mdify-cli 3.0.8__py3-none-any.whl → 3.1.1__py3-none-any.whl

mdify/__init__.py

@@ -1,3 +1,3 @@
 """mdify - Convert documents to Markdown via Docling container."""
 
-__version__ = "3.0.8"
+__version__ = "3.1.1"

mdify/cli.py

@@ -8,12 +8,14 @@ is lightweight and has no ML dependencies.
 """
 
 import argparse
+import asyncio
 import json
 import os
 import platform
 import shutil
 import subprocess
 import sys
+import tempfile
 import threading
 import time
 from pathlib import Path
@@ -1173,7 +1175,9 @@ def main_async_remote(args) -> int:
                 print(f"Error: Input file or directory not found: {args.input}", file=sys.stderr)
                 return 1
             
-            files_to_convert = get_files_to_convert(input_path.resolve(), args.glob, args.recursive)
+            # Store resolved path as base for relative path calculations
+            input_base = input_path.resolve()
+            files_to_convert = get_files_to_convert(input_base, args.glob, args.recursive)
             
             if not files_to_convert:
                 await ssh_client.disconnect()
@@ -1247,7 +1251,8 @@ def main_async_remote(args) -> int:
                 if isinstance(exc, SSHConnectionError):
                     return True
                 msg = str(exc).lower()
-                return "broken pipe" in msg or "connection closed" in msg
+                # Errno 32 = Broken pipe, Errno 54 = Connection reset by peer
+                return any(x in msg for x in ["broken pipe", "connection closed", "connection reset", "errno 32", "errno 54", "ssh connection"])
             
             try:
                 for idx, input_file in enumerate(files_to_convert, 1):
@@ -1282,19 +1287,10 @@ def main_async_remote(args) -> int:
                             
                             # Determine output path
                             output_dir = Path(args.out_dir)
+                            output_file = get_output_path(input_file, input_base, output_dir, args.flat)
                             
-                            # Preserve directory structure if not flat
-                            if not args.flat and input_path.is_dir():
-                                try:
-                                    rel_path = input_file.relative_to(input_path)
-                                    output_subdir = output_dir / rel_path.parent
-                                except ValueError:
-                                    output_subdir = output_dir
-                            else:
-                                output_subdir = output_dir
-                            
-                            output_subdir.mkdir(parents=True, exist_ok=True)
-                            output_file = output_subdir / f"{input_file.stem}.md"
+                            # Ensure output directory exists
+                            output_file.parent.mkdir(parents=True, exist_ok=True)
                             
                             # Check if output exists and skip if not overwrite
                             if output_file.exists() and not args.overwrite:
@@ -1336,7 +1332,10 @@ def main_async_remote(args) -> int:
                             while conversion_attempt < 3 and not conversion_success:
                                 try:
                                     if conversion_attempt > 0 and not args.quiet:
-                                        print(f"  ↻ Conversion retry {conversion_attempt}...", file=sys.stderr)
+                                        # Exponential backoff: 2s, 4s, 8s
+                                        backoff_delay = 2 ** conversion_attempt
+                                        print(f"  ↻ Conversion retry {conversion_attempt} (waiting {backoff_delay}s for server recovery)...", file=sys.stderr)
+                                        await asyncio.sleep(backoff_delay)
                                     
                                     conversion_output, _, conv_code = await ssh_client.run_command(convert_cmd, timeout=remote_conversion_timeout)
                                     
