mdify-cli 2.9.5__tar.gz → 2.11.4__tar.gz

{mdify_cli-2.9.5/mdify_cli.egg-info → mdify_cli-2.11.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mdify-cli
-Version: 2.9.5
+Version: 2.11.4
 Summary: Convert PDFs and document images into structured Markdown for LLM workflows
 Author: tiroq
 License-Expression: MIT

{mdify_cli-2.9.5 → mdify_cli-2.11.4}/mdify/__init__.py

@@ -1,3 +1,3 @@
 """mdify - Convert documents to Markdown via Docling container."""
 
-__version__ = "2.9.5"
+__version__ = "2.11.4"

{mdify_cli-2.9.5 → mdify_cli-2.11.4}/mdify/cli.py

@@ -38,6 +38,9 @@ SUPPORTED_RUNTIMES = ("docker", "podman", "orbstack", "colima", "container")
 MACOS_RUNTIMES_PRIORITY = ("container", "orbstack", "colima", "podman", "docker")
 OTHER_RUNTIMES_PRIORITY = ("docker", "podman")
 
+# Debug mode
+DEBUG = os.environ.get("MDIFY_DEBUG", "").lower() in ("1", "true", "yes")
+
 
 # =============================================================================
 # Update checking functions
@@ -294,8 +297,10 @@ def check_image_exists(runtime: str, image: str) -> bool:
                 try:
                     images = json.loads(result.stdout.decode())
                     # Check if image exists in the list
+                    # Apple Container returns format: [{"reference": "image:tag", "descriptor": {...}}]
                     for img in images:
-                        if img.get("name") == image or image in img.get("repoTags", []):
+                        reference = img.get("reference", "")
+                        if reference == image or reference.startswith(f"{image}:"):
                             return True
                 except json.JSONDecodeError:
                     pass
@@ -521,8 +526,8 @@ class Spinner:
         self.running = False
         if self.thread:
             self.thread.join(timeout=0.5)
-        # Clear the spinner line
-        print(f"\r{' ' * 80}\r", end="", flush=True)
+        # Clear the spinner line with enough spaces to cover the longest possible line
+        print(f"\r{' ' * 120}\r", end="", flush=True)
 
 
 # =============================================================================
@@ -813,6 +818,10 @@ def main() -> int:
 
     image_exists = check_image_exists(runtime, image)
 
+    if not args.quiet and image_exists:
+        print(f"Using cached image: {image}")
+        print()
+
     # NOTE: Docker Desktop on macOS/Windows uses a VM, so disk space checks may not
     # accurately reflect available space in the container's filesystem. Remote Docker
     # daemons (DOCKER_HOST) are also not supported. In these cases, the check will
@@ -977,11 +986,15 @@ def main() -> int:
                 start_time = time.time()
                 try:
                     # Convert via HTTP API
+                    if DEBUG:
+                        print(f"    DEBUG: Converting {input_file.name} via {container.base_url}/v1/convert/file", file=sys.stderr)
+                    
                     result = convert_file(
                         container.base_url, input_file, to_format="md"
                     )
                     elapsed = time.time() - start_time
 
+                    # Stop spinner before any output
                     if not args.quiet:
                         spinner.stop()
 
@@ -1004,12 +1017,58 @@ def main() -> int:
                 except Exception as e:
                     elapsed = time.time() - start_time
                     failed_count += 1
+                    # Stop spinner before printing error
                     if not args.quiet:
                         spinner.stop()
-                        print(
-                            f"{progress} {input_file.name} ✗ ({format_duration(elapsed)})"
-                        )
-                        print(f"    Error: {str(e)}", file=sys.stderr)
+                    
+                    # Check if container is still healthy
+                    error_msg = str(e)
+                    is_connection_error = "Connection refused" in error_msg or "Connection aborted" in error_msg or "RemoteDisconnected" in error_msg
+                    
+                    if is_connection_error:
+                        container_alive = container.is_ready()
+                        if not args.quiet:
+                            print(
+                                f"{progress} {input_file.name} ✗ ({format_duration(elapsed)})"
+                            )
+                            if container_alive:
+                                print(
+                                    "    Error: Connection lost (server may have crashed and restarted)",
+                                    file=sys.stderr,
+                                )
+                            else:
+                                print(
+                                    "    Error: Container crashed while processing file",
+                                    file=sys.stderr,
+                                )
+                                print(
+                                    "    File may be too complex, large, or malformed",
+                                    file=sys.stderr,
+                                )
+
+                            # Always show logs for connection errors to surface root cause
+                            print("    Retrieving container logs...", file=sys.stderr)
+                            logs = container.get_logs(tail=30)
+                            if logs:
+                                print("    Container logs (last 30 lines):", file=sys.stderr)
+                                for line in logs.strip().split("\n"):
+                                    print(f"      {line}", file=sys.stderr)
+                            else:
+                                print("    No logs available", file=sys.stderr)
+
+                            if not container_alive:
+                                print("    Stopping remaining conversions", file=sys.stderr)
+
+                        # Stop processing if container is dead
+                        if not container_alive:
+                            break
+                    else:
+                        # Non-connection error
+                        if not args.quiet:
+                            print(
+                                f"{progress} {input_file.name} ✗ ({format_duration(elapsed)})"
+                            )
+                            print(f"    Error: {error_msg}", file=sys.stderr)
 
