vfxreview/lib/batch-upload-parser.ts

/**
 * Utilities for parsing VFX filenames into shot / task metadata.
 *
 * Expected naming convention (any separator count is fine):
 *   {SHOW}_{EP}_{SCENE}_{CUT}_{TASK}_{ARTIST}_v{NNN}.{ext}
 *   e.g.  UNG_106_035_010_cmp_TT_v002.mp4
 *
 * The shot code is matched against the database – the parser itself
 * does NOT hard-code how many underscore segments belong to the shot.
 */

export interface ParsedFilename {
  originalName: string;
  /** Filename without the extension */
  nameWithoutExt: string;
  /** Lowercased extension without the leading dot (e.g. "mp4", "mov") */
  ext: string;
  /**
   * Task name base – the full name without the version suffix.
   * e.g. "UNG_106_035_010_cmp_TT"
   */
  taskNameBase: string;
  /**
   * Full task title including the version suffix (no extension).
   * This is what the task should be named after upload.
   * e.g. "UNG_106_035_010_cmp_TT_v002"
   */
  taskNameWithVersion: string;
  /**
   * Version string extracted from the filename, e.g. "v002".
   * Empty string if no version pattern was found.
   */
  version: string;
  type: "mp4" | "mov" | "other";
}

/**
 * Parse a VFX filename into its constituent parts.
 */
export function parseFilename(filename: string): ParsedFilename {
  const lastDot = filename.lastIndexOf(".");
  const ext = lastDot >= 0 ? filename.slice(lastDot + 1).toLowerCase() : "";
  const nameWithoutExt = lastDot >= 0 ? filename.slice(0, lastDot) : filename;

  // Version pattern: _v followed by one or more digits at the end (case insensitive)
  const versionMatch = nameWithoutExt.match(/_v(\d+)$/i);
  const version = versionMatch ? `v${versionMatch[1]}` : "";
  const taskNameBase = versionMatch
    ? nameWithoutExt.slice(0, nameWithoutExt.length - versionMatch[0].length)
    : nameWithoutExt;

  const type: "mp4" | "mov" | "other" =
    ext === "mp4" ? "mp4" : ext === "mov" ? "mov" : "other";

  return {
    originalName: filename,
    nameWithoutExt,
    ext,
    taskNameBase,
    taskNameWithVersion: nameWithoutExt,
    version,
    type,
  };
}

/**
 * Strip a trailing version suffix (_v###) from a task title.
 * e.g. "UNG_106_035_010_cmp_TT_v001" → "UNG_106_035_010_cmp_TT"
 */
export function stripVersionFromTitle(title: string): string {
  return title.replace(/_v\d+$/i, "");
}

/**
 * Given a taskNameBase and a list of shot codes, return the shot code that
 * is a prefix of the task name base (shot code + "_" separator).
 *
 * e.g. taskNameBase "UNG_106_035_010_cmp_TT" matches shot "UNG_106_035_010"
 */
export function findMatchingShotCode(
  taskNameBase: string,
  shotCodes: string[]
): string | null {
  // Sort by length descending so longer (more specific) codes are matched first
  const sorted = [...shotCodes].sort((a, b) => b.length - a.length);
  for (const code of sorted) {
    if (
      taskNameBase === code ||
      taskNameBase.toLowerCase().startsWith(code.toLowerCase() + "_")
    ) {
      return code;
    }
  }
  return null;
}

// ── Footage plates ────────────────────────────────────────────────────────────

/** Video extensions accepted by the footage plate batch upload. */
export const PLATE_EXTS = new Set(["mov", "mp4", "m4v", "mxf", "avi", "mkv", "webm"]);

const PLATE_CONTENT_TYPES: Record<string, string> = {
  mov: "video/quicktime",
  mp4: "video/mp4",
  m4v: "video/mp4",
  mxf: "video/mxf",
  avi: "video/x-msvideo",
  mkv: "video/x-matroska",
  webm: "video/webm",
};

/**
 * Derive a `video/*` content type from a plate filename's extension.
 *
 * Browsers routinely report an empty MIME type for .mov / .mxf files
 * (especially on Windows), and the presign endpoint rejects anything that
 * isn't `video/*`, so never trust `File.type` for these.
 */
export function plateContentType(fileName: string): string {
  const ext = fileName.split(".").pop()?.toLowerCase() ?? "";
  return PLATE_CONTENT_TYPES[ext] ?? "application/octet-stream";
}

export interface ParsedPlateFilename {
  /** Lowercased extension without the leading dot */
  ext: string;
  /** Filename without the extension */
  nameWithoutExt: string;
  /** Shot code this file belongs to, or null when nothing matched */
  shotCode: string | null;
  /**
   * The filename remainder after the shot code, used as the plate label.
   * e.g. "DH2_VFX_058_120_REF.mov" → "REF".
   * Empty when the filename is exactly the shot code.
   */
  label: string;
}

/**
 * Resolve a footage plate filename to a shot code and a plate label.
 *
 * e.g. "DH2_VFX_058_120_REF.mov" against shot "DH2_VFX_058_120"
 *      → { shotCode: "DH2_VFX_058_120", label: "REF" }
 */
export function parsePlateFilename(
  fileName: string,
  shotCodes: string[]
): ParsedPlateFilename {
  const lastDot = fileName.lastIndexOf(".");
  const ext = lastDot >= 0 ? fileName.slice(lastDot + 1).toLowerCase() : "";
  const nameWithoutExt = lastDot >= 0 ? fileName.slice(0, lastDot) : fileName;

  const shotCode = findMatchingShotCode(nameWithoutExt, shotCodes);
  const label = shotCode
    ? nameWithoutExt.slice(shotCode.length).replace(/^[_\-\s]+/, "").trim()
    : "";

  return { ext, nameWithoutExt, shotCode, label };
}