whattsreply/src/history.rs

//! SQLite-backed store for message history, chats (with autoreply flag) and contacts.

use std::path::Path;

use anyhow::{Context, Result};
use rusqlite::{Connection, OptionalExtension, Row, params};

const TABLE_MESSAGES: &str = "messages";
const TABLE_CHATS: &str = "chats";
const TABLE_CONTACTS: &str = "contacts";
const INDEX_MESSAGES_BY_CHAT: &str = "idx_messages_chat_timestamp";

/// Default `chats.last_timestamp` for chats that have no message yet.
const DEFAULT_LAST_TIMESTAMP: i64 = 0;
/// Default autoreply state of a chat that was never toggled.
const DEFAULT_AUTOREPLY: bool = false;

/// Separator between the user part and the server part of a WhatsApp jid.
const JID_SEPARATOR: char = '@';

/// A message persisted in the history.
#[derive(Debug, Clone, PartialEq)]
pub struct StoredMessage {
    pub message_id: String,
    pub chat_jid: String,
    pub sender_jid: String,
    pub sender_name: Option<String>,
    pub text: String,
    pub timestamp: i64,
    pub from_me: bool,
    pub is_ai: bool,
}

/// A chat as shown in the chat list.
#[derive(Debug, Clone, PartialEq)]
pub struct ChatSummary {
    pub jid: String,
    pub name: String,
    pub last_timestamp: i64,
    pub autoreply: bool,
}

/// A named contact.
#[derive(Debug, Clone, PartialEq)]
pub struct Contact {
    pub jid: String,
    pub name: String,
}

/// Local SQLite message history.
pub struct History {
    connection: Connection,
}

impl History {
    /// Purpose: open (or create) the history database file and ensure the schema exists.
    /// Inputs: `path` of the SQLite file; missing parent directories are created.
    /// Output: a ready-to-use `History`.
    /// Side effects: may create directories and the database file; creates missing tables.
    pub fn open(path: &Path) -> Result<History> {
        if let Some(parent) = path.parent().filter(|parent| !parent.as_os_str().is_empty()) {
            std::fs::create_dir_all(parent)
                .with_context(|| format!("create history directory {}", parent.display()))?;
        }
        let connection = Connection::open(path)
            .with_context(|| format!("open history database {}", path.display()))?;

        Self::from_connection(connection)
    }

    /// Purpose: create a throw-away history living only in memory (tests only).
    /// Inputs: none.
    /// Output: a ready-to-use `History`.
    /// Side effects: creates the schema in the in-memory database.
    #[cfg(test)]
    pub fn open_in_memory() -> Result<History> {
        let connection =
            Connection::open_in_memory().context("open in-memory history database")?;

        Self::from_connection(connection)
    }

    /// Purpose: wrap a connection and create the schema on it.
    /// Inputs: an open SQLite connection.
    /// Output: the `History` owning that connection.
    /// Side effects: executes the `CREATE TABLE IF NOT EXISTS` statements.
    fn from_connection(connection: Connection) -> Result<History> {
        connection
            .execute_batch(&schema_sql())
            .context("create history schema")?;

        Ok(History { connection })
    }

    /// Purpose: persist a message once, keeping the chat row up to date.
    /// Inputs: the message to store (deduplicated by `message_id`).
    /// Output: `true` if the message was newly inserted, `false` if it already existed.
    /// Side effects: inserts into messages; upserts chats (last_timestamp = max, name from
    /// the sender push name when the chat has none and the message is not from me), atomically.
    pub fn save_message(&self, message: &StoredMessage) -> Result<bool> {
        let context = || format!("save message {} in chat {}", message.message_id, message.chat_jid);
        let transaction = self.connection.unchecked_transaction().with_context(context)?;

        let inserted = insert_message(&transaction, message).with_context(context)?;
        if inserted {
            upsert_chat_for_message(&transaction, message).with_context(context)?;
        }
        transaction.commit().with_context(context)?;

