Refactor photo metadata workflow

tools/exiftool.py

@@ -2,8 +2,11 @@ from __future__ import annotations
 
 import json
 import os
+import re
 import subprocess
 import tempfile
+from collections.abc import Callable, Iterator
+from contextlib import contextmanager
 from pathlib import Path
 from typing import Any
 
@@ -18,15 +21,12 @@ def _use_windows_argument_file() -> bool:
     return os.name == "nt"
 
 
-def run_exiftool_command(
-    executable: str,
-    args: list[str],
-    **run_kwargs: Any,
-) -> subprocess.CompletedProcess:
-    """Run ExifTool with Unicode-safe filename arguments on Windows."""
+@contextmanager
+def _prepared_command(executable: str, args: list[str]) -> Iterator[tuple[list[str], str | None]]:
     command = [executable, "-charset", "filename=UTF8", *args]
     if not _use_windows_argument_file():
-        return subprocess.run(command, **run_kwargs)
+        yield command, None
+        return
 
     # ExifTool can't reliably receive arbitrary Unicode paths on the Windows
     # command line. The outer path is ASCII and the UTF-8 argument file holds
@@ -34,11 +34,20 @@ def run_exiftool_command(
     with tempfile.TemporaryDirectory(prefix="mbt-exiftool-") as temp_dir:
         argument_file = Path(temp_dir) / "arguments.txt"
         argument_file.write_text("\n".join(args) + "\n", encoding="utf-8")
-        return subprocess.run(
-            [executable, "-charset", "filename=UTF8", "-@", argument_file.name],
-            cwd=temp_dir,
-            **run_kwargs,
-        )
+        yield [executable, "-charset", "filename=UTF8", "-@", argument_file.name], temp_dir
+
+
+def run_exiftool_command(
+    executable: str,
+    args: list[str],
+    **run_kwargs: Any,
+) -> subprocess.CompletedProcess:
+    """Run ExifTool with Unicode-safe filename arguments on Windows."""
+    with _prepared_command(executable, args) as (command, cwd):
+        return subprocess.run(command, cwd=cwd, **run_kwargs)
+
+
+_PROGRESS_RE = re.compile(r"^======== .* \[\d+/(\d+)\]$")
 
 
 def run_exiftool_json(
@@ -46,6 +55,7 @@ def run_exiftool_json(
     paths: list[Path],
     *,
     quicktime_utc: bool = True,
+    progress: Callable[[], None] | None = None,
 ) -> list[dict]:
     results: list[dict] = []
     for start in range(0, len(paths), 80):
@@ -57,14 +67,45 @@ def run_exiftool_json(
         if quicktime_utc:
             args.extend(["-api", "QuickTimeUTC=1"])
         args.extend(str(path) for path in chunk)
-        completed = run_exiftool_command(
-            executable,
-            args,
-            check=True,
-            capture_output=True,
-            text=True,
-        )
-        results.extend(json.loads(completed.stdout or "[]"))
+        if progress is None:
+            completed = run_exiftool_command(
+                executable,
+                args,
+                check=True,
+                capture_output=True,
+                text=True,
+            )
+            results.extend(json.loads(completed.stdout or "[]"))
+            continue
+
+        with _prepared_command(executable, ["-progress", *args]) as (command, cwd):
+            with tempfile.TemporaryFile(mode="w+t", encoding="utf-8") as output:
+                process = subprocess.Popen(
+                    command,
+                    cwd=cwd,
+                    stdout=output,
+                    stderr=subprocess.PIPE,
+                    text=True,
+                )
+                assert process.stderr is not None
+                errors = []
+                started_file = False
+                for line in process.stderr:
+                    if _PROGRESS_RE.match(line.rstrip("\r\n")):
+                        # ExifTool emits this marker before reading the named
+                        # file, so it confirms that the previous one finished.
+                        if started_file:
+                            progress()
+                        started_file = True
+                    else:
+                        errors.append(line)
+                return_code = process.wait()
+                if return_code:
+                    raise subprocess.CalledProcessError(return_code, command, stderr="".join(errors))
+                if started_file:
+                    progress()
+                output.seek(0)
+                results.extend(json.load(output))
     return results
 
 
@@ -76,4 +117,13 @@ def run_exiftool_write(executable: str, path: Path, args: list[str]) -> None:
         *args,
         str(path),
     ]
-    run_exiftool_command(executable, command, check=True)
+    completed = run_exiftool_command(
+        executable,
+        command,
+        check=False,
+        capture_output=True,
+        text=True,
+    )
+    if completed.returncode:
+        detail = (completed.stderr or completed.stdout or "").strip()
+        raise RuntimeError(f"ExifTool failed: {detail or f'exit code {completed.returncode}'}")