allenai/olmocr

29.4

Adjusted Score

29.4

Raw Score

100%

Time Factor

2026-03-25

Last Push

19.1K

Stars

Python

Language

77.4K

Lines of Code

345

Files

1.5K

Pattern Hits

2026-07-14

Scan Date

0.19

HC Hit Rate

What These Metrics Mean

Adjusted Score: Primary synthetic code indicator. Raw score normalised per 1,000 lines of code and multiplied by the temporal discount factor. This is the definitive comparative metric — use it to rank repositories by AI authorship density.
Raw Score: The unmodified sum of all severity-weighted, context-multiplied pattern match scores before temporal discounting. Reflects the absolute signal strength independent of when the repository was last active.
Time Factor: The temporal discount multiplier (0–100%) applied to the raw score. Repositories last updated before ChatGPT's launch (Nov 2022) receive a 5% factor. Full signal is only assigned to repositories active in the post-adoption era (Jan 2024+).
Pattern Hits: Total count of individual pattern matches across all files and categories. A high hit count with a low score may indicate a very large codebase with isolated AI snippets; a low count with a high score indicates dense, concentrated AI signatures.
HC Hit Rate: High+Critical pattern hits per file, averaged across the repository. This orthogonal signal catches repositories where a few files are densely packed with high-severity AI tells — a strong indicator even when the normalised score appears moderate due to codebase size.
Lines of Code / Files: Total lines and files analysed. The scanner examines 94 file extensions. These denominators are used to normalise the score, enabling fair comparison between repositories of vastly different sizes.

Score History

This chart maps the temporal evolution of the adjusted synthetic code score across successive scan runs. An upward trajectory indicates ongoing incorporation of AI-generated code or expanding LLM-assisted scaffolding; a stable or declining trajectory may reflect active human refactoring, code removal, or the adoption of stricter authorship policies. The dashed secondary line (right axis) independently tracks total raw pattern hit count, which can diverge from the normalised score when codebase size changes significantly between scans.

Severity Breakdown

Classifies detected patterns by their diagnostic confidence and structural impact. CRITICAL patterns (coefficient 10) represent definitive synthetic signatures — hallucinated imports, explicit LLM attribution metadata — virtually never produced by human authors. HIGH (5) indicates strong structural tells such as cross-file repetition or cross-linguistic idioms. MEDIUM (2) covers recognisable conversational padding and AI-specific vocabulary. LOW (1) captures subtle indicators like tautological comments and generic boilerplate that require density to carry independent signal.

CRITICAL 0HIGH 65MEDIUM 326LOW 1079

Directory Score Breakdown

This horizontal bar chart decomposes the repository's raw synthetic code score by top-level directory, allowing you to pinpoint precisely which modules or components carry the highest AI authorship density. Directories with disproportionately high scores relative to their size warrant targeted manual review: concentrated AI signatures often trace back to mass-generated configuration layers, auto-ported test suites, LLM-scaffolded boilerplate classes, or entire subsystems authored under heavy copilot assistance. Use this view to prioritise your human code-review effort.

Pattern Findings

The scanner identified 1470 distinct pattern matches across 21 syntactic categories. Each entry below represents a discrete location in the source code where the engine recorded a statistically significant AI authorship indicator. Expand any category row to inspect the individual file paths, line numbers, code snippets, and the lexical context (CODE, COMMENT, or STRING) in which each match was detected.

Reading the findings table: The Severity column indicates the diagnostic confidence level (CRITICAL / HIGH / MEDIUM / LOW). The Context column identifies whether the match occurred inside executable code, an inline comment, or a string literal — comment-context matches receive a ×1.5 weight because LLMs systematically over-annotate. The ⚡ bolt icon marks clustered matches: three or more patterns within a 10-line window, each receiving an additional ×1.5 density multiplier as dense clusters constitute far stronger evidence of synthetic authorship than isolated hits.

