fix(sync): preserve bmignore rclone filters (#827)

Signed-off-by: phernandez <paul@basicmachines.co>

Author: phernandez

.github/workflows/test.yml

@@ -55,7 +55,7 @@ jobs:
 
   test-sqlite-unit:
     name: Test SQLite Unit (${{ matrix.os }}, Python ${{ matrix.python-version }})
-    timeout-minutes: 30
+    timeout-minutes: 45
     strategy:
       fail-fast: false
       matrix:
@@ -147,7 +147,7 @@ jobs:
 
   test-postgres-unit:
     name: Test Postgres Unit (Python ${{ matrix.python-version }})
-    timeout-minutes: 30
+    timeout-minutes: 60
     strategy:
       fail-fast: false
       matrix:

docs/cloud-cli.md

@@ -582,33 +582,62 @@ bm cloud logout
 **Default patterns:**
 
 ```gitignore
+# Hidden files and directories
+.*
+
+# Basic Memory internals
+*.db
+*.db-shm
+*.db-wal
+config.json
+
 # Version control
-.git/**
+.git
+.svn
 
 # Python
-__pycache__/**
+__pycache__
 *.pyc
-.venv/**
-venv/**
+*.pyo
+*.pyd
+.pytest_cache
+.coverage
+*.egg-info
+.tox
+.mypy_cache
+.ruff_cache
+
+# Virtual environments
+.venv
+venv
+env
+.env
 
 # Node.js
-node_modules/**
+node_modules
 
-# Basic Memory internals
-memory.db/**
-memory.db-shm/**
-memory.db-wal/**
-config.json/**
-watch-status.json/**
-.bmignore.rclone/**
+# Build artifacts
+build
+dist
+.cache
+
+# IDE
+.idea
+.vscode
 
 # OS files
-.DS_Store/**
-Thumbs.db/**
+.DS_Store
+Thumbs.db
+desktop.ini
+
+# Obsidian
+.obsidian
 
-# Environment files
-.env/**
-.env.local/**
+# Temporary files
+*.tmp
+*.swp
+*.swo
+*~
 ```
 
 **How it works:**
@@ -617,14 +646,19 @@ Thumbs.db/**
 3. Rclone uses filters during sync
 4. Same patterns used by all projects
 
+During conversion, file patterns exclude the direct match and recursive contents.
+For example, `config.json` becomes both `- config.json` and `- config.json/**`,
+while `.*` becomes both `- .*` and `- .*/**`. Directory-only patterns keep
+their trailing slash, so `cache/` becomes `- cache/` and `- cache/**`.
+
 **Customizing:**
 
 ```bash
 # Edit patterns
 code ~/.basic-memory/.bmignore
 
 # Add custom patterns
-echo "*.tmp/**" >> ~/.basic-memory/.bmignore
+echo "*.tmp" >> ~/.basic-memory/.bmignore
 
 # Next sync uses updated patterns
 bm project bisync --name research

src/basic_memory/cli/commands/cloud/bisync_commands.py

@@ -14,6 +14,24 @@ class BisyncError(Exception):
     pass
 
 
+def _rclone_exclude_filters(pattern: str) -> list[str]:
+    """Return rclone exclude filters for a gitignore-style pattern."""
+    if pattern.endswith("/"):
+        # Trigger: gitignore-style patterns ending in / are directory-only rules.
+        # Why: stripping the slash would also exclude a same-named file.
+        # Outcome: rclone keeps the directory rule and excludes recursive contents.
+        return [f"- {pattern}", f"- {pattern}**"]
+
+    path_pattern = pattern.removesuffix("/**")
+
+    # Trigger: rclone treats a directory contents filter separately from the
+    # directory/file path itself.
+    # Why: files like config.json and directory markers like .obsidian must both
+    # be excluded, along with anything below matching directories.
+    # Outcome: every ignore pattern excludes the direct match and recursive children.
+    return [f"- {path_pattern}", f"- {path_pattern}/**"]
+
+
 async def get_mount_info() -> TenantMountInfo:
     """Get current tenant information from cloud API."""
     try:
@@ -78,19 +96,11 @@ def convert_bmignore_to_rclone_filters() -> Path:
                     patterns.append(line)
                     continue
 
-                # Convert gitignore pattern to rclone filter syntax
-                # gitignore: node_modules  → rclone: - node_modules/**
-                # gitignore: *.pyc        → rclone: - *.pyc
-                if "*" in line:
-                    # Pattern already has wildcard, just add exclude prefix
-                    patterns.append(f"- {line}")
-                else:
-                    # Directory pattern - add /** for recursive exclude
-                    patterns.append(f"- {line}/**")
+                patterns.extend(_rclone_exclude_filters(line))
 
     except Exception:
         # If we can't read the file, create a minimal filter
-        patterns = ["# Error reading .bmignore, using minimal filters", "- .git/**"]
+        patterns = ["# Error reading .bmignore, using minimal filters", "- .git", "- .git/**"]
 
     # Write rclone filter file
     rclone_filter_path.write_text("\n".join(patterns) + "\n")

tests/cli/cloud/test_rclone_config_and_bmignore_filters.py

@@ -32,13 +32,48 @@ def test_convert_bmignore_to_rclone_filters_creates_and_converts(config_home):
     # Comments/empties preserved
     assert "# comment" in content
     assert "" in content
-    # Directory pattern becomes recursive exclude
+    # Plain and wildcard patterns exclude direct matches and recursive contents.
+    assert "- node_modules" in content
     assert "- node_modules/**" in content
-    # Wildcard pattern becomes simple exclude
     assert "- *.pyc" in content
+    assert "- *.pyc/**" in content
+    assert "- .git" in content
     assert "- .git/**" in content
 
 
+def test_convert_bmignore_to_rclone_filters_excludes_files_and_hidden_directory_contents(
+    config_home,
+):
+    bmignore = get_bmignore_path()
+    bmignore.parent.mkdir(parents=True, exist_ok=True)
+    bmignore.write_text("config.json\n.*\nnode_modules/**\n", encoding="utf-8")
+
+    rclone_filter = convert_bmignore_to_rclone_filters()
+    content = rclone_filter.read_text(encoding="utf-8").splitlines()
+
+    assert "- config.json" in content
+    assert "- config.json/**" in content
+    assert "- .*" in content
+    assert "- .*/**" in content
+    assert "- node_modules" in content
+    assert "- node_modules/**" in content
+
+
+def test_convert_bmignore_to_rclone_filters_preserves_directory_only_patterns(config_home):
+    bmignore = get_bmignore_path()
+    bmignore.parent.mkdir(parents=True, exist_ok=True)
+    bmignore.write_text("cache/\nconfig.json/**\n", encoding="utf-8")
+
+    rclone_filter = convert_bmignore_to_rclone_filters()
+    content = rclone_filter.read_text(encoding="utf-8").splitlines()
+
+    assert "- cache/" in content
+    assert "- cache/**" in content
+    assert "- cache" not in content
+    assert "- config.json" in content
+    assert "- config.json/**" in content
+
+
 def test_convert_bmignore_to_rclone_filters_is_cached_when_up_to_date(config_home):
     bmignore = get_bmignore_path()
     bmignore.parent.mkdir(parents=True, exist_ok=True)