rosetta-cli 2.0.4__tar.gz → 2.0.6__tar.gz

{rosetta_cli-2.0.4/rosetta_cli.egg-info → rosetta_cli-2.0.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: rosetta-cli
-Version: 2.0.4
+Version: 2.0.6
 Summary: Rosetta CLI for publishing knowledge base content to RAGFlow
 Author: Igor Solomatov
 License-Expression: Apache-2.0
@@ -89,6 +89,7 @@ Complete setup instructions are in [docs/QUICKSTART.md](../docs/QUICKSTART.md).
 ### Installed Usage
 
 ```bash
+uvx rosetta-cli@latest version
 uvx rosetta-cli@latest verify
 ```
 
@@ -105,6 +106,12 @@ venv/bin/rosetta-cli verify
 
 All commands support `--env <environment>` flag to override the active environment.
 
+### Version
+
+```bash
+uvx rosetta-cli@latest version
+```
+
 ### Publishing Commands
 
 #### Publish Knowledge Base Content

{rosetta_cli-2.0.4 → rosetta_cli-2.0.6}/README.md

@@ -61,6 +61,7 @@ Complete setup instructions are in [docs/QUICKSTART.md](../docs/QUICKSTART.md).
 ### Installed Usage
 
 ```bash
+uvx rosetta-cli@latest version
 uvx rosetta-cli@latest verify
 ```
 
@@ -77,6 +78,12 @@ venv/bin/rosetta-cli verify
 
 All commands support `--env <environment>` flag to override the active environment.
 
+### Version
+
+```bash
+uvx rosetta-cli@latest version
+```
+
 ### Publishing Commands
 
 #### Publish Knowledge Base Content

{rosetta_cli-2.0.4 → rosetta_cli-2.0.6}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "rosetta-cli"
-version = "2.0.4"
+version = "2.0.6"
 description = "Rosetta CLI for publishing knowledge base content to RAGFlow"
 readme = "README.md"
 requires-python = ">=3.12"

{rosetta_cli-2.0.4 → rosetta_cli-2.0.6}/rosetta_cli/cli.py

@@ -30,6 +30,11 @@ COMMAND_REGISTRY: dict[str, CommandClass] = {
 }
 
 
+def _print_version() -> None:
+    """Print the CLI version."""
+    print(f"Rosetta Version: {_CLI_VERSION}")
+
+
 def execute_command(command_name: str, args: CommandArgs, client: RAGFlowClient, config: IMSConfig) -> int:
     """
     Execute a command by name using the command registry.
@@ -60,6 +65,9 @@ def main() -> int:
         formatter_class=argparse.RawDescriptionHelpFormatter,
         epilog="""
 Examples:
+  # Print CLI version
+  rosetta-cli version
+
   # Publish knowledge base content from a folder (with timing)
   rosetta-cli publish ../instructions
 
@@ -217,6 +225,12 @@ Frontmatter Metadata (publish flow):
         default=None,
         help='Explicit path to a .env file'
     )
+
+    # Version command
+    subparsers.add_parser(
+        'version',
+        help='Print CLI version and exit'
+    )
     
     # List dataset command
     list_parser = subparsers.add_parser(
@@ -340,6 +354,10 @@ Frontmatter Metadata (publish flow):
     if not args.command:
         parser.print_help()
         return 1
+
+    if args.command == 'version':
+        _print_version()
+        return 0
     
     try:
         # Load configuration
@@ -348,7 +366,7 @@ Frontmatter Metadata (publish flow):
         # Validate configuration
         config.validate()
 
-        print(f"Rosetta Version: {_CLI_VERSION}")
+        _print_version()
         print(f"Rosetta Environment: {config.environment}")
         
         # Initialize RAGFlow client

{rosetta_cli-2.0.4 → rosetta_cli-2.0.6}/rosetta_cli/commands/cleanup_command.py

@@ -19,6 +19,10 @@ class CleanupCommand(BaseCommand):
     def execute(self, args: CommandArgs) -> int:
         """Execute cleanup-dataset command."""
         self._start_timing()
+
+        # Verify authentication before any dataset auto-detection touches the API.
+        from ..services.auth_service import AuthService
+        AuthService.verify_or_exit(self.client, self.config)
         
         # Resolve dataset name
         dataset_service = DatasetService(self.client, self.config)
@@ -32,10 +36,6 @@ class CleanupCommand(BaseCommand):
         print(f"Environment: {self.config.environment}")
         print(f"RAGFlow Instance: {self.config.base_url}\n")
         
-        # Verify authentication
-        from ..services.auth_service import AuthService
-        AuthService.verify_or_exit(self.client, self.config)
-        
         try:
             # Get dataset
             dataset = self.client.get_dataset(name=dataset_name)

{rosetta_cli-2.0.4 → rosetta_cli-2.0.6}/rosetta_cli/commands/list_command.py

@@ -15,6 +15,10 @@ class ListCommand(BaseCommand):
     def execute(self, args: CommandArgs) -> int:
         """Execute list-dataset command."""
         self._start_timing()
+
+        # Verify authentication before any dataset auto-detection touches the API.
+        from ..services.auth_service import AuthService
+        AuthService.verify_or_exit(self.client, self.config)
         
         # Resolve dataset name
         dataset_service = DatasetService(self.client, self.config)
@@ -30,10 +34,6 @@ class ListCommand(BaseCommand):
         print(f"RAGFlow Instance: {self.config.base_url}")
         print()
         