@@ -1344,19 +1343,39 @@ def main_async_remote(args) -> int:
                                         conversion_success = True
                                         break
                                     else:
-                                        conversion_attempt += 1
+                                        # Non-zero exit code - fail without retry for non-connection errors
+                                        break
                                 except Exception as conv_exc:
-                                    if is_connection_error(conv_exc) and conversion_attempt < 2:
+                                    is_conn_err = is_connection_error(conv_exc)
+                                    if is_conn_err and conversion_attempt < 2:
                                         conversion_attempt += 1
                                         if not args.quiet:
-                                            print(f"  ↻ Connection lost during conversion. Reconnecting (attempt {conversion_attempt})...", file=sys.stderr)
+                                            # Exponential backoff: 5s, 10s
+                                            backoff_delay = 5 * conversion_attempt
+                                            print(f"  ↻ Connection reset during conversion. Reconnecting in {backoff_delay}s...", file=sys.stderr)
+                                        
+                                        await asyncio.sleep(backoff_delay)
+                                        
                                         try:
                                             await ssh_client.disconnect()
                                         except Exception:
                                             pass
-                                        await ssh_client.connect()
+                                        
+                                        # Reconnect with retry
+                                        try:
+                                            await ssh_client.connect()
+                                        except Exception:
+                                            if not args.quiet:
+                                                print(f"  ⚠ Reconnection failed: retrying...", file=sys.stderr)
+                                            continue
                                     else:
-                                        conversion_attempt += 1
+                                        # Either not a connection error, or we've exhausted retries
+                                        if not args.quiet:
+                                            print(f"  [DEBUG] Breaking loop: not conn_err or exhausted retries", file=sys.stderr)
+                                        if conversion_attempt >= 2 and is_conn_err:
+                                            if not args.quiet:
+                                                print(f"  ↻ Connection error on final retry attempt", file=sys.stderr)
+                                        break
                             
                             if not conversion_success:
                                 print(f"  ✗ Failed: Conversion failed after {conversion_attempt} attempt(s)", file=sys.stderr)
@@ -1396,12 +1415,37 @@ def main_async_remote(args) -> int:
                                         # Ultimate fallback
                                         markdown_content = conversion_output
                                 
-                                # Write markdown content to remote file
-                                write_cmd = f"cat > {remote_output_path} << 'MDIFY_EOF'\n{markdown_content}\nMDIFY_EOF"
-                                _, _, write_code = await ssh_client.run_command(write_cmd, timeout=30)
+                                # Write markdown content to local temp file first, then upload via SFTP
+                                # (Piping large content through SSH here-documents can crash the connection)
+                                content_size_kb = len(markdown_content) / 1024
+                                if not args.quiet:
+                                    print(f"  {color.cyan('Writing')} {content_size_kb:.1f}KB markdown via SFTP...", file=sys.stderr)
                                 
-                                if write_code != 0:
-                                    print(f"  ✗ Failed to write markdown output", file=sys.stderr)
+                                try:
+                                    # Write to temporary local file
+                                    with tempfile.NamedTemporaryFile(mode='w', suffix='.md', delete=False) as temp_file:
+                                        temp_file.write(markdown_content)
+                                        temp_path = temp_file.name
+                                    
+                                    # Upload via SFTP (more reliable for large files)
+                                    await transfer_manager.upload_file(
+                                        local_path=temp_path,
+                                        remote_path=remote_output_path,
+                                        overwrite=True,
+                                        compress=False,
+                                    )
+                                    
+                                    # Cleanup temp file
+                                    try:
+                                        os.unlink(temp_path)
+                                    except Exception:
+                                        pass
+                                    
+                                    if not args.quiet:
+                                        print(f"  {color.green('✓')} Markdown written", file=sys.stderr)
+                                except Exception as write_exc:
+                                    if not args.quiet:
+                                        print(f"  ✗ Failed to write markdown: {write_exc}", file=sys.stderr)
                                     failed += 1
                                     break
                                 

{mdify_cli-3.0.8.dist-info → mdify_cli-3.1.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mdify-cli
-Version: 3.0.8
+Version: 3.1.1
 Summary: Convert PDFs and document images into structured Markdown for LLM workflows
 Author: tiroq
 License-Expression: MIT

