momentry_core/src/api/media_api.rs

use axum::{
    body::Body,
    extract::{Path, Query, State},
    http::{header, StatusCode},
    response::{IntoResponse, Response},
    routing::get,
    Router,
};
use once_cell::sync::Lazy;

use crate::core::db::{schema, PostgresDb};

static FFMPEG: Lazy<String> = Lazy::new(|| {
    std::env::var("MOMENTRY_FFMPEG").unwrap_or_else(|_| {
        let full = "/opt/homebrew/opt/ffmpeg-full/bin/ffmpeg";
        if std::path::Path::new(full).exists() { full.to_string() } else { "ffmpeg".to_string() }
    })
});

fn ffmpeg_cmd() -> std::process::Command {
    let mut cmd = std::process::Command::new(&*FFMPEG);
    let full_lib = "/opt/homebrew/opt/ffmpeg-full/lib";
    if std::path::Path::new(full_lib).exists() { cmd.env("DYLD_LIBRARY_PATH", full_lib); }
    cmd
}

pub fn bbox_routes() -> Router<crate::api::server::AppState> {
    Router::new()
        .route(
            "/api/v1/file/:file_uuid/video/bbox",
            get(bbox_overlay_video),
        )
        .route(
            "/api/v1/file/:file_uuid/trace/:trace_id/video",
            get(trace_video),
        )
        .route("/api/v1/file/:file_uuid/video", get(stream_video))
        .route("/api/v1/file/:file_uuid/thumbnail", get(face_thumbnail))
}

/// 5×7 bitmap font — each character 5 wide × 7 tall
/// Encoding: col 0=0x10, col 1=0x08, col 2=0x04, col 3=0x02, col 4=0x01
fn bitmap_char(c: char) -> [u8; 7] {
    match c.to_ascii_lowercase() {
        '0' => [0x0E, 0x11, 0x13, 0x15, 0x19, 0x11, 0x0E],
        '1' => [0x04, 0x0C, 0x04, 0x04, 0x04, 0x04, 0x0E],
        '2' => [0x0E, 0x11, 0x01, 0x02, 0x04, 0x08, 0x1F],
        '3' => [0x0E, 0x11, 0x01, 0x06, 0x01, 0x11, 0x0E],
        '4' => [0x02, 0x06, 0x0A, 0x12, 0x1F, 0x02, 0x02],
        '5' => [0x1F, 0x10, 0x1E, 0x01, 0x01, 0x11, 0x0E],
        '6' => [0x0E, 0x10, 0x1E, 0x11, 0x11, 0x11, 0x0E],
        '7' => [0x1F, 0x01, 0x02, 0x04, 0x08, 0x10, 0x10],
        '8' => [0x0E, 0x11, 0x11, 0x0E, 0x11, 0x11, 0x0E],
        '9' => [0x0E, 0x11, 0x11, 0x0F, 0x01, 0x11, 0x0E],
        'a' => [0x0E, 0x11, 0x11, 0x1F, 0x11, 0x11, 0x11],
        'b' => [0x1E, 0x11, 0x11, 0x1E, 0x11, 0x11, 0x1E],
        'c' => [0x0E, 0x11, 0x10, 0x10, 0x10, 0x11, 0x0E],
        'd' => [0x1C, 0x12, 0x11, 0x11, 0x11, 0x12, 0x1C],
        'e' => [0x1F, 0x10, 0x10, 0x1E, 0x10, 0x10, 0x1F],
        'f' => [0x1F, 0x10, 0x10, 0x1E, 0x10, 0x10, 0x10],
        't' => [0x04, 0x04, 0x1F, 0x04, 0x04, 0x04, 0x06],
        '_' => [0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x1F],
        ' ' => [0x00; 7],
        _ => [0x00; 7],
    }
}

