordering.
    pinning = query.sort in {"custom", "question-id"} and not query.order
    pin = "CASE WHEN q.slug = ? THEN 0 ELSE 1 END, " if pinning else ""
    pin_parameters: list[Any] = [content.DAILY_SLUG] if pinning else []
    with closing(store.connect()) as connection:
        total_row = connection.execute(
            "SELECT COUNT(*) FROM leetcode_questions q"
            " LEFT JOIN leetcode_progress p"
            " ON p.question_slug = q.slug AND p.subject_id = ?"
            f" WHERE {where}",
            [join_subject, *parameters],
        ).fetchone()
        total = int(total_row[0]) if total_row is not None else 0
        pages = max(1, -(-total // PER_PAGE))
        page = min(max(1, query.page), pages)
        rows = connection.execute(
            "SELECT q.*, p.status AS member_status,"
            " p.last_submitted_at AS member_last_submitted"
            " FROM leetcode_questions q"
            " LEFT JOIN leetcode_progress p"
            " ON p.question_slug = q.slug AND p.subject_id = ?"
            f" WHERE {where}"
            f" ORDER BY {pin}{column} {direction}, q.number ASC"
            " LIMIT ? OFFSET ?",
            [
                join_subject,
                *parameters,
                *pin_parameters,
                PER_PAGE,
                (page - 1) * PER_PAGE,
            ],
        ).fetchall()
        catalogue_row = connection.execute(
            "SELECT COUNT(*) FROM leetcode_questions"
        ).fetchone()
        catalogue_total = int(catalogue_row[0]) if catalogue_row is not None else 0
        solved = 0
        if subject_id:
            solved_row = connection.execute(
                "SELECT COUNT(*) FROM leetcode_progress"
                " WHERE subject_id = ? AND status = 'solved'",
                (subject_id,),
            ).fetchone()
            solved = int(solved_row[0]) if solved_row is not None else 0
    return ProblemPage(
        rows=[_public_row(row) for row in rows],
        total=total,
        page=page,
        solved=solved,
        catalogue_total=catalogue_total,
    )


def _public_row(row: Any) -> dict[str, Any]:
    slug = str(row["slug"])
    item = content.question(slug)
    return {
        "slug": slug,
        "number": int(row["number"]),
        "title": row["title"],
        "difficulty": row["difficulty"],
        "difficulty_label": content.difficulty_label(str(row["difficulty"])),
        "category": row["category"],
        "topics": json.loads(row["topics_json"]),
        "languages": json.loads(row["languages_json"]),
        "acceptance": float(row["acceptance"]),
        "frequency": int(row["frequency"]),
        "contest_point": int(row["contest_point"]),
        "status": row["member_status"],
        "last_submitted": row["member_last_submitted"],
        "is_daily": slug == content.DAILY_SLUG,
        "workspace_ready": bool(item and item.workspace_ready),
    }


def random_slug(query: ProblemQuery, subject_id: str | None, seed: str = "") -> str | None:
    """Pick one question from the active filter for the shuffle control."""

    import random

    where, parameters = _where(query, subject_id)
    with closing(store.connect()) as connection:
        slugs = [
            str(row["slug"])
            for row in connection.execute(
                "SELECT q.slug FROM leetcode_questions q"
                " LEFT JOIN leetcode_progress p"
                " ON p.question_slug = q.slug AND p.subject_id = ?"
                f" WHERE {where} ORDER BY q.number",
                [subject_id or "", *parameters],
            )
        ]
    if not slugs:
        return None
    chooser = random.Random(seed) if seed else random.Random()
    return chooser.choice(slugs)


def topic_counts(category: str = "all") -> list[tuple[str, int]]:
    counts: dict[str, int] = {}
    for question in content.iter_questions():
        if category != "all" and question.category != category:
            continue
        for topic in question.topics:
            counts[topic] = counts.get(topic, 0) + 1
    return sorted(counts.items(), key=lambda item: (-item[1], item[0]))


def category_counts() -> dict[str, int]:
    counts: dict[str, int] = {"all": 0}
    for question in content.iter_questions():
        counts["all"] += 1
        counts[question.category] = counts.get(question.category, 0) + 1
    return counts


def iter_pages(page: int, pages: int) -> Iterable[int]:
    window = {1, pages, page, page - 1, page + 1}
    return sorted(number for number in window if 1 <= number <= pages)