boxmaker/crates/memoryd/tests/honcho_read.rs

//! The workspaces `memoryd` reads (M5c spec, sections 3, 8 and 9): a reader for each
//! `[[honcho.read]]` workspace, the fleet's and the agent's own two, each with its own key, a
//! short wait for searches and a long one for chat; the label its results carry; what `recall`
//! searches and the names `ask_memory` may use. Against a fake Honcho. Do not edit.

#[path = "support/honcho.rs"]
mod honcho;

use std::collections::BTreeMap;
use std::ffi::OsString;
use std::time::Duration;

use bxcore::DataClass;
use bxcore::secrets::SecretSource;
use honcho::{FakeHoncho, Mode};
use memoryd::honcho::client::{CHAT_READ, CONNECT, classify};
use memoryd::honcho::config::{
    FLEET_KEY, HonchoConfig, HonchoSection, RESERVED, ReadSection, ReadWorkspace, TRUSTED_KEY,
    Trust, check, read_key,
};
use memoryd::honcho::read::{
    FLEET, Label, ReadTimeouts, Reader, Readers, SEARCH_READ, Source, UNTRUSTED,
};
use roncho::MessageSearch;

const FAST: ReadTimeouts = ReadTimeouts {
    connect: Duration::from_secs(5),
    search: Duration::from_millis(300),
    chat: Duration::from_millis(900),
};

/// Key values built from pieces, so that no file holds one whole.
fn key(which: &str) -> String {
    ["hk", which, "7a31c0"].join("-")
}

fn read(agent: &str, trust: Trust, class: DataClass) -> ReadWorkspace {
    ReadWorkspace {
        agent: agent.to_string(),
        workspace: format!("{agent}-trusted"),
        trust,
        class,
    }
}

/// `boxmaker0`, reading `boxmaker1` (labelled trusted, `secret`) and `boxmaker2` (the defaults).
/// Keys for `boxmaker1`, the fleet and the own trusted workspace; none for the others.
fn config(url: &str) -> HonchoConfig {
    let mut keys = BTreeMap::new();
    for (name, var) in [
        (read_key("boxmaker1"), "K_READ_1"),
        (FLEET_KEY.to_string(), "K_FLEET"),
        (TRUSTED_KEY.to_string(), "K_TRUSTED"),
    ] {
        keys.insert(name, SecretSource::Env(var.to_string()));
    }
    HonchoConfig {
        url: url.to_string(),
        agent: "boxmaker0".to_string(),
        owner: "kyle".to_string(),
        trusted: "boxmaker0-trusted".to_string(),
        untrusted: "boxmaker0-untrusted".to_string(),
        fleet: "fleet".to_string(),
        read: vec![
            read("boxmaker1", Trust::Trusted, DataClass::Secret),
            read("boxmaker2", Trust::Untrusted, DataClass::Public),
        ],
        keys,
    }
}

fn env(name: &str) -> Option<OsString> {
    match name {
        "K_READ_1" => Some(OsString::from(key("read1"))),
        "K_FLEET" => Some(OsString::from(key("fleet"))),
        "K_TRUSTED" => Some(OsString::from(key("trusted"))),
        _ => None,
    }
}

fn readers(url: &str) -> Readers {
    Readers::new(&config(url), &env, FAST).unwrap().0
}

/// Every reader: `recall`'s, then the own two.
fn all(r: &Readers) -> Vec<&Reader> {
    let mut all = r.recall_sources();
    all.push(r.resolve("own").unwrap());
    all.push(r.resolve("untrusted").unwrap());
    all
}

#[test]
fn a_reader_client_can_be_shared_across_threads() {
    fn shared<T: Send + Sync>() {}
    shared::<roncho::blocking::Client>();
    shared::<Reader>();
    shared::<Readers>();
}

