/** Picks a random tooltip string from a list of options */
export function randomTooltip(tooltips: string[]): string {
    return tooltips[Math.floor(Math.random() * tooltips.length)];
}


/** Minimal RFC4180-style split: handles quoted fields that contain the
 * delimiter or escaped quotes (""). Good enough for CSV exports that
 * quote string columns (e.g. a designation column with a comma in it). */
export function splitCSVLine(line: string): string[] {
    const result: string[] = [];
    let cur = "";
    let inQuotes = false;
    for (let i = 0; i < line.length; i++) {
        const c = line[i];
        if (inQuotes) {
            if (c === '"') {
                if (line[i + 1] === '"') { cur += '"'; i++; }
                else inQuotes = false;
            } else {
                cur += c;
            }
        } else {
            if (c === '"') inQuotes = true;
            else if (c === ",") { result.push(cur.trim()); cur = ""; }
            else cur += c;
        }
    }
    result.push(cur.trim());
    return result;
}

/**
 * FORTRAN-style format codes VizieR uses: F/E/D = floating point, I =
 * integer, A/a = text -- so the first letter alone tells us the type.
 *
 * NB: some "numeric" columns don't round-trip safely through JS numbers,
 * most notably Gaia's `Source` id (format I19, up to 19 digits) which
 * exceeds Number.MAX_SAFE_INTEGER. Values are always kept as raw strings
 * in `mainData`, so nothing is lost here -- just be careful not to pipe
 * an id column through Number()/parseFloat() downstream if you need
 * exact equality; compare the strings (or use BigInt) instead.
 */
function classifyFormat(format: string): "number" | "string" {
    const letter = format[0]?.toUpperCase();
    return letter === "F" || letter === "E" || letter === "D" || letter === "I" ? "number" : "string";
}


/**
 * Reads the `#Column NAME (FORMAT) DESCRIPTION [ucd=...]` lines VizieR
 * puts above the table. These are always tab-separated regardless of the
 * table's own delimiter, which is what makes them a reliable source for
 * descriptions and types even when we're not sure how the table below
 * is delimited yet.
 */
export function parseColumnComments(rawRows: string[]): Map<string, { description: string; type: "number" | "string" }> {
    const docs = new Map<string, { description: string; type: "number" | "string" }>();
    for (const line of rawRows) {
        if (!line.startsWith("#Column\t")) continue;
        const parts = line.split("\t");
        if (parts.length < 4) continue;

        const name = parts[1].trim();
        const format = parts[2].trim().replace(/^\(|\)$/g, "");
        let description = parts[3].trim();
        if (description.startsWith("?")) description = description.slice(1).trim(); // "?" marks a nullable field

        docs.set(name, { description, type: classifyFormat(format) });
    }
    return docs;
}


/** A column is "number" only if every non-empty value in it parses as
 * a finite number; empty/missing values don't disqualify it. */
export function inferType(values: string[]): "number" | "string" {
    const nonEmpty = values.map(v => v.trim()).filter(v => v.length > 0);
    if (nonEmpty.length === 0) return "string";
    return nonEmpty.every(v => Number.isFinite(Number(v))) ? "number" : "string";
}


/** Picks whichever candidate splits the header row into the most fields. */
export function detectDelimiter(headerLine: string): string {
    const candidates = ["\t", "|", ";", ","];
    let best = candidates[0];
    let bestCount = 1;
    for (const d of candidates) {
        const count = headerLine.split(d).length;
        if (count > bestCount) {
            bestCount = count;
            best = d;
        }
    }
    return best;
}

/** True for VizieR's "-------\t-------\t..." column-width separator row. */
export function isDashRow(line: string): boolean {
    return line.includes("-") && /^[-\s|;,\t]+$/.test(line);
}