             total_elapsed = time.time() - conversion_start
 

{mdify_cli-2.9.5 → mdify_cli-2.11.4}/mdify/container.py

@@ -125,6 +125,48 @@ class DoclingContainer:
                 check=False,
             )
 
+    def get_logs(self, tail: int = 50) -> str:
+        """Get container logs for debugging.
+
+        Args:
+            tail: Number of lines to retrieve from end of logs
+
+        Returns:
+            Container logs as string
+        """
+        if not self.container_name:
+            return ""
+        
+        try:
+            result = subprocess.run(
+                [self.runtime, "logs", "--tail", str(tail), self.container_name],
+                capture_output=True,
+                text=True,
+                check=False,
+            )
+            return result.stdout if result.returncode == 0 else ""
+        except Exception:
+            return ""
+
+    def is_running(self) -> bool:
+        """Check if container process is still running.
+
+        Returns:
+            True if container is running, False otherwise
+        """
+        if not self.container_name:
+            return False
+        
+        try:
+            result = subprocess.run(
+                [self.runtime, "ps", "-q", "-f", f"name={self.container_name}"],
+                capture_output=True,
+                check=False,
+            )
+            return result.returncode == 0 and bool(result.stdout.strip())
+        except Exception:
+            return False
+
     def is_ready(self) -> bool:
         """Check if container is healthy.
 
@@ -132,6 +174,10 @@ class DoclingContainer:
             True if container is healthy, False otherwise
         """
         try:
+            # First check if container is still running
+            if not self.is_running():
+                return False
+            # Then check health endpoint
             return check_health(self.base_url)
         except Exception:
             return False

{mdify_cli-2.9.5 → mdify_cli-2.11.4/mdify_cli.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mdify-cli
-Version: 2.9.5
+Version: 2.11.4
 Summary: Convert PDFs and document images into structured Markdown for LLM workflows
 Author: tiroq
 License-Expression: MIT

{mdify_cli-2.9.5 → mdify_cli-2.11.4}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "mdify-cli"
-version = "2.9.5"
+version = "2.11.4"
 description = "Convert PDFs and document images into structured Markdown for LLM workflows"
 readme = "README.md"
 requires-python = ">=3.8"

{mdify_cli-2.9.5 → mdify_cli-2.11.4}/tests/test_cli.py

@@ -995,12 +995,19 @@ class TestContainerRuntime:
         """Test check_image_exists uses 'image list' for Apple Container."""
         mock_result = Mock()
         mock_result.returncode = 0
+        # Use actual Apple Container response format with 'reference' field
         mock_result.stdout = json.dumps([
-            {"name": "test-image", "repoTags": ["test-image:latest"]},
-            {"name": "other-image", "repoTags": ["other-image:latest"]}
+            {
+                "reference": "ghcr.io/docling-project/docling-serve-cpu:main",
+                "descriptor": {
+                    "size": 1609,
+                    "mediaType": "application/vnd.oci.image.index.v1+json",
+                    "digest": "sha256:25e82dfa30371d17a0af17edc42261a4b9bedb37f0f337887c366184bc3ee291"
+                }
+            }
         ]).encode()
         with patch("mdify.cli.subprocess.run", return_value=mock_result) as mock_run:
-            result = check_image_exists("/usr/local/bin/container", "test-image")
+            result = check_image_exists("/usr/local/bin/container", "ghcr.io/docling-project/docling-serve-cpu:main")
         assert result is True
         mock_run.assert_called_once_with(
             ["/usr/local/bin/container", "image", "list", "--format", "json"],
@@ -1013,10 +1020,17 @@ class TestContainerRuntime:
         mock_result = Mock()
         mock_result.returncode = 0
         mock_result.stdout = json.dumps([
-            {"name": "other-image", "repoTags": ["other-image:latest"]}
+            {
+                "reference": "ghcr.io/other-project/other-image:latest",
+                "descriptor": {
+                    "size": 1234,
+                    "mediaType": "application/vnd.oci.image.index.v1+json",
+                    "digest": "sha256:abcd1234"
+                }
+            }
         ]).encode()
         with patch("mdify.cli.subprocess.run", return_value=mock_result):
-            result = check_image_exists("/usr/local/bin/container", "test-image")
+            result = check_image_exists("/usr/local/bin/container", "ghcr.io/docling-project/docling-serve-cpu:main")
         assert result is False