#[test]
fn the_timeouts() {
    assert_eq!(SEARCH_READ, Duration::from_secs(5));
    assert_eq!(CHAT_READ, Duration::from_secs(150));
    assert_eq!(
        ReadTimeouts::DEFAULT,
        ReadTimeouts {
            connect: CONNECT,
            search: SEARCH_READ,
            chat: CHAT_READ
        }
    );
    let fake = FakeHoncho::start();
    let r = Readers::new(&config(&fake.url), &env, ReadTimeouts::DEFAULT)
        .unwrap()
        .0;
    for reader in all(&r) {
        // `roncho`'s `Debug` shows the timeouts, and whether there is a key, never the key.
        let search = format!("{:?}", reader.search());
        let chat = format!("{:?}", reader.chat());
        for shown in [&search, &chat] {
            assert!(shown.contains("connect_timeout: 5s"), "{shown}");
        }
        assert!(search.contains("read_timeout: 5s"), "{search}");
        assert!(chat.contains("read_timeout: 150s"), "{chat}");
    }
}

#[test]
fn each_reader_reads_its_own_workspace_with_its_own_key() {
    let fake = FakeHoncho::start();
    let r = readers(&fake.url);
    let want = [
        ("boxmaker1-trusted", Some(key("read1"))),
        ("boxmaker2-trusted", None),
        ("fleet", Some(key("fleet"))),
        ("boxmaker0-trusted", Some(key("trusted"))),
        ("boxmaker0-untrusted", None),
    ];
    let readers = all(&r);
    assert_eq!(readers.len(), want.len());
    for (reader, (workspace, _)) in readers.iter().zip(&want) {
        assert_eq!(reader.search().workspace_id(), *workspace);
        assert_eq!(reader.chat().workspace_id(), *workspace);
        // The fake answers a search it has no recording for with 404: only the request matters.
        let _ = reader
            .search()
            .search_workspace(&MessageSearch::new("launch"));
        let _ = reader
            .chat()
            .search_workspace(&MessageSearch::new("launch"));
    }
    let requests = fake.requests();
    assert_eq!(requests.len(), 2 * want.len());
    for (n, request) in requests.iter().enumerate() {
        let (workspace, secret) = &want[n / 2];
        assert_eq!(request.path, format!("/v3/workspaces/{workspace}/search"));
        let bearer = secret.as_ref().map(|k| format!("Bearer {k}"));
        assert_eq!(
            request.header("authorization").map(str::to_string),
            bearer,
            "{workspace}"
        );
    }
}

#[test]
fn no_key_is_shown_by_a_reader_or_an_error() {
    let fake = FakeHoncho::start();
    let r = readers(&fake.url);
    fake.set(Mode::Status(401));
    let mut shown = vec![format!("{r:?}")];
    for reader in all(&r) {
        let err = reader
            .search()
            .search_workspace(&MessageSearch::new("launch"))
            .unwrap_err();
        let err = classify(&err);
        shown.push(format!("{err:?}"));
        shown.push(err.to_string());
    }
    for text in shown {
        for which in ["read1", "fleet", "trusted"] {
            assert!(!text.contains(&key(which)), "{text}");
        }
    }
    // A key that cannot be read stops the build, naming the key and never a value.
    let mut config = config(&fake.url);
    config.keys.insert(
        read_key("boxmaker2"),
        SecretSource::Env("K_MISSING".to_string()),
    );
    let err = Readers::new(&config, &env, FAST).unwrap_err();
    assert!(err.contains("honcho_read_boxmaker2"), "{err}");
    assert!(
        err.contains("see docs/runbook.md#secret-unavailable"),
        "{err}"
    );
    for which in ["read1", "fleet", "trusted"] {
        assert!(!err.contains(&key(which)), "{err}");
    }
}

/// `recall` searches the read workspaces, in the order written, then the fleet's; never the
/// agent's own two (M5c spec, section 8).
#[test]
fn recall_never_searches_the_agents_own_workspaces() {
    let fake = FakeHoncho::start();
    let r = readers(&fake.url);
    let sources: Vec<Source> = r
        .recall_sources()
        .iter()
        .map(|s| s.source.clone())
        .collect();
    assert_eq!(
        sources,
        [
            Source::Agent("boxmaker1".to_string()),
            Source::Agent("boxmaker2".to_string()),
            Source::Fleet
        ]
    );
    let workspaces: Vec<&str> = r
        .recall_sources()
        .iter()
        .map(|s| s.search().workspace_id())
        .collect();
    assert_eq!(
        workspaces,
        ["boxmaker1-trusted", "boxmaker2-trusted", "fleet"]
    );
    let mut none = config(&fake.url);
    none.read.clear();
    let r = Readers::new(&none, &env, FAST).unwrap().0;
    let sources: Vec<Source> = r
        .recall_sources()
        .iter()
        .map(|s| s.source.clone())
        .collect();
    assert_eq!(sources, [Source::Fleet]);
}

fn label(name: &str, trust: Trust, class: DataClass) -> Label {
    Label {
        name: name.to_string(),
        trust,
        class,
    }
}

