apply ruff

osm/_utils.py

@@ -5,7 +5,6 @@
 import time
 import types
 from pathlib import Path
-from time import sleep
 
 import pandas as pd
 import requests

osm/pipeline/core.py

@@ -15,7 +15,7 @@ def __init__(self, version: str = "0.0.1"):
         self._orm_model = None
 
     @abstractmethod
-    def _run(self, data: bytes|dict, **kwargs) -> Any:
+    def _run(self, data: bytes | dict, **kwargs) -> Any:
         """Abstract method that subclasses must implement."""
         pass
 
@@ -99,9 +99,11 @@ def __init__(
         self.xml_path = xml_path
         self.metrics_path = metrics_path
 
-    def run(self,user_managed_compose:bool=False):
+    def run(self, user_managed_compose: bool = False):
         for parser in self.parsers:
-            parsed_data = parser.run(self.file_data,user_managed_compose=user_managed_compose)
+            parsed_data = parser.run(
+                self.file_data, user_managed_compose=user_managed_compose
+            )
             if isinstance(parsed_data, bytes):
                 self.savers.save_file(parsed_data, self.xml_path)
             for extractor in self.extractors:

osm/pipeline/parsers.py

@@ -1,9 +1,11 @@
-import requests
+import io
 import time
+
+import requests
+
 from osm.schemas.custom_fields import LongBytes
 
 from .core import Component
-import io
 
 SCIENCEBEAM_URL = "http://localhost:8070/api/convert"
 
@@ -28,17 +30,19 @@ class PMCParser(NoopParser):
 
 
 class ScienceBeamParser(Component):
-    def _run(self, data: bytes,user_managed_compose=False) -> str:
+    def _run(self, data: bytes, user_managed_compose=False) -> str:
         self.sample = LongBytes(data)
         headers = {"Accept": "application/tei+xml", "Content-Type": "application/pdf"}
-        files = {'file': ('input.pdf', io.BytesIO(data), 'application/pdf')}
+        files = {"file": ("input.pdf", io.BytesIO(data), "application/pdf")}
         for attempt in range(5):
             try:
                 if not user_managed_compose:
                     time.sleep(10)
                 response = requests.post(SCIENCEBEAM_URL, files=files, headers=headers)
-            except requests.exceptions.RequestException as e:
-                print(f"Attempt {attempt + 1} for parsing the file failed. This can happen while the container is starting up. Retrying in 5 seconds.")
+            except requests.exceptions.RequestException:
+                print(
+                    f"Attempt {attempt + 1} for parsing the file failed. This can happen while the container is starting up. Retrying in 5 seconds."
+                )
                 continue
             if response.status_code == 200:
                 return response.content