-        # Verify authentication
-        from ..services.auth_service import AuthService
-        AuthService.verify_or_exit(self.client, self.config)
-        
         try:
             # Get dataset
             dataset = self.client.get_dataset(name=dataset_name)

{rosetta_cli-2.0.4 → rosetta_cli-2.0.6}/rosetta_cli/commands/parse_command.py

@@ -27,6 +27,9 @@ class ParseCommand(BaseCommand):
 
         # CLI flag must override config default for this run.
         self.config.parse_timeout = args.parse_timeout
+
+        # Verify authentication before any dataset auto-detection touches the API.
+        AuthService.verify_or_exit(self.client, self.config)
         
         # Resolve dataset name
         dataset_service = DatasetService(self.client, self.config)
@@ -44,9 +47,6 @@ class ParseCommand(BaseCommand):
             print("DRY-RUN MODE - No parsing will be triggered")
         print(f"Environment: {self.config.environment}")
         print(f"RAGFlow Instance: {self.config.base_url}\n")
-        
-        # Verify authentication
-        AuthService.verify_or_exit(self.client, self.config)
         print()
         
         try:

{rosetta_cli-2.0.4 → rosetta_cli-2.0.6}/rosetta_cli/ims_publisher.py

@@ -391,7 +391,7 @@ class ContentPublisher:
             
             
             if dry_run:
-                print(f"[DRY RUN] Would publish: {file.name}")
+                print(f"[DRY RUN] Would publish: {metadata.get('doc_title', metadata.get('original_path', file.name))}")
                 print(f"  Document ID: {ims_doc_id}")
                 print(f"  Dataset: {dataset_name}")
                 print(f"  File type: {'text' if is_text else 'binary'}")
@@ -665,22 +665,16 @@ class ContentPublisher:
                 new_name = f"{re.sub(r'\(\d+\)$', '', path.stem)}{path.suffix}"
                 return name[: -len(path.name)] + new_name
 
-            # Delete docs with no metadata (ims_doc_id absent) — these are RAGFlow
-            # auto-renamed leftovers (e.g. SKILL(1).md) from uploads where metadata
-            # was never written. They are unrecoverable and must be cleaned up first.
+            # Collect all docs to delete into one list: (doc, label)
+            # Start with unmanaged: incomplete metadata (ims_doc_id or original_path absent)
+            duplicates: list[tuple[DocumentLike, str]] = []
             for doc in all_docs:
                 meta = getattr(doc, "meta_fields", {}) or {}
                 ims_doc_id = meta.get("ims_doc_id") if isinstance(meta, dict) else getattr(meta, "ims_doc_id", None)
-                if not ims_doc_id:
+                doc_original_path = meta.get("original_path", "") if isinstance(meta, dict) else getattr(meta, "original_path", "")
+                if not ims_doc_id or not doc_original_path:
                     doc_name = getattr(doc, "name", "") or doc.id
-                    if dry_run:
-                        print(f"  [DRY RUN] Would delete unmanaged (no metadata): {doc_name}")
-                    else:
-                        try:
-                            dataset.delete_documents([doc.id])
-                            print(f"  Deleted unmanaged (no metadata): {doc_name}")
-                        except Exception as e:
-                            print(f"  Warning: Failed to delete unmanaged doc '{doc_name}': {e}")
+                    duplicates.append((doc, doc_name))
 
             managed_docs = []
             for doc in all_docs:
@@ -697,29 +691,12 @@ class ContentPublisher:
                 if original_path:
                     by_path[original_path].append(doc)
 
-            duplicates = []  # (doc, original_path) to delete
+            # All copies of a duplicated original_path are deleted; publish creates a fresh copy.
             for original_path, docs in by_path.items():
                 if len(docs) <= 1:
                     continue
-
-                canonical_ims_doc_id = canonical_ids.get(original_path)
-                # Find the authoritative doc (matching local ims_doc_id)
-                authoritative: DocumentLike | None = None
-                if canonical_ims_doc_id:
-                    for doc in docs:
-                        meta = getattr(doc, "meta_fields", {}) or {}
-                        doc_ims_id = meta.get("ims_doc_id", "") if isinstance(meta, dict) else getattr(meta, "ims_doc_id", "")
-                        if doc_ims_id == canonical_ims_doc_id:
-                            authoritative = doc
-                            break
-
-                # Fallback: keep most recent (first in desc order)
-                if authoritative is None:
-                    authoritative = docs[0]
-
                 for doc in docs:
-                    if doc.id != authoritative.id:
-                        duplicates.append((doc, original_path))
+                    duplicates.append((doc, original_path))
 
             # Name duplicates: foo.md + foo(1).md + foo(2).md ...
             name_groups: dict[str, list[DocumentLike]] = defaultdict(list)

{rosetta_cli-2.0.4 → rosetta_cli-2.0.6/rosetta_cli.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: rosetta-cli
-Version: 2.0.4
+Version: 2.0.6
 Summary: Rosetta CLI for publishing knowledge base content to RAGFlow
 Author: Igor Solomatov
 License-Expression: Apache-2.0
@@ -89,6 +89,7 @@ Complete setup instructions are in [docs/QUICKSTART.md](../docs/QUICKSTART.md).
 ### Installed Usage
 
 ```bash
+uvx rosetta-cli@latest version
 uvx rosetta-cli@latest verify
 ```
 
@@ -105,6 +106,12 @@ venv/bin/rosetta-cli verify
 
 All commands support `--env <environment>` flag to override the active environment.
 
+### Version
+
+```bash
+uvx rosetta-cli@latest version
+```
+
 ### Publishing Commands
 
 #### Publish Knowledge Base Content