portal/src/content/validate.rs

//! What a content repo must satisfy before it serves: every action
//! lands on a page, every transition is one the state machine allows,
//! every recorded bucket is read back somewhere, and the built-in
//! aggregates the runtime owns are always present.

#[allow(unused_imports)] // the browser build sees fewer items
use super::*;

/// A relay must be a ws:// or wss:// URL - shared between site.yaml's
/// hero and `Requirement.relay` validation.
#[cfg(feature = "ssr")]
pub fn validate_relay_url(relay: &str) -> anyhow::Result<()> {
    let parsed = url::Url::parse(relay)
        .map_err(|e| anyhow::anyhow!("relay {relay:?} is not a valid URL: {e}"))?;
    if !matches!(parsed.scheme(), "ws" | "wss") {
        anyhow::bail!("relay {relay:?} must use the ws:// or wss:// scheme");
    }
    Ok(())
}

/// Validates every declared transition target (`SelfTransition.to`,
/// `ResourceSpec.transitions[].to`) against `aggregates` - the real
/// state graph loaded from `aggregates.yaml` - for every bucket that
/// has one declared. A bucket with no entry in `aggregates` is left
/// alone entirely (no validation applied): not every bucket needs to
/// be event-sourced to keep working. Called on every content
/// load/reload (`watch_for_reload`, `main.rs`'s boot path) and by the
/// standalone `question-lint` binary, so a YAML typo becomes a caught,
/// logged rejection instead of a silently-accepted, later-broken
/// string.
/// Bucket the runtime keeps for announced pages (`Question.event`):
/// one record per event question, `announced` while the window is
/// open, `awaiting_summary` once it closes - the "post what happened"
/// task a desk reading this bucket picks up - and `summarized` when an
/// owner marks it done. Declared here, not in content, because the
/// runtime's sweeper (`announce.rs`) is what moves the records.
pub const EVENTS_BUCKET: &str = "portal_events";

#[cfg(feature = "ssr")]
const BUILTIN_AGGREGATES_YAML: &str = "\
aggregates:
  - bucket: portal_events
    initial: announced
    states:
      announced: { event: announced }
      awaiting_summary: { event: ended }
      summarized: { event: summarized }
    transitions:
      announced: [awaiting_summary, summarized]
      awaiting_summary: [summarized]
  - bucket: portal_invites
    attended_by: the inviting group's own desk, or the link in the invite event
    initial: open
    states:
      open: { event: invited }
      done: { event: done }
    transitions:
      open: [done]
";

/// Adds the runtime's own state graphs to a content repo's. Content
/// wins if it declares the same bucket (say, to add states).
#[cfg(feature = "ssr")]
pub fn with_builtin_aggregates(
    mut aggregates: std::collections::HashMap<String, crate::aggregates::AggregateSchema>,
) -> std::collections::HashMap<String, crate::aggregates::AggregateSchema> {
    let builtin = crate::aggregates::parse_aggregates_yaml(BUILTIN_AGGREGATES_YAML)
        .expect("builtin aggregates yaml is valid");
    for (bucket, schema) in builtin {
        aggregates.entry(bucket).or_insert(schema);
    }
    aggregates
}

/// True when some kv resource in the content reads `bucket` - the
/// attended-bucket check, exposed so the lint can warn about the
/// runtime's own `portal_events` (a warning, not a failure: a repo
/// with no event pages has nothing to attend).
#[cfg(feature = "ssr")]
pub fn bucket_is_read(questions: &std::collections::HashMap<String, Question>, bucket: &str) -> bool {
    questions.values().any(|q| {
        q.alternatives.iter().any(|a| {
            a.features
                .iter()
                .any(|f| f.resource.as_ref().and_then(|r| r.bucket()).is_some_and(|b| b == bucket))
        })
    })
}