/// A result is tagged `memory` only from a workspace the owner labelled trusted, or the agent's
/// own trusted one; the fleet and everything else is `memory-tainted`.
#[test]
fn each_source_has_its_label() {
    let fake = FakeHoncho::start();
    let r = readers(&fake.url);
    let got: Vec<(Label, &str, bool)> = all(&r)
        .iter()
        .map(|reader| {
            (
                reader.label.clone(),
                reader.label.tag(),
                reader.label.untrusted(),
            )
        })
        .collect();
    assert_eq!(
        got,
        [
            (
                label("boxmaker1", Trust::Trusted, DataClass::Secret),
                "memory",
                false
            ),
            (
                label("boxmaker2", Trust::Untrusted, DataClass::Public),
                "memory-tainted",
                true
            ),
            (
                label("fleet", Trust::Untrusted, DataClass::Private),
                "memory-tainted",
                true
            ),
            (
                label("own", Trust::Trusted, DataClass::Private),
                "memory",
                false
            ),
            (
                label("untrusted", Trust::Untrusted, DataClass::Private),
                "memory-tainted",
                true
            ),
        ]
    );
}

#[test]
fn ask_memory_names_resolve_to_their_readers() {
    let fake = FakeHoncho::start();
    let r = readers(&fake.url);
    for (name, source, workspace) in [
        ("own", Source::Own, "boxmaker0-trusted"),
        (UNTRUSTED, Source::OwnUntrusted, "boxmaker0-untrusted"),
        (FLEET, Source::Fleet, "fleet"),
        (
            "boxmaker1",
            Source::Agent("boxmaker1".to_string()),
            "boxmaker1-trusted",
        ),
        (
            "boxmaker2",
            Source::Agent("boxmaker2".to_string()),
            "boxmaker2-trusted",
        ),
    ] {
        let reader = r.resolve(name).unwrap();
        assert_eq!(reader.source, source, "{name}");
        assert_eq!(reader.chat().workspace_id(), workspace, "{name}");
    }
    for unknown in ["boxmaker9", "boxmaker0", "Own", "boxmaker1-trusted"] {
        assert_eq!(
            r.resolve(unknown).unwrap_err(),
            format!(
                "no workspace is named {unknown}; ask own, untrusted, fleet, boxmaker1 or boxmaker2"
            )
        );
    }
    let mut none = config(&fake.url);
    none.read.clear();
    let r = Readers::new(&none, &env, FAST).unwrap().0;
    assert_eq!(
        r.resolve("boxmaker1").unwrap_err(),
        "no workspace is named boxmaker1; ask own, untrusted or fleet"
    );
}

/// `own`, `untrusted` and `fleet` name this agent's workspaces in `ask_memory`, so no
/// `[[honcho.read]]` agent may be called that: the fleet can never be labelled trusted.
#[test]
fn no_read_agent_takes_a_reserved_name() {
    assert_eq!(RESERVED, ["own", "untrusted", "fleet"]);
    for agent in RESERVED {
        let section = HonchoSection {
            url: "http://127.0.0.1:9".to_string(),
            agent: "boxmaker0".to_string(),
            owner: "kyle".to_string(),
            trusted: "boxmaker0-trusted".to_string(),
            untrusted: "boxmaker0-untrusted".to_string(),
            fleet: "fleet".to_string(),
            read: vec![ReadSection {
                agent: agent.to_string(),
                workspace: "someone-trusted".to_string(),
                trust: Some("trusted".to_string()),
                class: None,
            }],
        };
        assert_eq!(
            check(Some(&section), &BTreeMap::new()).unwrap_err(),
            format!(
                "[[honcho.read]] agent {agent:?} is a name ask_memory gives this agent's own \
                 workspaces: own, untrusted and fleet cannot be agents' names"
            )
        );
    }
}