        Ok(inserted)
    }

    /// Purpose: fetch the latest messages of a chat.
    /// Inputs: `chat_jid` and the maximum number of messages (`limit`).
    /// Output: up to `limit` most recent messages, in ascending timestamp order.
    /// Side effects: none (read only).
    pub fn recent_messages(&self, chat_jid: &str, limit: usize) -> Result<Vec<StoredMessage>> {
        let context = || format!("read last {limit} messages of chat {chat_jid}");
        let sql = format!(
            "SELECT message_id, chat_jid, sender_jid, sender_name, text, timestamp, from_me, is_ai \
             FROM {TABLE_MESSAGES} WHERE chat_jid = ?1 \
             ORDER BY timestamp DESC, rowid DESC LIMIT ?2"
        );
        let limit = i64::try_from(limit).unwrap_or(i64::MAX);

        let mut statement = self.connection.prepare_cached(&sql).with_context(context)?;
        let mut messages = statement
            .query_map(params![chat_jid, limit], message_from_row)
            .with_context(context)?
            .collect::<rusqlite::Result<Vec<_>>>()
            .with_context(context)?;
        messages.reverse();

        Ok(messages)
    }

    /// Purpose: list all chats for the UI.
    /// Inputs: none.
    /// Output: chats ordered by `last_timestamp` descending; name = contact name, else chat
    /// name, else the user part of the jid.
    /// Side effects: none (read only).
    pub fn list_chats(&self) -> Result<Vec<ChatSummary>> {
        let context = || "list chats";
        let sql = format!(
            "SELECT c.jid, COALESCE(k.name, c.name), c.last_timestamp, c.autoreply \
             FROM {TABLE_CHATS} c LEFT JOIN {TABLE_CONTACTS} k ON k.jid = c.jid \
             ORDER BY c.last_timestamp DESC, c.jid"
        );

        let mut statement = self.connection.prepare_cached(&sql).with_context(context)?;
        let chats = statement
            .query_map([], chat_summary_from_row)
            .with_context(context)?
            .collect::<rusqlite::Result<Vec<_>>>()
            .with_context(context)?;

        Ok(chats)
    }

    /// Purpose: store or rename a contact.
    /// Inputs: the contact; a blank name is ignored.
    /// Output: `()`.
    /// Side effects: upserts the contacts row unless the name is blank.
    pub fn upsert_contact(&self, contact: &Contact) -> Result<()> {
        if contact.name.trim().is_empty() {
            return Ok(());
        }
        let sql = format!(
            "INSERT INTO {TABLE_CONTACTS} (jid, name) VALUES (?1, ?2) \
             ON CONFLICT(jid) DO UPDATE SET name = excluded.name"
        );

        self.connection
            .execute(&sql, params![contact.jid, contact.name])
            .with_context(|| format!("upsert contact {}", contact.jid))?;

        Ok(())
    }

    /// Purpose: list all known contacts.
    /// Inputs: none.
    /// Output: contacts ordered by name, case-insensitive.
    /// Side effects: none (read only).
    pub fn list_contacts(&self) -> Result<Vec<Contact>> {
        let context = || "list contacts";
        let sql = format!(
            "SELECT jid, name FROM {TABLE_CONTACTS} ORDER BY name COLLATE NOCASE, jid"
        );

        let mut statement = self.connection.prepare_cached(&sql).with_context(context)?;
        let contacts = statement
            .query_map([], |row| {
                Ok(Contact {
                    jid: row.get(0)?,
                    name: row.get(1)?,
                })
            })
            .with_context(context)?
            .collect::<rusqlite::Result<Vec<_>>>()
            .with_context(context)?;

