Fix fileinfo for arweave file

ocean_provider/file_types/definitions.py

@@ -72,7 +72,10 @@ def check_details(self, with_checksum=False):
                 if result:
                     status_code = result.status_code
                     headers = copy.deepcopy(result.headers)
-                    files_url = copy.deepcopy(result.url)
+                    if "arweave" in result.url:
+                        files_url = ""
+                    else:
+                        files_url = copy.deepcopy(result.url)
                     # always close requests session, see https://requests.readthedocs.io/en/latest/user/advanced/#body-content-workflow
                     result.close()
                     if status_code == 200:
@@ -85,18 +88,7 @@ def check_details(self, with_checksum=False):
                 file_name = None
 
                 if files_url:
-                    filename = urlparse(files_url).path.split("/")[-1]
-                    try:
-                        if not self._validate_filename(filename):
-                            msg = "Invalid file name format. It was not possible to get the file name."
-                            logger.error(msg)
-                            return False, {"error": msg}
-
-                        file_name = filename
-                    except Exception as e:
-                        msg = f"It was not possible to get the file name. {e}"
-                        logger.warning(msg)
-                        return False, {"error": msg}
+                    file_name = urlparse(files_url).path.split("/")[-1]
 
                 if not content_length and content_range:
                     # sometimes servers send content-range instead
@@ -128,10 +120,6 @@ def check_details(self, with_checksum=False):
 
         return False, {}
 
-    def _validate_filename(self, header: str) -> bool:
-        pattern = re.compile(r"\\|\.\.|/")
-        return not bool(pattern.findall(header))
-
     def _get_result_from_url(self, with_checksum=False):
         func, func_args = self._get_func_and_args()
 

ocean_provider/file_types/file_types.py

@@ -1,6 +1,7 @@
 import json
 import logging
 import os
+import re
 from typing import Any, Optional, Tuple
 from urllib.parse import urljoin
 from uuid import uuid4
@@ -34,11 +35,22 @@ def validate_dict(self) -> Tuple[bool, Any]:
         if self.method not in ["get", "post"]:
             return False, f"Unsafe method {self.method}."
 
+        if not self.validate_url(self.url):
+            msg = "Invalid file name format. It was not possible to get the file name."
+            logger.error(msg)
+            return False, msg
+
         return True, self
 
     def get_download_url(self):
         return self.url
 
+    def validate_url(self, url: str) -> bool:
+        pattern = re.compile(r"^(.+)\/([^/]+)$")
+        if url.startswith("http://") or url.startswith("https://"):
+            return True
+        return not bool(pattern.findall(url))
+
     @enforce_types
     def get_filename(self) -> str:
         return self.url.split("/")[-1]

ocean_provider/utils/test/test_util.py

@@ -390,6 +390,17 @@ def test_validate_url_object():
     assert result is False
     assert message == "Unsafe method delete."
 
+    url_object = {
+        "url": "./../dir",
+        "type": "url",
+        "method": "GET",
+    }
+    result, message = FilesTypeFactory.validate_and_create(url_object)
+    assert result is False
+    assert (
+        message == "Invalid file name format. It was not possible to get the file name."
+    )
+
 
 @pytest.mark.unit
 def test_build_download_response_ipfs():

tests/test_fileinfo.py

@@ -171,6 +171,7 @@ def test_check_arweave_good(client):
         assert file_info["contentType"] == "application/octet-stream"
         assert file_info["valid"] is True
         assert file_info["type"] == "arweave"
+        assert file_info["name"] == ""
 
 
 @pytest.mark.unit