{mdify_cli-3.0.8.dist-info → mdify_cli-3.1.1.dist-info}/RECORD

@@ -1,7 +1,7 @@
 assets/mdify.png,sha256=qUj7WXWqNwpI2KNXOW79XJwqFqa-UI0JEkmt1mmy4Rg,1820418
-mdify/__init__.py,sha256=iUfoOWZ_3PErQD0XOXLNy8UyG2lK1eTPJFlW9l5NDdc,90
+mdify/__init__.py,sha256=_C45QsMJxMtjDM6sIpFOC7912SJbhfrBhrZ06ZjrAO8,90
 mdify/__main__.py,sha256=bhpJ00co6MfaVOdH4XLoW04NtLYDa_oJK7ODzfLrn9M,143
-mdify/cli.py,sha256=-HYbPCEjRU4iRcurfRpmMCdbHtQS6oNah-zajkzjojA,78732
+mdify/cli.py,sha256=Ro_v7i1o6rszfAUTu3W52t5OO5kR_dda1CWbnjfXgE8,81843
 mdify/container.py,sha256=BjL5ZR__n1i_WHifXKllTPoqO7IuOUdPDo5esuNg0Iw,8213
 mdify/docling_client.py,sha256=xuQR6sC1v3EPloOSwExoHCqT4uUxE8myYq-Yeby3C2I,7975
 mdify/formatting.py,sha256=lJKhMbDPcaWCdyEa7aKwAm_desaWvkfDc8C3EP7LWp4,790
@@ -10,9 +10,9 @@ mdify/ssh/client.py,sha256=nO7gAQ6eWxIXFIOplW2F2ya0-1ZEFlLmz3ovi1TEFTg,14997
 mdify/ssh/models.py,sha256=jpbDS1yGhd7Xwq2tW7bZv14mTBlR8DCfhT4x-Xf2Wq4,17676
 mdify/ssh/remote_container.py,sha256=KB8rrsp3h5s_2BKPBXhzkOdQFafyPGQhSnFbK20b4yQ,10230
 mdify/ssh/transfer.py,sha256=aQuWa_B81McrgZBBWo_CxnjwoiGHmoxAoqObm19JAk8,10940
-mdify_cli-3.0.8.dist-info/licenses/LICENSE,sha256=NWM66Uv-XuSMKaU-gaPmvfyk4WgE6zcIPr78wyg6GAo,1065
-mdify_cli-3.0.8.dist-info/METADATA,sha256=ap0J1JAN1eHngW32_OFd_5jQJrZrm4gcvCEcOMHNGNI,14766
-mdify_cli-3.0.8.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
-mdify_cli-3.0.8.dist-info/entry_points.txt,sha256=0Xki8f5lADQUtwdt6Eq_FEaieI6Byhk8UE7BuDhChMg,41
-mdify_cli-3.0.8.dist-info/top_level.txt,sha256=qltzf7h8owHq7dxCdfCkSHY8gT21hn1_E8P-VWS_OKM,6
-mdify_cli-3.0.8.dist-info/RECORD,,
+mdify_cli-3.1.1.dist-info/licenses/LICENSE,sha256=NWM66Uv-XuSMKaU-gaPmvfyk4WgE6zcIPr78wyg6GAo,1065
+mdify_cli-3.1.1.dist-info/METADATA,sha256=KFklqUB5LjJqiDqynre9fX00KhTwIDdaFAOL8_uwPtE,14766
+mdify_cli-3.1.1.dist-info/WHEEL,sha256=wUyA8OaulRlbfwMtmQsvNngGrxQHAvkKcvRmdizlJi0,92
+mdify_cli-3.1.1.dist-info/entry_points.txt,sha256=0Xki8f5lADQUtwdt6Eq_FEaieI6Byhk8UE7BuDhChMg,41
+mdify_cli-3.1.1.dist-info/top_level.txt,sha256=qltzf7h8owHq7dxCdfCkSHY8gT21hn1_E8P-VWS_OKM,6
+mdify_cli-3.1.1.dist-info/RECORD,,