        Ok(contacts)
    }

    /// Purpose: resolve the human-readable name of a jid.
    /// Inputs: `jid` (chat or contact).
    /// Output: contact name, else chat name, else the user part of the jid.
    /// Side effects: none (read only).
    pub fn display_name(&self, jid: &str) -> Result<String> {
        let sql = format!(
            "SELECT COALESCE((SELECT name FROM {TABLE_CONTACTS} WHERE jid = ?1), \
                             (SELECT name FROM {TABLE_CHATS} WHERE jid = ?1))"
        );
        let stored_name: Option<String> = self
            .connection
            .query_row(&sql, params![jid], |row| row.get(0))
            .with_context(|| format!("resolve display name of {jid}"))?;

        Ok(stored_name.unwrap_or_else(|| jid_user_part(jid).to_string()))
    }

    /// Purpose: enable or disable the AI autoreply for a chat.
    /// Inputs: `chat_jid` and the new state.
    /// Output: `()`.
    /// Side effects: upserts the chats row (created with default values if missing).
    pub fn set_autoreply(&self, chat_jid: &str, enabled: bool) -> Result<()> {
        let sql = format!(
            "INSERT INTO {TABLE_CHATS} (jid, autoreply) VALUES (?1, ?2) \
             ON CONFLICT(jid) DO UPDATE SET autoreply = excluded.autoreply"
        );

        self.connection
            .execute(&sql, params![chat_jid, enabled])
            .with_context(|| format!("set autoreply={enabled} for chat {chat_jid}"))?;

        Ok(())
    }

    /// Purpose: tell whether the AI autoreply is enabled for a chat.
    /// Inputs: `chat_jid`.
    /// Output: the stored flag, or the default (`false`) if the chat has no row.
    /// Side effects: none (read only).
    pub fn is_autoreply(&self, chat_jid: &str) -> Result<bool> {
        let sql = format!("SELECT autoreply FROM {TABLE_CHATS} WHERE jid = ?1");
        let enabled: Option<bool> = self
            .connection
            .query_row(&sql, params![chat_jid], |row| row.get(0))
            .optional()
            .with_context(|| format!("read autoreply of chat {chat_jid}"))?;

        Ok(enabled.unwrap_or(DEFAULT_AUTOREPLY))
    }
}

/// Purpose: build the `CREATE TABLE/INDEX IF NOT EXISTS` script of the whole schema.
/// Inputs: none.
/// Output: the SQL script.
/// Side effects: none.
fn schema_sql() -> String {
    format!(
        "CREATE TABLE IF NOT EXISTS {TABLE_MESSAGES} (
            message_id  TEXT PRIMARY KEY,
            chat_jid    TEXT NOT NULL,
            sender_jid  TEXT NOT NULL,
            sender_name TEXT,
            text        TEXT NOT NULL,
            timestamp   INTEGER NOT NULL,
            from_me     INTEGER NOT NULL,
            is_ai       INTEGER NOT NULL
        );
        CREATE INDEX IF NOT EXISTS {INDEX_MESSAGES_BY_CHAT}
            ON {TABLE_MESSAGES} (chat_jid, timestamp);
        CREATE TABLE IF NOT EXISTS {TABLE_CHATS} (
            jid            TEXT PRIMARY KEY,
            name           TEXT,
            last_timestamp INTEGER NOT NULL DEFAULT {DEFAULT_LAST_TIMESTAMP},
            autoreply      INTEGER NOT NULL DEFAULT {default_autoreply}
        );
        CREATE TABLE IF NOT EXISTS {TABLE_CONTACTS} (
            jid  TEXT PRIMARY KEY,
            name TEXT NOT NULL
        );",
        default_autoreply = i64::from(DEFAULT_AUTOREPLY),
    )
}

