chores

Author: Integer-Ctrl

README.md

@@ -163,7 +163,7 @@ databusclient download 'PREFIX dcat: <http://www.w3.org/ns/dcat#> SELECT ?x WHER
 databusclient deploy --help
 ```
 ```
-Usage: databusclient deploy [OPTIONS] [INPUTS]...
+Usage: databusclient deploy [OPTIONS] [DISTRIBUTIONS]...
 
  Flexible deploy to databus command:
 
@@ -174,7 +174,7 @@ Usage: databusclient deploy [OPTIONS] [INPUTS]...
  - Upload & deploy via Nextcloud
 
 Arguments:
-  INPUTS...  Depending on mode:
+  DISTRIBUTIONS...  Depending on mode:
              - Classic mode: List of distributions in the form
                URL|CV|fileext|compression|sha256sum:contentlength
                (where URL is the download URL and CV the key=value pairs,
@@ -200,7 +200,7 @@ Options:
   
 ```
 #### Examples of using deploy command
-Mode 1: Classic Deploy (Distributions)
+##### Mode 1: Classic Deploy (Distributions)
 ```
 databusclient deploy --version-id https://databus.dbpedia.org/user1/group1/artifact1/2022-05-18 --title title1 --abstract abstract1 --description description1 --license http://dalicc.net/licenselibrary/AdaptivePublicLicense10 --apikey MYSTERIOUS 'https://raw.githubusercontent.com/dbpedia/databus/master/server/app/api/swagger.yml|type=swagger'  
 ```
@@ -215,7 +215,7 @@ A few more notes for CLI usage:
   * If other parameters are used, you need to leave them empty like `https://raw.githubusercontent.com/dbpedia/databus/master/server/app/api/swagger.yml||yml|7a751b6dd5eb8d73d97793c3c564c71ab7b565fa4ba619e4a8fd05a6f80ff653:367116`
 
 
-Mode 2: Deploy with Metadata File
+##### Mode 2: Deploy with Metadata File
 
 Use a JSON metadata file to define all distributions.
 The metadata.json should list all distributions and their metadata.
@@ -251,7 +251,7 @@ Metadata file structure (file_format and compression are optional):
 ```
 
 
-Mode 3: Upload & Deploy via Nextcloud
+##### Mode 3: Upload & Deploy via Nextcloud
 
 Upload local files or folders to a WebDAV/Nextcloud instance and automatically deploy to DBpedia Databus. 
 Rclone is required.

databusclient/cli.py

@@ -1,7 +1,6 @@
 #!/usr/bin/env python3
 import json
 import os
-import re
 
 import click
 from typing import List
@@ -34,30 +33,29 @@ def app():
 @click.option("--remote", help="rclone remote name (e.g., 'nextcloud')")
 @click.option("--path", help="Remote path on Nextcloud (e.g., 'datasets/mydataset')")
 
-@click.argument("inputs", nargs=-1)
+@click.argument("distributions", nargs=-1)
 def deploy(version_id, title, abstract, description, license_url, apikey,
-           metadata_file, webdav_url, remote, path, inputs: List[str]):
+           metadata_file, webdav_url, remote, path, distributions: List[str]):
     """
-    Flexible deploy to databus command:\n
-    - Classic dataset deployment\n
-    - Metadata-based deployment\n
-    - Upload & deploy via Nextcloud
+    Flexible deploy to Databus command supporting three modes:\n
+    - Classic deploy (distributions as arguments)\n
+    - Metadata-based deploy (--metadata <file>)\n
+    - Upload & deploy via Nextcloud (--webdav-url, --remote, --path)
     """
 
-    # === Mode 1: Upload & Deploy (Nextcloud) ===
-    if webdav_url and remote and path:
-        if not inputs:
-            raise click.UsageError("Please provide files to upload when using WebDAV/Nextcloud mode.")
+    # Sanity checks for conflicting options
+    if metadata_file and any([distributions, webdav_url, remote, path]):
+        raise click.UsageError("Invalid combination: when using --metadata, do not provide --webdav-url, --remote, --path, or distributions.")
+    if any([webdav_url, remote, path]) and not all([webdav_url, remote, path]):
+        raise click.UsageError("Invalid combination: when using WebDAV/Nextcloud mode, please provide --webdav-url, --remote, and --path together.")
 