/// Width of one character in pixels (5 cols × 3px/dot = 15px)
const CHAR_W: i32 = 5 * 3;
/// Spacing between characters (px)
const CHAR_GAP: i32 = 4;
/// Total advance per character
const CHAR_ADVANCE: i32 = CHAR_W + CHAR_GAP;

fn render_text(
    parts: &mut Vec<String>,
    text: &str,
    origin_x: i32,
    origin_y: i32,
    enable: Option<i32>,
) -> i32 {
    let mut px = origin_x;
    for ch in text.chars() {
        let bm = bitmap_char(ch);
        for (row, bits) in bm.iter().enumerate() {
            for col in 0..5 {
                if bits & (1 << (4 - col)) != 0 {
                    let x = px + col as i32 * 3;
                    let y = origin_y + row as i32 * 3;
                    if let Some(offset) = enable {
                        parts.push(format!(
                            "drawbox=x={}:y={}:w=3:h=3:color=white@1.0:t=fill:enable='eq(n,{})'",
                            x, y, offset
                        ));
                    } else {
                        parts.push(format!(
                            "drawbox=x={}:y={}:w=3:h=3:color=white@1.0:t=fill",
                            x, y
                        ));
                    }
                }
            }
        }
        px += CHAR_ADVANCE;
    }
    px
}

#[derive(Debug, serde::Deserialize)]
struct BboxParams {
    start: Option<i32>,
    end: Option<i32>,
    face_uuid: Option<String>,
    duration: Option<f64>,
}

async fn bbox_overlay_video(
    State(state): State<crate::api::server::AppState>,
    Path(file_uuid): Path<String>,
    Query(p): Query<BboxParams>,
) -> Result<impl IntoResponse, StatusCode> {
    let videos_table = schema::table_name("videos");
    let row: Option<(String,)> = sqlx::query_as(&format!(
        "SELECT file_path FROM {} WHERE file_uuid = $1",
        videos_table
    ))
    .bind(&file_uuid)
    .fetch_optional(state.db.pool())
    .await
    .map_err(|_| StatusCode::INTERNAL_SERVER_ERROR)?;
    let (video_path,) = row.ok_or(StatusCode::NOT_FOUND)?;

    let start_f = p.start.unwrap_or(0);
    let end_f = p.end.unwrap_or(i32::MAX);
    let face_fuid = p.face_uuid.as_deref().unwrap_or(&file_uuid);
    let duration = p.duration.unwrap_or(10.0);

    // Get FPS
    let fps: f64 = sqlx::query_scalar(&format!(
        "SELECT COALESCE(fps, 24.0) FROM {} WHERE file_uuid = $1",
        videos_table
    ))
    .bind(&file_uuid)
    .fetch_optional(state.db.pool())
    .await
    .map_err(|_| StatusCode::INTERNAL_SERVER_ERROR)?
    .unwrap_or(24.0);

    let start_sec = start_f as f64 / fps;

    // Get face bboxes
    let face_table = schema::table_name("face_detections");
    let rows: Vec<(i32, i32, i32, i32, i32, Option<i32>, Option<String>)> = sqlx::query_as(
        &format!("SELECT frame_number, x, y, width, height, trace_id, face_id FROM {} WHERE file_uuid = $1 AND frame_number BETWEEN $2 AND $3 ORDER BY frame_number", face_table)
    )
    .bind(face_fuid).bind(start_f).bind(end_f)
    .fetch_all(state.db.pool()).await
    .unwrap_or_else(|e| { tracing::error!("bbox query error: {}", e); vec![] });

    // Build filters
    let mut parts: Vec<String> = Vec::new();
    let mut is_first = true;
    for (frame, x, y, w, h, trace_id, _) in &rows {
        let text = format!("t{}", trace_id.unwrap_or(0));

        if is_first {
            is_first = false;
            // Persistent bbox: thin pale red border
            parts.push(format!(
                "drawbox=x={}:y={}:w={}:h={}:color=red@0.3:thickness=4",
                x, y, w, h
            ));
            // Always-on text: top-left of bbox with padding
            let tx = *x + 6;
            let ty = *y + 6;
            render_text(&mut parts, &text, tx, ty, None);
        } else {
            let offset = frame - start_f;
            // Per-frame bbox: thick bright red
            parts.push(format!(
                "drawbox=x={}:y={}:w={}:h={}:color=red@0.8:thickness=8:enable='eq(n,{})'",
                x, y, w, h, offset
            ));
            // Per-frame text
            let tx = *x + 6;
            let ty = *y + 6;
            render_text(&mut parts, &text, tx, ty, Some(offset));
        }
    }

    let vf = if parts.is_empty() {
        "null".to_string()
    } else {
        parts.join(",")
    };