Excessive Try-Catch Wrapping466 hits · 677 pts

Severity	File	Line	Snippet	Context
LOW	tests/test_tests.py	1578	except Exception as e:	CODE
LOW	tests/test_tests.py	1600	except Exception as e:	CODE
LOW	tests/test_tests.py	1612	except Exception as e:	CODE
LOW	tests/test_tests.py	1625	except Exception as e:	CODE
LOW	tests/test_tests.py	1645	except Exception as e:	CODE
LOW	tests/test_tests.py	2214	except Exception as e:	CODE
LOW⚡	olmocr/image_utils.py	49	except Exception as e:	CODE
MEDIUM⚡	olmocr/image_utils.py	50	print(f"Error: {e}")	CODE
LOW⚡	olmocr/image_utils.py	59	except Exception as e:	CODE
MEDIUM⚡	olmocr/image_utils.py	60	print(f"Error: {e}")	CODE
MEDIUM	olmocr/image_utils.py	44	def is_png(file_path):	CODE
MEDIUM	olmocr/image_utils.py	54	def is_jpeg(file_path):	CODE
LOW	olmocr/s3_utils.py	141	except Exception as e:	CODE
LOW	olmocr/s3_utils.py	243	except Exception as e:	CODE
LOW	olmocr/s3_utils.py	281	except Exception as e:	CODE
LOW	olmocr/s3_utils.py	323	except Exception as e:	CODE
LOW	olmocr/s3_utils.py	349	except Exception as e:	CODE
MEDIUM	olmocr/s3_utils.py	277	def download_blob(blob, local_file_path):	CODE
MEDIUM	olmocr/pipeline.py	845	def _kill_proc():	CODE
MEDIUM	olmocr/pipeline.py	1115	def process_output_file(s3_path):	CODE
LOW	olmocr/pipeline.py	228	except Exception as e:	CODE
LOW	olmocr/pipeline.py	466	except Exception as e:	CODE
LOW	olmocr/pipeline.py	570	except Exception as e:	CODE
LOW	olmocr/pipeline.py	803	except Exception as e:	CODE
LOW	olmocr/pipeline.py	890	except Exception as ex:	CODE
LOW	olmocr/pipeline.py	948	except Exception:	CODE
LOW	olmocr/pipeline.py	974	except Exception:	CODE
LOW	olmocr/pipeline.py	1149	except Exception as e:	CODE
LOW	olmocr/pipeline.py	1395	except Exception as e:	CODE
LOW	olmocr/work_queue.py	213	except Exception as e:	CODE
LOW	olmocr/bench/benchmark.py	111	except Exception as e:	CODE
LOW	olmocr/bench/benchmark.py	121	except Exception as e:	CODE
MEDIUM	olmocr/bench/benchmark.py	203	print("Error: /pdfs folder must exist in your data directory.", file=sys.stderr)	CODE
MEDIUM	olmocr/bench/benchmark.py	209	print(f"Error: No PDF files found in {pdf_folder}", file=sys.stderr)	CODE
MEDIUM	olmocr/bench/benchmark.py	220	print(f"Error: No .jsonl files found in {input_folder}.", file=sys.stderr)	CODE
MEDIUM	olmocr/bench/benchmark.py	275	print("Error: No candidate pipeline folders found (subdirectories besides 'pdfs').", file=sys.stderr)	CODE
LOW	olmocr/bench/convert.py	84	except Exception as ex:	CODE
LOW	olmocr/bench/convert.py	172	except Exception as e:	CODE
LOW	olmocr/bench/convert.py	215	except Exception as e:	CODE
LOW	olmocr/bench/convert.py	226	except Exception as e:	CODE
MEDIUM⚡	olmocr/bench/review_app.py	267	print(f"Error: Dataset not found: {args.dataset_file}")	CODE
MEDIUM⚡	olmocr/bench/review_app.py	276	print(f"Error: PDF directory not found: {pdf_dir}")	CODE
LOW⚡	olmocr/bench/review_app.py	282	except Exception as e:	CODE
MEDIUM⚡	olmocr/bench/review_app.py	283	print(f"Error loading dataset: {str(e)}")	CODE
MEDIUM⚡	olmocr/bench/review_app_latex.py	689	print(f"Error: Dataset not found: {args.dataset_file}")	CODE
MEDIUM⚡	olmocr/bench/review_app_latex.py	697	print(f"Error: PDF directory not found: {pdf_dir}")	CODE
LOW⚡	olmocr/bench/review_app_latex.py	702	except Exception as e:	CODE
MEDIUM⚡	olmocr/bench/review_app_latex.py	703	print(f"Error loading dataset: {str(e)}")	CODE
MEDIUM⚡	olmocr/bench/tests.py	834	print(f"Error parsing JSON on line {line_number}: {e}")	CODE
MEDIUM⚡	olmocr/bench/tests.py	837	print(f"Error on line {line_number}: {e}")	CODE
LOW⚡	olmocr/bench/tests.py	839	except Exception as e:	CODE
LOW	olmocr/bench/report.py	267	except Exception as e:	CODE
LOW	olmocr/bench/report.py	279	except Exception as e:	CODE
LOW	olmocr/bench/katex/render.py	179	except Exception:	CODE
LOW	olmocr/bench/katex/render.py	183	except Exception:	CODE
LOW	olmocr/bench/katex/render.py	224	except Exception:	CODE
MEDIUM	olmocr/bench/katex/render.py	304	print(f"Error rendering equation: '{equation}'")	STRING
LOW	olmocr/bench/katex/render.py	431	except Exception as e:	STRING
MEDIUM	olmocr/bench/katex/render.py	432	print("Error parsing MathML with BeautifulSoup:", e)	STRING
LOW	olmocr/bench/scripts/screen_pdfs.py	93	except Exception as e:	CODE
406 more matches not shown…