#[cfg(feature = "ssr")]
pub fn validate_questions(
    questions: &std::collections::HashMap<String, Question>,
    aggregates: &std::collections::HashMap<String, crate::aggregates::AggregateSchema>,
) -> anyhow::Result<()> {
    // A mail to the submitter needs an address to go to: every form
    // that records into that bucket has to collect a required email.
    for (bucket, schema) in aggregates {
        let submitter_states: Vec<&String> = schema.mail_for_state.iter().filter(|(_, m)| m.to_submitter()).map(|(s, _)| s).collect();
        if submitter_states.is_empty() {
            continue;
        }
        for question in questions.values() {
            for alternative in question.alternatives.iter().filter(|a| a.record_as.as_deref() == Some(bucket.as_str())) {
                let collects_email = alternative
                    .features
                    .iter()
                    .flat_map(|f| &f.requirements)
                    .any(|r| r.name == "email" && r.kind == "email" && !r.optional);
                if !collects_email {
                    anyhow::bail!(
                        "question {:?} alternative {:?}: records into {bucket:?}, whose state(s) {:?} mail the submitter, so it must collect a required `email` field of type email",
                        question.id, alternative.name, submitter_states
                    );
                }
            }
        }
    }
    for question in questions.values() {
        if let Some(event) = &question.event {
            if question.is_dynamic() {
                anyhow::bail!("question {:?}: a dynamic page can't carry an event", question.id);
            }
            if let Err(e) = event.window_ms() {
                anyhow::bail!("question {:?}: event: {e}", question.id);
            }
        }
        if let Some(place) = &question.place {
            if let Err(e) = place.validate() {
                anyhow::bail!("question {:?}: place: {e}", question.id);
            }
        }
        if let Some(target) = &question.requires_chain {
            if !questions.contains_key(target) {
                anyhow::bail!(
                    "question {:?}: requires_chain {:?} does not match any declared question id",
                    question.id,
                    target
                );
            }
        }
        for alternative in &question.alternatives {
            // A dangling action is a literal dead end: the submit
            // button navigates to "Nothing here".
            if let Some(action) = &alternative.action {
                if action.contains('{') {
                    // Placeholders are filled at submit time; the
                    // shape must land on a dynamic page once filled.
                    let probe = template_action(action, &serde_json::json!({}));
                    let filled = {
                        let mut a = action.clone();
                        while let (Some(i), Some(j)) = (a.find('{'), a.find('}')) {
                            a.replace_range(i..=j, "x");
                        }
                        a
                    };
                    let _ = probe;
                    // Match the pattern itself: resolve_question would
                    // hand back a clone whose id is already concrete.
                    let lands = questions
                        .values()
                        .any(|q| q.is_dynamic() && path_matches(&q.id, &filled).is_some());
                    if lands {
                        continue;
                    }
                    anyhow::bail!(
                        "question {:?} alternative {:?}: templated action {:?} must land on a dynamic page ([name].yaml)",
                        question.id, alternative.name, action
                    );
                }
                match questions.get(action) {
                    None => anyhow::bail!(
                        "question {:?} alternative {:?}: action {:?} does not match any declared question id",
                        question.id, alternative.name, action
                    ),
                    // A dynamic page needs a concrete segment value to
                    // be a URL - a submit button can't supply one.
                    Some(target) if target.is_dynamic() => anyhow::bail!(
                        "question {:?} alternative {:?}: action {:?} targets a dynamic page - actions must name a concrete question",
                        question.id, alternative.name, action
                    ),
                    Some(_) => {}
                }
            }
            if let Some(invite) = &alternative.invite {
                if question.qualifies.is_none() {
                    anyhow::bail!(
                        "question {:?} alternative {:?}: an invite into {:?} must sit on a page gated by `qualifies` - that group is who may extend it",
                        question.id, alternative.name, invite.group
                    );
                }
                let has = |name: &str| {
                    alternative
                        .features
                        .iter()
                        .flat_map(|f| &f.requirements)
                        .any(|r| r.name == name && !r.optional)
                };
                if !has("name") || !has("email") {
                    anyhow::bail!(
                        "question {:?} alternative {:?}: an invite must collect required `name` and `email` fields",
                        question.id, alternative.name
                    );
                }
            }
            if let Some(st) = &alternative.self_transition {
                if let Some(schema) = aggregates.get(&st.bucket) {
                    if !schema.has_state(&st.to) {
                        anyhow::bail!(
                            "question {:?} alternative {:?}: self_transition.to {:?} is not a declared state for bucket {:?}",
                            question.id, alternative.name, st.to, st.bucket
                        );
                    }
                }
            }
            // Bind targets resolve across the whole alternative, same
            // scope as the signal maps the renderer builds.
            let sibling_names: std::collections::HashSet<&str> = alternative
                .features
                .iter()
                .flat_map(|f| f.requirements.iter())
                .map(|r| r.name.as_str())
                .collect();
            for feature in &alternative.features {
                for requirement in &feature.requirements {
                    if requirement.kind == "select"
                        && requirement.resource.is_none()
                        && requirement.options.is_empty()
                    {
                        anyhow::bail!(
                            "question {:?} alternative {:?} feature {:?}: requirement {:?} is type: select but declares neither a resource nor inline options",
                            question.id, alternative.name, feature.name, requirement.name
                        );
                    }
                    if let Some(relay) = &requirement.relay {
                        if requirement.kind != "gesture" && requirement.kind != "voice" {
                            anyhow::bail!(
                                "question {:?} alternative {:?} feature {:?}: requirement {:?} declares relay but is type {:?} - relay only makes sense on type: gesture or voice",
                                question.id, alternative.name, feature.name, requirement.name, requirement.kind
                            );
                        }
                        validate_relay_url(relay).map_err(|e| anyhow::anyhow!(
                            "question {:?} alternative {:?} feature {:?}: requirement {:?}: {e}",
                            question.id, alternative.name, feature.name, requirement.name
                        ))?;
                    }
                    if requirement.kind == "gesture"
                        && (requirement.resource.is_some() || requirement.bind.is_some())
                    {
                        anyhow::bail!(
                            "question {:?} alternative {:?} feature {:?}: requirement {:?} is type: gesture - it draws its value, it can't also load one from a resource or bind",
                            question.id, alternative.name, feature.name, requirement.name
                        );
                    }
                    if requirement.kind == "location" {
                        if requirement.resource.is_some() || requirement.bind.is_some() {
                            anyhow::bail!(
                                "question {:?} alternative {:?} feature {:?}: requirement {:?} is type: location - the visitor puts the pin, it can't also load one from a resource or bind",
                                question.id, alternative.name, feature.name, requirement.name
                            );
                        }
                        // `value` opens the map somewhere useful. It is
                        // the content author's own coordinates, so a
                        // typo here is a map of the Atlantic.
                        if let Some(value) = &requirement.value {
                            let parts: Vec<&str> = value.split(',').map(str::trim).collect();
                            let ok = (2..=3).contains(&parts.len())
                                && parts.iter().all(|p| p.parse::<f64>().is_ok())
                                && parts[0].parse::<f64>().is_ok_and(|lat| (-90.0..=90.0).contains(&lat))
                                && parts[1].parse::<f64>().is_ok_and(|lon| (-180.0..=180.0).contains(&lon));
                            if !ok {
                                anyhow::bail!(
                                    "question {:?} alternative {:?} feature {:?}: requirement {:?} is type: location - its `value` opens the map and must be `lat,lon` or `lat,lon,zoom`, not {value:?}",
                                    question.id, alternative.name, feature.name, requirement.name
                                );
                            }
                        }
                    }
                    if let Some(bind) = &requirement.bind {
                        if bind.field == requirement.name
                            || !sibling_names.contains(bind.field.as_str())
                        {
                            anyhow::bail!(
                                "question {:?} alternative {:?} feature {:?}: requirement {:?} binds to {:?}, which is not another requirement on this alternative",
                                question.id, alternative.name, feature.name, requirement.name, bind.field
                            );
                        }
                    }
                }
                let Some(resource) = &feature.resource else {
                    continue;
                };
                let Some(bucket) = resource.bucket() else {
                    continue;
                };
                let Some(schema) = aggregates.get(bucket) else {
                    continue;
                };
                for transition in &resource.transitions {
                    if !schema.has_state(&transition.from) {
                        anyhow::bail!(
                            "question {:?} alternative {:?} feature {:?}: transition.from {:?} is not a declared state for bucket {:?}",
                            question.id, alternative.name, feature.name, transition.from, bucket
                        );
                    }
                    if !schema.has_state(&transition.to) {
                        anyhow::bail!(
                            "question {:?} alternative {:?} feature {:?}: transition.to {:?} is not a declared state for bucket {:?}",
                            question.id, alternative.name, feature.name, transition.to, bucket
                        );
                    }
                    if !schema.allowed(&transition.from).iter().any(|s| s == &transition.to) {
                        anyhow::bail!(
                            "question {:?} alternative {:?} feature {:?}: transition {:?} -> {:?} is not a declared edge for bucket {:?}",
                            question.id, alternative.name, feature.name, transition.from, transition.to, bucket
                        );
                    }
                }
            }
        }
    }