    let tmp = std::env::temp_dir().join(format!("bbox_{}.mp4", uuid::Uuid::new_v4()));
    let tmp_str = tmp.to_str().unwrap_or("").to_string();
    let status = ffmpeg_cmd()
        .args([
            "-ss",
            &start_sec.to_string(),
            "-i",
            &video_path,
            "-t",
            &duration.to_string(),
            "-vf",
            &vf,
            "-c:v",
            "libx264",
            "-preset",
            "ultrafast",
            "-crf",
            "28",
            "-an",
            "-movflags",
            "+faststart",
            "-y",
            &tmp_str,
        ])
        .status()
        .map_err(|_| StatusCode::INTERNAL_SERVER_ERROR)?;
    if !status.success() {
        let _ = std::fs::remove_file(&tmp);
        return Err(StatusCode::INTERNAL_SERVER_ERROR);
    }

    let data = tokio::fs::read(&tmp)
        .await
        .map_err(|_| StatusCode::INTERNAL_SERVER_ERROR)?;
    let _ = std::fs::remove_file(&tmp);
    Ok(Response::builder()
        .header(header::CONTENT_TYPE, "video/mp4")
        .header(header::CONTENT_LENGTH, data.len())
        .body(Body::from(data))
        .unwrap())
}

fn parse_range(range: &str, file_size: u64) -> (u64, u64) {
    let r = range.trim_start_matches("bytes=");
    let parts: Vec<&str> = r.split('-').collect();
    let start = parts[0].parse::<u64>().unwrap_or(0);
    let end = if parts.len() > 1 && !parts[1].is_empty() {
        parts[1].parse::<u64>().unwrap_or(file_size - 1)
    } else {
        file_size - 1
    };
    (start.min(file_size - 1), end.min(file_size - 1))
}

async fn trace_video(
    State(state): State<crate::api::server::AppState>,
    Path((file_uuid, trace_id)): Path<(String, i32)>,
    Query(params): Query<std::collections::HashMap<String, String>>,
) -> Result<impl IntoResponse, StatusCode> {
    use axum::http::header;

    let videos_table = schema::table_name("videos");
    let row: Option<(String,)> = sqlx::query_as(&format!(
        "SELECT file_path FROM {} WHERE file_uuid = $1",
        videos_table
    ))
    .bind(&file_uuid)
    .fetch_optional(state.db.pool())
    .await
    .map_err(|_| StatusCode::INTERNAL_SERVER_ERROR)?;
    let (video_path,) = row.ok_or(StatusCode::NOT_FOUND)?;

    let fps: f64 = sqlx::query_scalar(&format!(
        "SELECT COALESCE(fps, 24.0) FROM {} WHERE file_uuid = $1",
        videos_table
    ))
    .bind(&file_uuid)
    .fetch_optional(state.db.pool())
    .await
    .map_err(|_| StatusCode::INTERNAL_SERVER_ERROR)?
    .unwrap_or(24.0);

    // Get all detections for this trace_id
    let face_table = schema::table_name("face_detections");
    let rows: Vec<(i32, i32, i32, i32, i32)> = sqlx::query_as(&format!(
        "SELECT frame_number, x, y, width, height FROM {} WHERE file_uuid = $1 AND trace_id = $2 ORDER BY frame_number",
        face_table
    ))
    .bind(&file_uuid).bind(trace_id)
    .fetch_all(state.db.pool()).await
    .unwrap_or_else(|e| { tracing::error!("trace query error: {}", e); vec![] });

    if rows.is_empty() {
        return Err(StatusCode::NOT_FOUND);
    }

    let first_frame = rows[0].0;
    let last_frame = rows[rows.len() - 1].0;
    let start_sec = first_frame as f64 / fps;
    let padding = params.get("padding").and_then(|s| s.parse().ok()).unwrap_or(2.0);
    let duration = (last_frame - first_frame) as f64 / fps + padding * 2.0;
    let seek = (start_sec - padding).max(0.0);