Self-Referential Comments119 hits · 346 pts

Severity	File	Line	Snippet	Context
MEDIUM	tests/test_mine_html_templates_mine_tests.py	931	# Create a temporary file to store the rendered PDF	STRING
MEDIUM⚡	tests/test_tests.py	570	# Create a table with irregular pipe placement at the end	COMMENT
MEDIUM⚡	tests/test_tests.py	603	# Create a very long text before the table	COMMENT
MEDIUM	tests/test_grpo.py	44	# Create a minimal valid PDF	COMMENT
MEDIUM	tests/test_grpo.py	189	# Create an empty JSONL file	COMMENT
MEDIUM	tests/test_grpo.py	208	# Create a JSONL with some malformed entries	COMMENT
MEDIUM	tests/test_grpo.py	275	# Create a sample JSONL test file with different test types	COMMENT
MEDIUM	tests/test_pipeline.py	172	# Create a more complex test image	COMMENT
MEDIUM	olmocr/metrics.py	159	# Create the table header	COMMENT
MEDIUM	olmocr/metrics.py	163	# Create the table rows	COMMENT
MEDIUM	olmocr/pipeline.py	691	# Create the output markdown path	COMMENT
MEDIUM	olmocr/pipeline.py	750	# Define the output S3 path using the work_hash	COMMENT
MEDIUM	olmocr/pipeline.py	775	# Create the directory structure if it doesn't exist	COMMENT
MEDIUM	olmocr/pipeline.py	1070	# Create the experiment spec	COMMENT
MEDIUM	olmocr/bench/benchmark.py	79	# Define an inner function to evaluate a single test	COMMENT
MEDIUM	olmocr/bench/review_app_latex.py	248	# Create the review_latex.html template with MathJax support	COMMENT
MEDIUM	olmocr/bench/review_app_latex.py	611	# Create the all_done_latex.html template	STRING
MEDIUM	olmocr/bench/katex/render.py	243	# Create a new page.	COMMENT
MEDIUM	olmocr/bench/scripts/split_table_tests.py	91	# Create a new test dict for each relationship	COMMENT
MEDIUM⚡	olmocr/bench/miners/mine_reading_order.py	93	# Create a new PDF with just the selected page	COMMENT
MEDIUM	olmocr/bench/miners/mine_reading_order.py	34	# Create a thread-safe lock for writing to output files	COMMENT
MEDIUM	olmocr/bench/miners/mine_reading_order.py	343	# Create a thread-specific temp directory to avoid conflicts	COMMENT
MEDIUM	olmocr/bench/miners/mine_reading_order.py	442	# Create a summary file	COMMENT
MEDIUM	olmocr/bench/miners/mine_reading_order.py	448	# Create a ThreadPoolExecutor	COMMENT
MEDIUM⚡	olmocr/bench/miners/mine_tables_gpt.py	89	# Create a new PDF with just the selected page	COMMENT
MEDIUM⚡	olmocr/bench/miners/mine_tables_gemini.py	97	# Create a new PDF with just the selected page	COMMENT
MEDIUM	olmocr/bench/miners/mine_tables_gemini.py	37	# Create a thread-safe lock for writing to the output file	COMMENT
MEDIUM	olmocr/bench/miners/mine_tables_gemini.py	238	# Create an image part from the rendered pdf image	COMMENT
MEDIUM	olmocr/bench/miners/mine_tables_gemini.py	308	# Create a thread-specific temp directory to avoid conflicts	COMMENT
MEDIUM	olmocr/bench/miners/mine_tables_gemini.py	414	# Create a ThreadPoolExecutor	COMMENT
MEDIUM⚡	olmocr/bench/miners/mine_long_tiny_text.py	57	# Create a new PDF with just the selected page	COMMENT
MEDIUM	olmocr/bench/miners/mine_long_tiny_text.py	214	# Create a placeholder test to ensure the PDF is represented	COMMENT
