Ruff fixes to CI

olmocr/bench/convert.py

@@ -2,7 +2,6 @@ import argparse
 import asyncio
 import glob
 import importlib
-import inspect
 import os
 
 from tqdm import tqdm
@@ -83,7 +82,7 @@ if __name__ == "__main__":
         "methods",
         nargs="+",
         help="Methods to run in the format method[:key=value ...]. "
-        "Example: gotocr mineru:temperature=2 marker:runs=3. "
+        "Example: gotocr mineru:temperature=2 marker:u=3. "
         "Use 'name=folder_name' to specify a custom output folder name.",
     )
     parser.add_argument("--repeats", type=int, default=1, help="Number of times to repeat the conversion for each PDF.")

olmocr/bench/miners/mine_diffs.py

@@ -7,7 +7,6 @@ from collections import Counter
 from difflib import SequenceMatcher
 
 import syntok.segmenter as segmenter
-import syntok.tokenizer as tokenizer
 from google import genai
 from google.genai import types
 

olmocr/bench/runners/run_olmocr.py

@@ -1,12 +1,7 @@
-import argparse
 import asyncio
 import json
 import logging
-import os
-import tempfile
 from dataclasses import dataclass
-from functools import partial
-from typing import Optional
 
 # Import necessary components from olmocr
 from olmocr.pipeline import (
@@ -58,13 +53,13 @@ async def run_olmocr(pdf_path: str, page_num: int = 1, temperature: float = 0.8)
     semaphore = asyncio.Semaphore(1)
 
     # Ensure server is running
-    server_task = None
+    _server_task = None
     try:
         await asyncio.wait_for(sglang_server_ready(), timeout=5)
         print("Using existing sglang server")
     except Exception:
         print("Starting new sglang server")
-        server_task = asyncio.create_task(sglang_server_host(args, semaphore))
+        _server_task = asyncio.create_task(sglang_server_host(args, semaphore))
         await sglang_server_ready()
 
     try:

olmocr/eval/dolma_refine/metrics.py

@@ -4,7 +4,7 @@ from typing import Type
 import regex as re
 from tqdm import tqdm
 
-from .aligners import BaseAligner
+from .aligners import BaseAligner, AlignerRegistry
 from .registry import BaseRegistry
 from .segmenters import BaseSegmenter, SegmenterRegistry
 

olmocr/eval/dolma_refine/segmenters.py

@@ -1,7 +1,5 @@
-import re
 from typing import Type
 
-import torch
 from spacy.lang.en import English
 
 from .registry import BaseRegistry

scripts/infinigram_count.py

@@ -50,7 +50,7 @@ def query_infinigram(ngram, index="v4_rpj_llama_s4", retries=3):
                 result = response.json()
                 if "count" in result:
                     return result["count"]
-        except Exception as e:  # type: ignore
+        except Exception:  # type: ignore
             time.sleep(1)
     return 0