/// Purpose: insert a message unless its id is already stored.
/// Inputs: a connection (or transaction) and the message.
/// Output: `true` if a row was inserted, `false` if ignored as duplicate.
/// Side effects: writes into messages.
fn insert_message(connection: &Connection, message: &StoredMessage) -> rusqlite::Result<bool> {
    let sql = format!(
        "INSERT OR IGNORE INTO {TABLE_MESSAGES} \
         (message_id, chat_jid, sender_jid, sender_name, text, timestamp, from_me, is_ai) \
         VALUES (?1, ?2, ?3, ?4, ?5, ?6, ?7, ?8)"
    );
    let changed = connection.execute(
        &sql,
        params![
            message.message_id,
            message.chat_jid,
            message.sender_jid,
            message.sender_name,
            message.text,
            message.timestamp,
            message.from_me,
            message.is_ai,
        ],
    )?;

    Ok(changed > 0)
}

/// Purpose: upsert the chat row touched by a newly stored message.
/// Inputs: a connection (or transaction) and the message.
/// Output: `()`.
/// Side effects: writes into chats; keeps the max `last_timestamp` and never overwrites an
/// existing non-empty name.
fn upsert_chat_for_message(connection: &Connection, message: &StoredMessage) -> rusqlite::Result<()> {
    let sql = format!(
        "INSERT INTO {TABLE_CHATS} (jid, name, last_timestamp) VALUES (?1, ?2, ?3) \
         ON CONFLICT(jid) DO UPDATE SET \
             last_timestamp = MAX(last_timestamp, excluded.last_timestamp), \
             name = COALESCE(NULLIF(name, ''), excluded.name)"
    );
    connection.execute(
        &sql,
        params![message.chat_jid, chat_name_candidate(message), message.timestamp],
    )?;

    Ok(())
}

/// Purpose: pick the chat name a message may provide.
/// Inputs: the message.
/// Output: the trimmed sender push name for incoming messages, `None` otherwise or if blank.
/// Side effects: none.
fn chat_name_candidate(message: &StoredMessage) -> Option<&str> {
    if message.from_me {
        return None;
    }

    message
        .sender_name
        .as_deref()
        .map(str::trim)
        .filter(|name| !name.is_empty())
}

/// Purpose: extract the user part of a jid (text before `@`).
/// Inputs: `jid`.
/// Output: the user part, or the whole jid if it has no separator.
/// Side effects: none.
fn jid_user_part(jid: &str) -> &str {
    jid.split_once(JID_SEPARATOR).map_or(jid, |(user, _)| user)
}

/// Purpose: map a messages row (columns in `recent_messages` order) to a `StoredMessage`.
/// Inputs: the SQLite row.
/// Output: the message, or the column conversion error.
/// Side effects: none.
fn message_from_row(row: &Row<'_>) -> rusqlite::Result<StoredMessage> {
    Ok(StoredMessage {
        message_id: row.get(0)?,
        chat_jid: row.get(1)?,
        sender_jid: row.get(2)?,
        sender_name: row.get(3)?,
        text: row.get(4)?,
        timestamp: row.get(5)?,
        from_me: row.get(6)?,
        is_ai: row.get(7)?,
    })
}

/// Purpose: map a chats row (jid, resolved name, last_timestamp, autoreply) to a `ChatSummary`.
/// Inputs: the SQLite row; a NULL name falls back to the user part of the jid.
/// Output: the summary, or the column conversion error.
/// Side effects: none.
fn chat_summary_from_row(row: &Row<'_>) -> rusqlite::Result<ChatSummary> {
    let jid: String = row.get(0)?;
    let stored_name: Option<String> = row.get(1)?;
    let name = stored_name.unwrap_or_else(|| jid_user_part(&jid).to_string());

    Ok(ChatSummary {
        jid,
        name,
        last_timestamp: row.get(2)?,
        autoreply: row.get(3)?,
    })
}

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

    const CHAT_A: &str = "393331111111@s.whatsapp.net";
    const CHAT_B: &str = "393332222222@s.whatsapp.net";
    const GROUP: &str = "120363000000@g.us";