    // Build filters: bbox+drawtext (1 filter + 1 drawtext per detection)
    let mut parts: Vec<String> = Vec::new();
    for (i, (frame, x, y, w, h)) in rows.iter().enumerate() {
        let next_frame = if i + 1 < rows.len() { rows[i + 1].0 } else { last_frame + (padding * fps) as i32 };
        let start_offset = frame - first_frame + (padding * fps) as i32;
        let end_offset = next_frame - first_frame + (padding * fps) as i32;
        // Bbox
        parts.push(format!(
            "drawbox=x={}:y={}:w={}:h={}:color=red@0.8:thickness=8:enable='between(n,{},{})'",
            x, y, w, h, start_offset, end_offset - 1
        ));
        // Text label (drawtext, 1 filter vs ~175 bitmap drawboxes)
        parts.push(format!(
            "drawtext=text='{}':x={}:y={}:fontsize=20:fontcolor=white:box=1:boxcolor=red@0.8:enable='between(n,{},{})'",
            trace_id, x + 4, y + 4, start_offset, end_offset - 1
        ));
    }

    // Write filter to temp file to bypass ARG_MAX
    let filter_text = parts.join(",");
    let filter_file = std::env::temp_dir().join(format!("vf_{}.txt", uuid::Uuid::new_v4()));
    let _ = std::fs::write(&filter_file, &filter_text);
    let filter_path = filter_file.to_str().unwrap_or("");

    let tmp = std::env::temp_dir().join(format!("trace_{}.mp4", uuid::Uuid::new_v4()));
    let tmp_str = tmp.to_str().unwrap_or("").to_string();
    let result = ffmpeg_cmd()
        .args([
            "-ss", &seek.to_string(), "-i", &video_path,
            "-t", &duration.to_string(),
            "-/filter_complex", &filter_path,
            "-c:v", "libx264", "-preset", "ultrafast", "-crf", "28",
            "-an", "-movflags", "+faststart", "-y", &tmp_str,
        ])
        .output()
        .map_err(|e| { tracing::error!("ffmpeg spawn: {}", e); StatusCode::INTERNAL_SERVER_ERROR })?;
    if !result.status.success() {
        let stderr = String::from_utf8_lossy(&result.stderr);
        tracing::error!("ffmpeg failed: {}", &stderr[..stderr.len().min(300)]);
        let _ = std::fs::remove_file(&filter_file);
        let _ = std::fs::remove_file(&tmp);
        return Err(StatusCode::INTERNAL_SERVER_ERROR);
    }

    let data = tokio::fs::read(&tmp)
        .await
        .map_err(|_| StatusCode::INTERNAL_SERVER_ERROR)?;
    let _ = std::fs::remove_file(&filter_file);
    let _ = std::fs::remove_file(&tmp);
    Ok(Response::builder()
        .header(header::CONTENT_TYPE, "video/mp4")
        .header(header::CONTENT_LENGTH, data.len())
        .body(Body::from(data))
        .unwrap())
}

async fn stream_video(
    State(state): State<crate::api::server::AppState>,
    Path(file_uuid): Path<String>,
    Query(params): Query<std::collections::HashMap<String, String>>,
    request: axum::http::Request<Body>,
) -> Result<impl IntoResponse, StatusCode> {
    use tokio::io::{AsyncReadExt, AsyncSeekExt};

    let videos_table = schema::table_name("videos");
    let row: Option<(String,)> = sqlx::query_as(&format!(
        "SELECT file_path FROM {} WHERE file_uuid = $1",
        videos_table
    ))
    .bind(&file_uuid)
    .fetch_optional(state.db.pool())
    .await
    .map_err(|_| StatusCode::INTERNAL_SERVER_ERROR)?;
    let (file_path,) = row.ok_or(StatusCode::NOT_FOUND)?;

    let src = std::path::PathBuf::from(&file_path);
    if !src.exists() {
        return Err(StatusCode::NOT_FOUND);
    }

    // Chunk extraction with start/end params
    if let (Some(s), Some(e)) = (params.get("start"), params.get("end")) {
        let start: f64 = s.parse().unwrap_or(0.0);
        let end: f64 = e.parse().unwrap_or(0.0);
        let dur = end - start;
        if dur <= 0.0 {
            return Err(StatusCode::BAD_REQUEST);
        }