MEDIUM	olmocr/bench/miners/cleanup_data.py	56	# Create a new single-page PDF	COMMENT
MEDIUM⚡	olmocr/bench/miners/cleanup_data.py	93	# Create a dataset.jsonl with only the tests for pages we're keeping	COMMENT
MEDIUM⚡	olmocr/bench/miners/mine_headers_footers.py	89	# Create a new PDF with just the selected page	COMMENT
MEDIUM⚡	olmocr/bench/miners/mine_old_scans_math.py	57	# Create a new PDF with just the selected page	COMMENT
MEDIUM	olmocr/bench/miners/pick_mediod.py	141	# Create the output path, preserving the directory structure relative to input_dir	COMMENT
MEDIUM⚡	olmocr/bench/runners/run_mistral.py	32	# Create a new PDF with just the requested page	COMMENT
MEDIUM⚡	olmocr/bench/runners/run_mineru.py	26	# Create a new PDF with just the requested page	COMMENT
MEDIUM	olmocr/bench/runners/run_marker.py	17	# Create a configuration dictionary with the necessary settings	COMMENT
MEDIUM⚡	olmocr/bench/runners/run_marker.py	47	# Create a new PDF with just the requested page	COMMENT
MEDIUM⚡	olmocr/synth/mine_html_templates.py	530	# Create a temporary PDF file	COMMENT
MEDIUM⚡	olmocr/synth/mine_html_templates.py	670	# Create a new PDF with just the selected page	COMMENT
MEDIUM	olmocr/synth/mine_html_templates.py	224	# Create a temporary soup with just this element to get its HTML	COMMENT
MEDIUM	olmocr/synth/mine_html_templates.py	311	# Create a new soup with just the body content	COMMENT
MEDIUM	olmocr/synth/mine_html_templates.py	328	# Create an img tag with placeholder src and appropriate alt text	COMMENT
MEDIUM	olmocr/synth/mine_html_templates.py	1825	# Create a unique folder for each PDF in the temp directory; include PID and PDF path hash to avoid cross-run collis	COMMENT
MEDIUM	olmocr/synth/mine_html_templates.py	972	# Create a TableTest with relevant relationships	STRING
MEDIUM	olmocr/synth/mine_html_templates.py	1016	# Create the actual test object	STRING
MEDIUM	olmocr/synth/mine_html_templates.py	1441	# Create a format test	STRING
MEDIUM	olmocr/synth/mine_html_templates.py	1694	# Create a signature for the test by using all fields except 'id'	STRING
MEDIUM	olmocr/synth/mine_html_templates.py	1717	# Create a baseline test object to check if disallowed characters are present	STRING
MEDIUM	olmocr/synth/rotate_html_templates.py	54	# Create the same symlink in destination	COMMENT
MEDIUM	olmocr/synth/rotate_html_templates.py	283	# Create a temporary file for the rotated PDF	COMMENT
MEDIUM	olmocr/train/config.py	330	# Create main config	COMMENT
MEDIUM	olmocr/train/compress_checkpoint.py	254	# Define a oneshot data collator for multimodal inputs.	COMMENT
MEDIUM	olmocr/train/grpo_train.py	1429	# Create a wrapper function that calls olmocr_bench_reward with macro_average=True	COMMENT
MEDIUM	olmocr/train/grpo_train.py	1474	# Create a wrapper function with proper __name__ attribute	COMMENT
MEDIUM	olmocr/train/dataloader.py	926	# Create a new PageResponse with the updated text (since it's frozen)	COMMENT
MEDIUM	olmocr/train/dataloader.py	1075	# Create a new PageResponse with the updated text (since it's frozen)	COMMENT
59 more matches not shown…

