clawmates/crates/cm-api/src/routes/files.rs

use std::collections::HashMap;
use std::path::Path;

use axum::extract::{Query, State};
use axum::Json;
use cm_domain::{AgentId, FileDrive, FileNode, WorkspaceId};
use serde::Deserialize;
use sqlx::PgPool;

use crate::routes::claws::workspace_agent;
use crate::{ApiError, AppState, Authed};

/// Recursively collect files under `dir` as (path-relative-to-dir → size).
fn walk_files<'a>(
    dir: std::path::PathBuf,
    base: std::path::PathBuf,
    out: &'a mut HashMap<String, u64>,
) -> std::pin::Pin<Box<dyn std::future::Future<Output = ()> + Send + 'a>> {
    Box::pin(async move {
        let mut rd = match tokio::fs::read_dir(&dir).await {
            Ok(rd) => rd,
            Err(_) => return,
        };
        while let Ok(Some(entry)) = rd.next_entry().await {
            let path = entry.path();
            let Ok(ft) = entry.file_type().await else {
                continue;
            };
            if ft.is_dir() {
                walk_files(path, base.clone(), out).await;
            } else if ft.is_file() {
                if let Ok(rel) = path.strip_prefix(&base) {
                    let rel = rel.to_string_lossy().replace('\\', "/");
                    let size = entry.metadata().await.map(|m| m.len()).unwrap_or(0);
                    out.insert(rel, size);
                }
            }
        }
    })
}

/// Reconcile `file_nodes` with what's actually in the drive's blob directory, so
/// files the Terminal created/edited/removed are reflected in the Files app
/// (Local backend only; the blob layout is `{root}/{ws}/{drive}/{scope}/{path}`).
async fn reconcile_drive(
    pool: &PgPool,
    root: &Path,
    ws: WorkspaceId,
    drive: FileDrive,
    agent: AgentId,
) {
    let scope = if drive.is_agent_scoped() {
        agent.to_string()
    } else {
        "shared".to_string()
    };
    let dir = root.join(format!("{ws}/{}/{scope}", drive.as_str()));
    let mut disk = HashMap::new();
    walk_files(
        dir,
        root.join(format!("{ws}/{}/{scope}", drive.as_str())),
        &mut disk,
    )
    .await;

    let db = match cm_db::repo::files::list(pool, ws, drive, agent).await {
        Ok(d) => d,
        Err(_) => return,
    };
    // Upsert files present on disk but new/changed in the index.
    for (path, size) in &disk {
        let existing = db.iter().find(|n| &n.path == path);
        if existing.map(|n| n.size as u64) == Some(*size) {
            continue;
        }
        let node = FileNode {
            id: existing.map(|n| n.id).unwrap_or_else(uuid::Uuid::now_v7),
            workspace_id: ws,
            agent_id: drive.is_agent_scoped().then_some(agent),
            drive,
            path: path.clone(),
            size: *size as i64,
            blob_ref: format!("{ws}/{}/{scope}/{path}", drive.as_str()),
        };
        let _ = cm_db::repo::files::upsert(pool, &node).await;
    }
    // Drop index rows whose blob no longer exists on disk.
    for n in &db {
        if n.blob_ref.is_empty() {
            continue;
        }
        if !root.join(&n.blob_ref).exists() {
            let _ = cm_db::repo::files::delete(pool, n.id).await;
        }
    }
}

#[derive(Deserialize)]
pub struct FilesQuery {
    #[serde(rename = "clawId")]
    claw_id: AgentId,
    /// `documents` (default) or `received` for the per-agent drives.
    drive: Option<String>,
}

/// GET /api/openclaw/files?clawId=&drive= — the agent's personal drives (§7.4).
pub async fn openclaw_files(
    State(state): State<AppState>,
    Authed(user): Authed,
    Query(query): Query<FilesQuery>,
) -> Result<Json<Vec<FileNode>>, ApiError> {
    let agent = workspace_agent(&state, &user, query.claw_id).await?;
    let drive: FileDrive = query
        .drive
        .as_deref()
        .unwrap_or("documents")
        .parse()
        .map_err(|_| ApiError::NotFound)?;
    if !drive.is_agent_scoped() {
        return Err(ApiError::NotFound); // shared drive has its own route
    }
    if let Some(root) = &state.file_root {
        reconcile_drive(&state.pool, root, user.workspace_id, drive, agent.id).await;
    }
    let nodes = cm_db::repo::files::list(&state.pool, user.workspace_id, drive, agent.id).await?;
    Ok(Json(nodes))
}

#[derive(Deserialize)]
pub struct SharedQuery {
    #[serde(rename = "clawId")]
    claw_id: AgentId,
}

#[derive(Deserialize)]
pub struct ContentQuery {
    #[serde(rename = "clawId")]
    claw_id: AgentId,
    drive: String,
    path: String,
}

/// GET /api/openclaw/files/content?clawId=&drive=&path= — a file's text content
/// from a drive (used by the Obsidian vault viewer to render a note).
pub async fn file_content(
    State(state): State<AppState>,
    Authed(user): Authed,
    Query(query): Query<ContentQuery>,
) -> Result<String, ApiError> {
    let agent = workspace_agent(&state, &user, query.claw_id).await?;
    let drive: FileDrive = query.drive.parse().map_err(|_| ApiError::NotFound)?;
    let scope = if drive.is_agent_scoped() {
        agent.id.to_string()
    } else {
        "shared".to_string()
    };
    let key = format!(
        "{}/{}/{}/{}",
        user.workspace_id,
        drive.as_str(),
        scope,
        query.path
    );
    let bytes = state
        .runtime
        .blob()
        .get(&key)
        .await
        .map_err(|_| ApiError::NotFound)?;
    Ok(String::from_utf8_lossy(&bytes).to_string())
}

/// GET /api/shared-drive/files?clawId= — the team-wide ClawDrive (§7.4).
pub async fn shared_files(
    State(state): State<AppState>,
    Authed(user): Authed,
    Query(query): Query<SharedQuery>,
) -> Result<Json<Vec<FileNode>>, ApiError> {
    let agent = workspace_agent(&state, &user, query.claw_id).await?;
    if let Some(root) = &state.file_root {
        reconcile_drive(
            &state.pool,
            root,
            user.workspace_id,
            FileDrive::Shared,
            agent.id,
        )
        .await;
    }
    let nodes =
        cm_db::repo::files::list(&state.pool, user.workspace_id, FileDrive::Shared, agent.id)
            .await?;
    Ok(Json(nodes))
}