kestrove-core/crates/kv-import/src/csv.rs

//! CSV import: 1Password's CSV export (`Title,Url,Username,Password,OTPAuth,Favorite,Archived,Tags,Notes`)
//! and any CSV whose header names columns like `title,url,username,password,otp,notes`
//! (case-insensitive; see [`column_for`]). Columns with no field of their own go into the notes.
//!
//! RFC 4180 parsing: quoted fields, `""` escapes, embedded newlines, CRLF or LF, a UTF-8 BOM.

use kv_model::ItemKind;
use zeroize::Zeroize;

use crate::{malformed, new_item, set_opt, Extras, ImportError, Limits, Parsed, Result};

#[derive(Debug, Clone, Copy, PartialEq, Eq)]
enum Column {
    Title,
    Url,
    Username,
    Password,
    Otp,
    Notes,
    Favorite,
    Tags,
    Archived,
    Other,
}

/// Header aliases (lowercased, non-alphanumerics removed). Covers 1Password 7/8, Bitwarden, Chrome.
fn column_for(header: &str) -> Column {
    let h: String = header.chars().filter(|c| c.is_alphanumeric()).flat_map(char::to_lowercase).collect();
    match h.as_str() {
        "title" | "name" => Column::Title,
        "url" | "urls" | "website" | "loginuri" | "loginurl" | "uri" => Column::Url,
        "username" | "loginusername" | "user" | "login" => Column::Username,
        "password" | "loginpassword" => Column::Password,
        "otp" | "otpauth" | "totp" | "logintotp" | "onetimepassword" => Column::Otp,
        "notes" | "note" | "notesplain" | "comments" => Column::Notes,
        "favorite" | "favourite" | "fav" => Column::Favorite,
        "tags" | "folder" => Column::Tags,
        "archived" => Column::Archived,
        _ => Column::Other,
    }
}

/// A parsed CSV whose cells are wiped on drop.
pub struct Records(pub Vec<Vec<String>>);

impl Drop for Records {
    fn drop(&mut self) {
        for row in &mut self.0 {
            row.iter_mut().for_each(Zeroize::zeroize);
        }
    }
}

/// RFC 4180 reader. Blank lines are skipped. An unterminated quote is an error.
pub fn read_records(text: &str) -> Result<Records> {
    let mut rows = Records(vec![]);
    let mut row: Vec<String> = vec![];
    let mut field = String::new();
    let mut in_quotes = false;
    let mut chars = text.chars().peekable();
    while let Some(c) = chars.next() {
        if in_quotes {
            match c {
                '"' if chars.peek() == Some(&'"') => {
                    chars.next();
                    field.push('"');
                }
                '"' => in_quotes = false,
                _ => field.push(c),
            }
            continue;
        }
        match c {
            '"' if field.is_empty() => in_quotes = true,
            ',' => row.push(std::mem::take(&mut field)),
            '\r' if chars.peek() == Some(&'\n') => {}
            '\n' | '\r' => {
                row.push(std::mem::take(&mut field));
                push_row(&mut rows.0, std::mem::take(&mut row));
            }
            _ => field.push(c),
        }
    }
    if in_quotes {
        field.zeroize();
        row.iter_mut().for_each(Zeroize::zeroize);
        return Err(malformed("CSV has an unterminated quoted field"));
    }
    if !field.is_empty() || !row.is_empty() {
        row.push(field);
        push_row(&mut rows.0, row);
    }
    Ok(rows)
}

fn push_row(rows: &mut Vec<Vec<String>>, mut row: Vec<String>) {
    if row.iter().all(|f| f.trim().is_empty()) {
        row.iter_mut().for_each(Zeroize::zeroize);
    } else {
        rows.push(row);
    }
}

fn truthy(s: &str) -> bool {
    matches!(s.trim().to_lowercase().as_str(), "1" | "true" | "yes" | "y" | "x")
}

pub fn parse(bytes: &[u8], now: i64, limits: &Limits) -> Result<Parsed> {
    let bytes = bytes.strip_prefix(b"\xEF\xBB\xBF").unwrap_or(bytes);
    let text = std::str::from_utf8(bytes).map_err(|_| malformed("CSV is not UTF-8"))?;
    let records = read_records(text)?;
    let Some((header, rows)) = records.0.split_first() else { return Err(malformed("CSV is empty")) };
    let columns: Vec<Column> = header.iter().map(|h| column_for(h)).collect();
    let has = |c: Column| columns.contains(&c);
    if !(has(Column::Title) || has(Column::Url))
        || !(has(Column::Password) || has(Column::Username) || has(Column::Notes))
    {
        return Err(ImportError::Unsupported(
            "CSV header must name a title or url column and a username, password or notes column".into(),
        ));
    }
    limits.check_items(rows.len())?;

    let mut out = Parsed::default();
    let mut archived = 0u32;
    for row in rows {
        let cell = |c: Column| columns.iter().position(|x| *x == c).and_then(|i| row.get(i)).map_or("", |s| s.as_str());
        let mut item = new_item(ItemKind::Login, cell(Column::Title), now);
        if cell(Column::Title).trim().is_empty() {
            if let Some(host) =
                url::Url::parse(cell(Column::Url).trim()).ok().and_then(|u| u.host_str().map(String::from))
            {
                item.title = host;
            }
        }
        set_opt(&mut item.username, cell(Column::Username));
        set_opt(&mut item.password, cell(Column::Password));
        set_opt(&mut item.totp, cell(Column::Otp).trim());
        set_opt(&mut item.notes, cell(Column::Notes));
        item.urls = cell(Column::Url).lines().map(str::trim).filter(|u| !u.is_empty()).map(String::from).collect();
        item.favorite = truthy(cell(Column::Favorite));
        item.tags =
            cell(Column::Tags).split([',', ';']).map(str::trim).filter(|t| !t.is_empty()).map(String::from).collect();
        if truthy(cell(Column::Archived)) {
            item.tags.push("Archived".into());
            archived += 1;
        }
        let mut extras = Extras::default();
        for (i, value) in row.iter().enumerate() {
            match columns.get(i) {
                Some(Column::Other) => extras.push(&header[i], value),
                Some(_) => {}
                None => extras.push(&format!("Column {}", i + 1), value),
            }
        }
        extras.append_to(&mut item, "— Imported fields —");
        out.items.push(item);
    }
    if archived > 0 {
        out.warnings.push(format!("{archived} archived item(s) were imported with the tag \"Archived\""));
    }
    Ok(out)
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn rfc4180() {
        let r = read_records("a,\"b,c\",\"d \"\"q\"\"\"\r\n\r\n\"multi\nline\",,x\n").unwrap();
        assert_eq!(r.0, vec![vec!["a", "b,c", "d \"q\""], vec!["multi\nline", "", "x"]]);
        let r = read_records("a,b").unwrap();
        assert_eq!(r.0, vec![vec!["a", "b"]]);
        assert!(read_records("a,\"open").is_err());
    }

    #[test]
    fn header_aliases() {
        assert_eq!(column_for("Login_URI"), Column::Url);
        assert_eq!(column_for("OTPAuth"), Column::Otp);
        assert_eq!(column_for(" Title "), Column::Title);
        assert_eq!(column_for("Security question"), Column::Other);
    }
}