Hyper-Verbose Identifiers312 hits · 285 pts

Severity	File	Line	Snippet	Context
LOW⚡	tests/test_olmocrmix.py	10	def test_repackage_and_prepare_olmocrmix():	CODE
LOW	tests/test_s3_work_queue.py	29	def test_compute_workgroup_hash(self):	CODE
LOW	tests/test_s3_work_queue.py	73	async def test_populate_queue_new_items(self):	CODE
LOW	tests/test_s3_work_queue.py	94	async def test_populate_queue_existing_items(self):	CODE
LOW	tests/test_mine_html_templates_markdown.py	17	def test_title_tag_excluded_from_markdown(self):	CODE
LOW	tests/test_mine_html_templates_markdown.py	44	def test_image_with_data_description(self):	STRING
LOW	tests/test_mine_html_templates_markdown.py	65	def test_image_without_data_description(self):	STRING
LOW	tests/test_mine_html_templates_markdown.py	80	def test_headers_footers_excluded(self):	STRING
LOW	tests/test_mine_html_templates_markdown.py	109	def test_no_body_tag_fallback(self):	STRING
LOW	tests/test_mine_html_templates_markdown.py	124	def test_removes_triple_dashes_from_content(self):	STRING
LOW	tests/test_mine_html_templates_markdown.py	222	def test_mixed_super_and_subscripts(self):	STRING
LOW	tests/test_mine_html_templates_markdown.py	314	def test_frontmatter_preserved(self):	STRING
LOW	tests/test_mine_html_templates_markdown.py	351	def test_empty_super_subscripts(self):	STRING
LOW	tests/test_mine_html_templates_markdown.py	369	def test_complex_math_expression(self):	STRING
LOW	tests/test_mine_html_templates_markdown.py	391	def test_non_numerical_footnote(self):	STRING
LOW	tests/test_mine_html_templates_markdown.py	406	def test_line_numbers_markdown(self):	STRING
LOW	tests/test_dataloader.py	25	def test_markdown_table_detection(self):	CODE
LOW	tests/test_dataloader.py	99	def test_malformed_html_table_unclosed_tags(self):	STRING
LOW	tests/test_dataloader.py	128	def test_malformed_html_table_missing_closing(self):	STRING
LOW	tests/test_dataloader.py	197	def test_complex_markdown_table_variations(self):	CODE
LOW	tests/test_dataloader.py	247	def test_br_tags_outside_tables(self):	STRING
LOW	tests/test_dataloader.py	277	def test_simple_bold_conversion(self):	CODE
LOW	tests/test_dataloader.py	293	def test_simple_italic_conversion(self):	CODE
LOW	tests/test_dataloader.py	309	def test_bold_and_italic_conversion(self):	CODE
LOW	tests/test_dataloader.py	325	def test_latex_in_math_equation_preserved(self):	CODE
LOW	tests/test_dataloader.py	373	def test_multiple_occurrences(self):	CODE
LOW	tests/test_dataloader.py	389	def test_latex_in_parenthesis_delimiter(self):	CODE
LOW	tests/test_dataloader.py	405	def test_latex_in_bracket_delimiter(self):	CODE
LOW	tests/test_dataloader.py	460	def test_complex_latex_with_parenthesis_delimiters(self):	CODE