    /// Builds an incoming message with the given id, chat and timestamp.
    fn message(message_id: &str, chat_jid: &str, timestamp: i64) -> StoredMessage {
        StoredMessage {
            message_id: message_id.to_string(),
            chat_jid: chat_jid.to_string(),
            sender_jid: chat_jid.to_string(),
            sender_name: None,
            text: format!("text {message_id}"),
            timestamp,
            from_me: false,
            is_ai: false,
        }
    }

    /// Builds a contact from jid and name.
    fn contact(jid: &str, name: &str) -> Contact {
        Contact {
            jid: jid.to_string(),
            name: name.to_string(),
        }
    }

    #[test]
    fn duplicate_message_id_is_ignored() {
        let history = History::open_in_memory().unwrap();
        let original = message("m1", CHAT_A, 10);
        let mut duplicate = message("m1", CHAT_A, 20);
        duplicate.text = "changed".to_string();

        assert!(history.save_message(&original).unwrap());
        assert!(!history.save_message(&duplicate).unwrap());

        let stored = history.recent_messages(CHAT_A, 10).unwrap();
        assert_eq!(stored, vec![original]);
    }

    #[test]
    fn recent_messages_respects_limit_and_ascending_order() {
        let history = History::open_in_memory().unwrap();
        for (id, timestamp) in [("m3", 30), ("m1", 10), ("m4", 40), ("m2", 20)] {
            history.save_message(&message(id, CHAT_A, timestamp)).unwrap();
        }
        history.save_message(&message("other", CHAT_B, 50)).unwrap();

        let ids: Vec<String> = history
            .recent_messages(CHAT_A, 3)
            .unwrap()
            .into_iter()
            .map(|stored| stored.message_id)
            .collect();

        assert_eq!(ids, ["m2", "m3", "m4"]);
        assert!(history.recent_messages(CHAT_A, 0).unwrap().is_empty());
    }

    #[test]
    fn message_flags_and_sender_name_roundtrip() {
        let history = History::open_in_memory().unwrap();
        let mut sent = message("m1", CHAT_A, 10);
        sent.from_me = true;
        sent.is_ai = true;
        sent.sender_name = Some("Me".to_string());
        history.save_message(&sent).unwrap();

        assert_eq!(history.recent_messages(CHAT_A, 1).unwrap(), vec![sent]);
    }

    #[test]
    fn list_chats_orders_by_last_timestamp_desc_and_keeps_max() {
        let history = History::open_in_memory().unwrap();
        history.save_message(&message("a1", CHAT_A, 100)).unwrap();
        history.save_message(&message("b1", CHAT_B, 200)).unwrap();
        history.save_message(&message("a0", CHAT_A, 50)).unwrap();

        let chats = history.list_chats().unwrap();
        let summary: Vec<(&str, i64)> = chats
            .iter()
            .map(|chat| (chat.jid.as_str(), chat.last_timestamp))
            .collect();

        assert_eq!(summary, [(CHAT_B, 200), (CHAT_A, 100)]);
    }

    #[test]
    fn chat_name_falls_back_contact_then_chat_name_then_jid_user_part() {
        let history = History::open_in_memory().unwrap();
        history.save_message(&message("a1", CHAT_A, 30)).unwrap();
        let mut named = message("b1", CHAT_B, 20);
        named.sender_name = Some("Push Bob".to_string());
        history.save_message(&named).unwrap();
        history.save_message(&message("g1", GROUP, 10)).unwrap();

        let names = |history: &History| -> Vec<String> {
            history.list_chats().unwrap().into_iter().map(|chat| chat.name).collect()
        };
        assert_eq!(names(&history), ["393331111111", "Push Bob", "120363000000"]);

        history.upsert_contact(&contact(CHAT_B, "Roberto")).unwrap();
        assert_eq!(names(&history), ["393331111111", "Roberto", "120363000000"]);
        assert_eq!(history.display_name(CHAT_A).unwrap(), "393331111111");
        assert_eq!(history.display_name(CHAT_B).unwrap(), "Roberto");
        assert_eq!(history.display_name("unknown@s.whatsapp.net").unwrap(), "unknown");
    }