-        #Check that all given paths exist and are files or directories.#
-        invalid = [f for f in inputs if not os.path.exists(f)]
-        if invalid:
-            raise click.UsageError(f"The following input files or folders do not exist: {', '.join(invalid)}")
+    # === Mode 1: Classic Deploy ===
+    if distributions and not (metadata_file or webdav_url or remote or path):
+        click.echo("[MODE] Classic deploy with distributions")
+        click.echo(f"Deploying dataset version: {version_id}")
 
-        click.echo("[MODE] Upload & Deploy to DBpedia Databus via Nextcloud")
-        click.echo(f"→ Uploading to: {remote}:{path}")
-        metadata = upload.upload_to_nextcloud(inputs, remote, path, webdav_url)
-        client.deploy_from_metadata(metadata, version_id, title, abstract, description, license_url, apikey)
+        dataid = client.create_dataset(version_id, title, abstract, description, license_url, distributions)
+        client.deploy(dataid=dataid, api_key=apikey)
         return
 
     # === Mode 2: Metadata File ===
@@ -67,20 +65,21 @@ def deploy(version_id, title, abstract, description, license_url, apikey,
             metadata = json.load(f)
         client.deploy_from_metadata(metadata, version_id, title, abstract, description, license_url, apikey)
         return
+    
+    # === Mode 3: Upload & Deploy (Nextcloud) ===
+    if webdav_url and remote and path:
+        if not distributions:
+            raise click.UsageError("Please provide files to upload when using WebDAV/Nextcloud mode.")
 
-    # === Mode 3: Classic Deploy ===
-    if inputs:
-        invalid = client.validate_distributions(inputs)
+        #Check that all given paths exist and are files or directories.#
+        invalid = [f for f in distributions if not os.path.exists(f)]
         if invalid:
-            raise click.UsageError(
-                f"The following distributions are not in a valid format:\n"
-                + "\n".join(invalid)
-                + "\nExpected format example:\n"
-                  "https://example.com/file.ttl|format=ttl|gzip|abcdef123456789:12345"
-            )
-        click.echo("[MODE] Classic deploy with distributions")
-        dataid = client.create_dataset(version_id, title, abstract, description, license_url, inputs)
-        client.deploy(dataid=dataid, api_key=apikey)
+            raise click.UsageError(f"The following input files or folders do not exist: {', '.join(invalid)}")
+
+        click.echo("[MODE] Upload & Deploy to DBpedia Databus via Nextcloud")
+        click.echo(f"→ Uploading to: {remote}:{path}")
+        metadata = upload.upload_to_nextcloud(distributions, remote, path, webdav_url)
+        client.deploy_from_metadata(metadata, version_id, title, abstract, description, license_url, apikey)
         return
 
     raise click.UsageError(

databusclient/client.py

@@ -7,7 +7,6 @@
 from SPARQLWrapper import SPARQLWrapper, JSON
 from hashlib import sha256
 import os
-import re
 
 __debug = False
 
@@ -256,26 +255,6 @@ def create_distributions_from_metadata(metadata: List[Dict[str, Union[str, int]]
         counter += 1
     return distributions
 
-
-def validate_distributions(distros: List[str]) -> List[str]:
-    """
-    Check that all distributions follow the pattern:
-    url|key=value|[format]|[compression]|[sha256:len]
-
-    Parameters
-    ----------
-    distros: List[str]
-        List of distribution identifiers to validate
-
-    Returns
-    -------
-    List[str]
-        List of invalid distribution identifier strings
-    """
-    pattern = re.compile(r"^https?://[^|]+\|.+$")
-    return [d for d in distros if not pattern.match(d)]
-
-
 def create_dataset(
     version_id: str,
     title: str,
@@ -754,7 +733,7 @@ def __download_list__(urls: List[str],
 def __get_databus_id_parts__(uri: str) -> Tuple[Optional[str], Optional[str], Optional[str], Optional[str], Optional[str], Optional[str]]:
     uri = uri.removeprefix("https://").removeprefix("http://")
     parts = uri.strip("/").split("/")
-    parts += [None] * (6 - len(parts))  # pad fwith None if less than 6 parts
+    parts += [None] * (6 - len(parts))  # pad with None if less than 6 parts
     return tuple(parts[:6])  # return only the first 6 parts