        let tmp = std::env::temp_dir().join(format!("chunk_{}.mp4", uuid::Uuid::new_v4()));
        let tmp_str = tmp.to_str().unwrap_or("").to_string();
        let status = ffmpeg_cmd()
            .args([
                "-ss",
                &start.to_string(),
                "-i",
                &file_path,
                "-t",
                &dur.to_string(),
                "-c",
                "copy",
                "-movflags",
                "+faststart",
                "-y",
                &tmp_str,
            ])
            .status()
            .map_err(|_| StatusCode::INTERNAL_SERVER_ERROR)?;
        if !status.success() {
            let _ = std::fs::remove_file(&tmp);
            return Err(StatusCode::INTERNAL_SERVER_ERROR);
        }
        let data = tokio::fs::read(&tmp)
            .await
            .map_err(|_| StatusCode::INTERNAL_SERVER_ERROR)?;
        let _ = std::fs::remove_file(&tmp);
        return Ok(Response::builder()
            .header(header::CONTENT_TYPE, "video/mp4")
            .header(header::CONTENT_LENGTH, data.len())
            .body(Body::from(data))
            .unwrap());
    }

    let file_size = src.metadata().map(|m| m.len()).unwrap_or(0);
    let content_type = "video/mp4";

    let range_hdr = request
        .headers()
        .get(header::RANGE)
        .and_then(|v| v.to_str().ok());
    if let Some(range_str) = range_hdr {
        let (start, end) = parse_range(range_str, file_size);
        let length = end - start + 1;
        let mut file = tokio::fs::File::open(&src)
            .await
            .map_err(|_| StatusCode::INTERNAL_SERVER_ERROR)?;
        file.seek(std::io::SeekFrom::Start(start))
            .await
            .map_err(|_| StatusCode::INTERNAL_SERVER_ERROR)?;
        let take = file.take(length);
        let stream = tokio_util::io::ReaderStream::new(take);
        let body = Body::from_stream(stream);
        Ok(Response::builder()
            .status(StatusCode::PARTIAL_CONTENT)
            .header(header::CONTENT_TYPE, content_type)
            .header(
                header::CONTENT_RANGE,
                format!("bytes {}-{}/{}", start, end, file_size),
            )
            .header(header::CONTENT_LENGTH, length)
            .body(body)
            .unwrap())
    } else {
        let file = tokio::fs::File::open(&src)
            .await
            .map_err(|_| StatusCode::INTERNAL_SERVER_ERROR)?;
        let stream = tokio_util::io::ReaderStream::new(file);
        let body = Body::from_stream(stream);
        Ok(Response::builder()
            .header(header::CONTENT_TYPE, content_type)
            .header(header::CONTENT_LENGTH, file_size)
            .header(header::ACCEPT_RANGES, "bytes")
            .body(body)
            .unwrap())
    }
}

#[derive(Debug, serde::Deserialize)]
struct ThumbQuery {
    frame: i64,
    x: Option<i32>,
    y: Option<i32>,
    w: Option<i32>,
    h: Option<i32>,
}

async fn face_thumbnail(
    State(state): State<crate::api::server::AppState>,
    Path(file_uuid): Path<String>,
    Query(q): Query<ThumbQuery>,
) -> Result<impl IntoResponse, StatusCode> {
    let videos_table = schema::table_name("videos");
    let row: Option<(String,)> = sqlx::query_as(&format!(
        "SELECT file_path FROM {} WHERE file_uuid = $1",
        videos_table
    ))
    .bind(&file_uuid)
    .fetch_optional(state.db.pool())
    .await
    .map_err(|_| StatusCode::INTERNAL_SERVER_ERROR)?;
    let (file_path,) = row.ok_or(StatusCode::NOT_FOUND)?;

    let select = format!("select=eq(n\\,{})", q.frame);
    let vf = if let (Some(x), Some(y), Some(w), Some(h)) = (q.x, q.y, q.w, q.h) {
        format!("{},crop={}:{}:{}:{}", select, w, h, x, y)
    } else {
        select
    };

    let output = ffmpeg_cmd()
        .args([
            "-i",
            &file_path,
            "-vf",
            &vf,
            "-frames:v",
            "1",
            "-f",
            "image2pipe",
            "-vcodec",
            "mjpeg",
            "-",
        ])
        .output()
        .map_err(|_| StatusCode::INTERNAL_SERVER_ERROR)?;

    if !output.status.success() {
        return Err(StatusCode::INTERNAL_SERVER_ERROR);
    }

    Ok(Response::builder()
        .status(StatusCode::OK)
        .header(header::CONTENT_TYPE, "image/jpeg")
        .header(header::CACHE_CONTROL, "public, max-age=86400")
        .body(Body::from(output.stdout))
        .unwrap())
}