    #[test]
    fn chat_name_is_set_once_and_never_from_own_messages() {
        let history = History::open_in_memory().unwrap();
        let mut own = message("m1", CHAT_A, 10);
        own.from_me = true;
        own.sender_name = Some("Me".to_string());
        history.save_message(&own).unwrap();
        assert_eq!(history.display_name(CHAT_A).unwrap(), "393331111111");

        let mut first = message("m2", CHAT_A, 20);
        first.sender_name = Some("Alice".to_string());
        let mut second = message("m3", CHAT_A, 30);
        second.sender_name = Some("Alice Renamed".to_string());
        history.save_message(&first).unwrap();
        history.save_message(&second).unwrap();

        assert_eq!(history.display_name(CHAT_A).unwrap(), "Alice");
    }

    #[test]
    fn autoreply_defaults_to_false_and_toggles() {
        let history = History::open_in_memory().unwrap();
        assert!(!history.is_autoreply(CHAT_A).unwrap());

        history.set_autoreply(CHAT_A, true).unwrap();
        assert!(history.is_autoreply(CHAT_A).unwrap());
        assert!(!history.is_autoreply(CHAT_B).unwrap());

        history.set_autoreply(CHAT_A, false).unwrap();
        assert!(!history.is_autoreply(CHAT_A).unwrap());
    }

    #[test]
    fn autoreply_survives_incoming_messages_and_shows_in_chat_list() {
        let history = History::open_in_memory().unwrap();
        history.set_autoreply(CHAT_A, true).unwrap();
        history.save_message(&message("m1", CHAT_A, 10)).unwrap();

        let chats = history.list_chats().unwrap();
        assert_eq!(chats.len(), 1);
        assert!(chats[0].autoreply);
        assert_eq!(chats[0].last_timestamp, 10);
    }

    #[test]
    fn list_contacts_is_ordered_case_insensitive() {
        let history = History::open_in_memory().unwrap();
        history.upsert_contact(&contact("3@s.whatsapp.net", "bruno")).unwrap();
        history.upsert_contact(&contact("1@s.whatsapp.net", "Carla")).unwrap();
        history.upsert_contact(&contact("2@s.whatsapp.net", "Anna")).unwrap();

        let names: Vec<String> = history
            .list_contacts()
            .unwrap()
            .into_iter()
            .map(|stored| stored.name)
            .collect();

        assert_eq!(names, ["Anna", "bruno", "Carla"]);
    }

    #[test]
    fn upsert_contact_ignores_blank_names_and_renames_existing() {
        let history = History::open_in_memory().unwrap();
        history.upsert_contact(&contact(CHAT_A, "")).unwrap();
        history.upsert_contact(&contact(CHAT_A, "   ")).unwrap();
        assert!(history.list_contacts().unwrap().is_empty());

        history.upsert_contact(&contact(CHAT_A, "Anna")).unwrap();
        history.upsert_contact(&contact(CHAT_A, "")).unwrap();
        history.upsert_contact(&contact(CHAT_A, "Anna B")).unwrap();
        assert_eq!(history.list_contacts().unwrap(), vec![contact(CHAT_A, "Anna B")]);
    }

    #[test]
    fn open_persists_data_on_disk_and_creates_parent_directory() {
        let directory = std::env::temp_dir().join(format!("wh-history-test-{}", std::process::id()));
        let path = directory.join("nested").join("history.db");
        {
            let history = History::open(&path).unwrap();
            history.save_message(&message("m1", CHAT_A, 10)).unwrap();
        }

        let reopened = History::open(&path).unwrap();
        assert_eq!(reopened.recent_messages(CHAT_A, 10).unwrap().len(), 1);

        drop(reopened);
        std::fs::remove_dir_all(&directory).unwrap();
    }
}