    // Attended buckets: no publicly collected answer may land somewhere
    // nothing reads. Every `record_as` bucket must either be read back
    // by some Kv resource in this same content repo (a desk or listing)
    // or carry an explicit `attended_by:` annotation in aggregates.yaml
    // naming the automation that consumes it. This is a contract, not a
    // convention, because convention already failed once: a question
    // was dropped and its bucket - answers included - silently fell out
    // of every reader's view.
    let mut read_buckets = std::collections::HashSet::new();
    let note_resource = |spec: &ResourceSpec, set: &mut std::collections::HashSet<String>| {
        if let ResourceSource::Kv { bucket } = &spec.source {
            set.insert(bucket.clone());
        }
    };
    for question in questions.values() {
        for alternative in &question.alternatives {
            for feature in &alternative.features {
                if let Some(spec) = &feature.resource {
                    note_resource(spec, &mut read_buckets);
                }
                for requirement in &feature.requirements {
                    if let Some(spec) = &requirement.resource {
                        note_resource(spec, &mut read_buckets);
                    }
                    if let Some(bind) = &requirement.bind {
                        note_resource(&bind.resource, &mut read_buckets);
                    }
                }
            }
        }
    }
    for question in questions.values() {
        for alternative in &question.alternatives {
            if let Some(bucket) = &alternative.record_as {
                let attended = read_buckets.contains(bucket)
                    || aggregates
                        .get(bucket)
                        .is_some_and(|schema| schema.attended_by.is_some());
                if !attended {
                    anyhow::bail!(
                        "question {:?} alternative {:?}: record_as {bucket:?} is unattended - no page reads that bucket back, and aggregates.yaml declares no attended_by for it. Add a desk/listing resource over it, or annotate the automation that consumes it.",
                        question.id, alternative.name
                    );
                }
            }
        }
    }
    Ok(())
}