handle optim errors

.gitignore

@@ -8,4 +8,6 @@ data
 dist
 docs
 
-*.log
+*.log
+*.err*
+*.swp

src/__main__.py

@@ -126,6 +126,7 @@ def main():
     logging.info(
         f"total optimization ({optimized}/{len(result.orig)}): {percent:.2f}% -> {size:.2f} Mb"  # noqa
     )
+    result.check_errors()
 
 
 if __name__ == "__main__":

src/optimizer.py

@@ -4,7 +4,7 @@ import signal
 import subprocess
 import time
 from concurrent.futures import ProcessPoolExecutor, as_completed
-from dataclasses import dataclass
+from dataclasses import dataclass, field
 from typing import Optional
 
 from tqdm import tqdm
@@ -15,7 +15,9 @@ __all__ = ["ImgOptimizer", "OptimizerResult"]
 
 
 # TODO(rmanach): add argument to set the size or leave it empty for loseless optim
-def _jpeg_optim(dest_dir: str, file: File) -> tuple["File", Optional["File"]] | None:
+def _jpeg_optim(
+    dest_dir: str, file: File
+) -> tuple["File", Optional["File"], str | None]:
     """
     Optimize the `file` with `jpegoptim` and put the result in
     `dest_dir` directory keeping file path.
@@ -35,22 +37,26 @@ def _jpeg_optim(dest_dir: str, file: File) -> tuple["File", Optional["File"]] |
     try:
         _ = subprocess.run(cmd, shell=True, check=True)
     except subprocess.CalledProcessError as e:
-        logging.error("error while running command: %s, err: %s", cmd, e.stderr or "?")
-        return None
+        logging.debug("error while running command: %s, err: %s", cmd, e.stderr or "?")
+        return file, None, f"command: {cmd} failed, err: {e.stderr}"
     except Exception:
-        logging.error("unexpected error while running command: %s", cmd, exc_info=True)
-        return None
+        logging.debug("unexpected error while running command: %s", cmd, exc_info=True)
+        return (
+            file,
+            None,
+            f"command: {cmd} failed unexpectedly, set debug mode for more details",
+        )
 
     try:
         file_optim = File.from_directory(dest_dir, file.name)
     except Exception as e:
         logging.debug("unable to get file: %s after optimization: %s", file, e)
-        return file, None
+        return file, None, None
 
-    return file, file_optim
+    return file, file_optim, None
 
 
-def _optim(dest_dir: str, file: File) -> tuple["File", Optional["File"]] | None:
+def _optim(dest_dir: str, file: File) -> tuple["File", Optional["File"], str | None]:
     """
     Entry point of `file` optimization selection the handler.
     NOTE: Must be launched in separated process.
@@ -59,7 +65,27 @@ def _optim(dest_dir: str, file: File) -> tuple["File", Optional["File"]] | None:
     signal.signal(signal.SIGINT, signal.SIG_IGN)
     if file.mimetype == FileImgMimetype.JPEG.value:
         return _jpeg_optim(dest_dir, file)
-    return None
+    return file, None, f"mimetype: {file.mimetype} not supported for optimization"
+
+
+@dataclass(slots=True, frozen=True)
+class OptimizerErrors:
+    _data: dict[str, str] = field(default_factory=dict)
+
+    def save(self):
+        if not len(self._data):
+            return
+
+        with open("optimg.err.csv", "w") as f:
+            f.write("filepath;err\n")
+            for filepath, error in self._data.items():
+                f.write(f"{filepath};{error}\n")
+
+    def has_errors(self) -> bool:
+        return len(self._data) > 0
+
+    def add(self, filepath: str, error: str):
+        self._data[filepath] = error
 
 
 @dataclass(slots=True, frozen=True)
@@ -73,6 +99,7 @@ class OptimizerResult:
     orig: FileGroup
     opti: FileGroup
     optimized: int
+    err: OptimizerErrors
 
     def stats(self) -> tuple[int, float, float]:
         """
@@ -85,6 +112,13 @@ class OptimizerResult:
         size = self.orig._size - self.opti._size
         return (self.optimized, percent, size)
 
+    def check_errors(self):
+        if self.err.has_errors():
+            logging.error(
+                "some errors detected during optimization, check 'optimg.err.csv' for details"  # noqa
+            )
+            self.err.save()
+
 
 @dataclass(slots=True, frozen=True)
 class ImgOptimizer:
@@ -94,8 +128,8 @@ class ImgOptimizer:
 
     Example:
     ```python
-        optimizer = ImgOptimizer("mypath")
-        optimizer.optimize()
+        optimizer = ImgOptimizer()
+        res = optimizer.optimize(file_group)
     ```
     """
 
@@ -134,20 +168,24 @@ class ImgOptimizer:
 
             file_group_optim = FileGroup()
             optimized = 0
+            errors = OptimizerErrors()
             for f in tqdm(
                 as_completed(futures), total=len(futures), desc="Optimizing..."
             ):
-                if (res := f.result()) and res is not None:
-                    match res:
-                        case (orig, None):
-                            logging.debug(f"no optimization for file: {orig}")
-                            file_group_optim.add(orig)
-                        case (orig, opti):
-                            optimized += 1
-                            logging.debug(
-                                f"optimization for file: {orig} -> {(1 - (opti.size / orig.size)) * 100:.2f}%"  # noqa
-                            )
-                            file_group_optim.add(opti)
+                match f.result():
+                    case (orig, None, err):
+                        logging.debug(f"optim error for file: {orig}, err: {err}")
+                        file_group_optim.add(orig)
+                        errors.add(orig.path, err)
+                    case (orig, None, None):
+                        logging.debug(f"no optimization for file: {orig}")
+                        file_group_optim.add(orig)
+                    case (orig, opti, None):
+                        optimized += 1
+                        logging.debug(
+                            f"optimization for file: {orig} -> {(1 - (opti.size / orig.size)) * 100:.2f}%"  # noqa
+                        )
+                        file_group_optim.add(opti)
 
         logging.info(f"optimization finished in {time.perf_counter() - start:.2f}s")
-        return OptimizerResult(file_group, file_group_optim, optimized)
+        return OptimizerResult(file_group, file_group_optim, optimized, errors)