LOW	tests/test_dataloader.py	536	def test_rotation_correction_needed(self):	CODE
LOW	tests/test_dataloader.py	567	def test_dollar_to_parentheses(self):	CODE
LOW	tests/test_dataloader.py	584	def test_double_dollar_to_brackets(self):	CODE
LOW	tests/test_dataloader.py	601	def test_mixed_latex_delimiters(self):	CODE
LOW	tests/test_dataloader.py	667	def test_parse_yaml_front_matter(self, mock_read_text):	CODE
LOW	tests/test_dataloader.py	715	def test_preserve_existing_markdown_content(self, mock_read_text):	STRING
LOW	tests/test_mine_html_templates_mine_tests.py	1044	def test_complex_markdown_formatting(self):	CODE
LOW	tests/test_mine_html_templates_mine_tests.py	1104	def test_complex_formatting_two(self):	STRING
LOW	tests/test_mine_html_templates_footnotes.py	46	def test_single_marker_generates_marker_only(self):	CODE
LOW	tests/test_mine_html_templates_footnotes.py	71	def test_marker_with_definition_includes_text(self):	STRING
LOW	tests/test_mine_html_templates_footnotes.py	99	def test_multiple_markers_mixed_occurrences(self):	STRING
LOW	tests/test_mine_html_templates_footnotes.py	211	def test_footnotes_in_footer_ptags(self):	STRING
LOW	tests/test_mine_html_templates_footnotes.py	314	def test_footnotes_footer_again(self):	CODE
LOW⚡	tests/test_tests.py	136	def test_present_text_exact_match(self):	CODE
LOW⚡	tests/test_tests.py	142	def test_present_text_not_found(self):	CODE
LOW⚡	tests/test_tests.py	149	def test_present_text_with_max_diffs(self):	CODE
LOW⚡	tests/test_tests.py	182	def test_absent_text_not_found(self):	CODE
LOW⚡	tests/test_tests.py	188	def test_case_insensitive_present(self):	CODE
LOW⚡	tests/test_tests.py	194	def test_case_insensitive_absent(self):	CODE
LOW⚡	tests/test_tests.py	526	def test_with_stripped_content(self):	CODE
LOW⚡	tests/test_tests.py	534	def test_table_at_end_of_file(self):	CODE
LOW⚡	tests/test_tests.py	542	def test_table_at_end_with_no_trailing_newline(self):	CODE
LOW⚡	tests/test_tests.py	550	def test_table_at_end_with_extra_spaces(self):	CODE
LOW⚡	tests/test_tests.py	559	def test_table_at_end_with_mixed_whitespace(self):	CODE
LOW⚡	tests/test_tests.py	567	def test_malformed_table_at_end(self):	CODE
LOW⚡	tests/test_tests.py	580	def test_incomplete_table_at_end(self):	CODE
LOW⚡	tests/test_tests.py	592	def test_table_with_excessive_blank_lines_at_end(self):	CODE
LOW⚡	tests/test_tests.py	600	def test_table_at_end_after_long_text(self):	CODE
LOW⚡	tests/test_tests.py	609	def test_valid_table_at_eof_without_newline(self):	CODE
LOW⚡	tests/test_tests.py	2077	def test_empty_appears_before_validation(self):	CODE
LOW⚡	tests/test_tests.py	2082	def test_empty_appears_after_validation(self):	CODE
252 more matches not shown…

