Merge pull request #124 from AI4WA/feature/cli

Feature/cli

Docs2KG/cli.py

@@ -38,29 +38,18 @@ def get_supported_formats(cls) -> str:
         return ", ".join(cls.PROCESSORS.keys())
 
 
-def setup_environment(config_path: str = None):
-    """Setup the environment with the config file."""
-    if config_path:
-        os.environ["CONFIG_FILE"] = str(Path(config_path).resolve())
-    else:
-        os.environ["CONFIG_FILE"] = str(Path.cwd() / "config.yml")
-
-
 @click.group()
-@click.option(
-    "--config",
-    "-c",
-    type=click.Path(exists=True),
-    help="Path to the configuration file (default: ./config.yml)",
-)
-def cli(config):
+def cli():
     """Docs2KG - Document to Knowledge Graph conversion tool.
 
     Supports multiple document formats: PDF, DOCX, HTML, and EPUB.
     """
-    setup_environment(config)
-    logger.info(f"Using configuration: {os.environ.get('CONFIG_FILE')}")
-    logger.info(PROJECT_CONFIG.data)
+    logger.info("Welcome to Docs2KG!")
+    logger.info(f"Configuration loaded from: {os.environ.get('CONFIG_FILE')}")
+    logger.info(f"Input directory: {PROJECT_CONFIG.data.input_dir}")
+    logger.info(f"Output directory: {PROJECT_CONFIG.data.output_dir}")
+    logger.info(f"Ontology directory: {PROJECT_CONFIG.data.ontology_dir}")
+    logger.info("---")
 
 
 def process_single_file(
@@ -154,7 +143,11 @@ def process_document(file_path, project_id, agent_name, agent_type):
 
 
 @cli.command()
-@click.argument("input_dir", type=click.Path(exists=True))
+@click.argument(
+    "input_dir",
+    type=click.Path(exists=True),
+    default=PROJECT_CONFIG.data.input_dir.as_posix(),
+)
 @click.option(
     "--project-id",
     "-p",

Docs2KG/digitization/native/html_parser.py

@@ -17,16 +17,16 @@ class HTMLDocling(DigitizationBase):
     HTMLDocling class for processing HTML content from files or URLs to markdown.
     """
 
-    def __init__(self, source: Union[str, Path]):
-        self.is_url = isinstance(source, str) and self._is_valid_url(source)
+    def __init__(self, file_path: Union[str, Path]):
+        self.is_url = isinstance(file_path, str) and self._is_valid_url(file_path)
 
         if self.is_url:
             # Create a filename from the URL
-            url_path = urlparse(source).path
+            url_path = urlparse(file_path).path
             url_filename = (
                 unquote(Path(url_path).name)
                 if url_path and Path(url_path).name
-                else urlparse(source).netloc
+                else urlparse(file_path).netloc
             )
             self.html_filename = (
                 f"{url_filename}.html"
@@ -36,16 +36,16 @@ def __init__(self, source: Union[str, Path]):
 
             # Download and save the HTML content
             self.html_path = PROJECT_CONFIG.data.input_dir / self.html_filename
-            self._download_and_save_html(source)
+            self._download_and_save_html(file_path)
 
             # Use the saved file path as the source
-            source = self.html_path
+            file_path = self.html_path
 
         super().__init__(
-            file_path=source,
+            file_path=file_path,
             supported_formats=["html", "htm"],
         )
-        self.source = source
+        self.source = file_path
 
     def _download_and_save_html(self, url: str) -> None:
         """

README.md

@@ -38,6 +38,8 @@ You have two ways to run the package:
 ### Command Line
 
 ```bash
+# first setup the CONFIG_FILE environment variable to local one
+export CONFIG_FILE=config.yml # or any other path for the configuration file
 docs2kg # this command will tell you how to use the package
 
 # we currently support the following commands

docs/index.md

@@ -30,6 +30,9 @@ You have two ways to run the package:
 ### Command Line
 
 ```bash
+# first setup the CONFIG_FILE environment variable to local one
+export CONFIG_FILE=config.yml # or any other path for the configuration file
+
 docs2kg # this command will tell you how to use the package
 
 # we currently support the following commands