Redundant / Tautological Comments144 hits · 224 pts

Severity	File	Line	Snippet	Context
LOW	tests/test_mine_html_templates_markdown.py	288	# Check if tags are preserved in table cells	STRING
LOW	tests/test_mine_html_templates_mine_tests.py	938	# Check if rendering was successful	STRING
LOW	olmocr/s3_utils.py	76	# Check if it's actually a folder with contents	COMMENT
LOW	olmocr/pipeline.py	1347	# Check if this is a tar.gz file (local)	COMMENT
LOW	olmocr/pipeline.py	1410	# Add tarballs to the queue - each tarball is one work item	COMMENT
LOW	olmocr/bench/tests.py	319	# Check if the text appears in any of the formatted texts using fuzzy matching	COMMENT
LOW	olmocr/bench/tests.py	715	# Check if appears_before_marker is at the end of this text (using fuzzy matching)	COMMENT
LOW	olmocr/bench/tests.py	741	# Check if appears_after_marker is at the beginning of this text (using fuzzy matching)	COMMENT
LOW	olmocr/bench/table_parsing.py	284	# Check if this line has pipe characters (a table row indicator)	COMMENT
LOW	olmocr/bench/table_parsing.py	359	# Check if this looks like a separator row (contains mostly dashes)	COMMENT
LOW	olmocr/bench/scripts/convert_all.sh	121	# Check if environment exists	COMMENT
LOW	olmocr/bench/scripts/convert_all.sh	149	# Check if the server process is running	COMMENT
LOW	olmocr/bench/scripts/rotate_pdfs_random.sh	6	# Check if directory exists	COMMENT
LOW	olmocr/bench/scripts/rotate_pdfs_random.sh	12	# Check if qpdf is installed (preferred for PDF rotation)	COMMENT
LOW	olmocr/bench/scripts/rotate_pdfs_random.sh	28	# Check if any PDF files exist	COMMENT
LOW⚡	olmocr/bench/miners/mine_reading_order.py	88	# Check if page number is valid	COMMENT
LOW	olmocr/bench/miners/mine_reading_order.py	470	# Check if we've reached the maximum number of documents	COMMENT
LOW⚡	olmocr/bench/miners/mine_tables_gpt.py	84	# Check if page number is valid	COMMENT
LOW⚡	olmocr/bench/miners/mine_blank_pages_gpt.py	119	# Check if natural_text is null	COMMENT
LOW	olmocr/bench/miners/mine_blank_pages_gpt.py	172	# Check if the page has null natural_text	COMMENT
LOW⚡	olmocr/bench/miners/mine_tables_gemini.py	92	# Check if page number is valid	COMMENT
LOW	olmocr/bench/miners/mine_tables_gemini.py	432	# Check if we've reached the maximum number of tests	COMMENT
LOW	olmocr/bench/miners/mine_multilingual_gpt.py	169	# Check if the page contains a table	COMMENT
LOW⚡	olmocr/bench/miners/mine_long_tiny_text.py	52	# Check if page number is valid	COMMENT
LOW⚡	olmocr/bench/miners/mine_headers_footers.py	84	# Check if page number is valid	COMMENT
LOW	olmocr/bench/miners/delete_rejected.py	38	# Check if the test is marked as rejected	COMMENT
LOW	olmocr/bench/miners/delete_rejected.py	95	# Check if all tests for this PDF are in the rejected list	COMMENT
LOW⚡	olmocr/bench/miners/mine_footnotes_gpt.py	125	# Check if there are any footnotes in the list	COMMENT
LOW	olmocr/bench/miners/mine_footnotes_gpt.py	179	# Check if the page contains footnotes	COMMENT
LOW⚡	olmocr/bench/miners/mine_old_scans_math.py	52	# Check if page number is valid	COMMENT
LOW	olmocr/bench/miners/mine_tables_gpt_simple.py	200	# Check if the page contains a table	COMMENT
LOW⚡	olmocr/bench/runners/run_mistral.py	28	# Check if the requested page exists	COMMENT
LOW⚡	olmocr/bench/runners/run_mistral.py	39	temp_file.close() # Close the file but keep the name	CODE
LOW⚡	olmocr/bench/runners/run_mineru.py	22	# Check if the requested page exists	COMMENT
LOW⚡	olmocr/bench/runners/run_mineru.py	33	temp_file.close() # Close the file but keep the name	CODE
LOW⚡	olmocr/bench/runners/run_marker.py	43	# Check if the requested page exists	COMMENT
LOW⚡	olmocr/bench/runners/run_marker.py	54	temp_file.close() # Close the file but keep the name	CODE
LOW⚡	olmocr/synth/mine_html_templates.py	665	# Check if page number is valid	COMMENT
LOW	olmocr/synth/mine_html_templates.py	157	# Check if it's a local path	COMMENT
LOW	olmocr/synth/mine_html_templates.py	461	# Check if response was complete	COMMENT
LOW	olmocr/synth/mine_html_templates.py	508	# Check if response was complete	COMMENT
LOW	olmocr/synth/mine_html_templates.py	603	# Check if refinement response was complete	COMMENT
LOW	olmocr/synth/mine_html_templates.py	1866	# Check if this PDF has already been processed (skip logic for resume)	COMMENT
LOW	olmocr/synth/mine_html_templates.py	885	# Check if this element has no children from our target tags	STRING
LOW	olmocr/synth/mine_html_templates.py	1194	# Check if it's a number	STRING
LOW	olmocr/synth/mine_html_templates.py	1198	# Check if it's a word (has at least one alphabetic character)	STRING
LOW	olmocr/synth/rotate_html_templates.py	50	# Check if it's a symlink	COMMENT
LOW	olmocr/synth/rotate_html_templates.py	117	# Check if file starts with FrontMatter	COMMENT
LOW	olmocr/train/train.py	119	# Check if we have any valid samples	COMMENT
LOW	olmocr/train/train.py	654	# Check if we should do a gradient update	COMMENT
LOW	olmocr/train/train.py	712	# Check if we've reached our training limit	COMMENT
LOW⚡	olmocr/train/dataloader.py	439	# Check if line looks like a table row	COMMENT
LOW⚡	olmocr/train/dataloader.py	441	# Check if next line is a separator (for header rows)	COMMENT
LOW⚡	olmocr/train/dataloader.py	446	# Check if previous line is a separator (for data rows)	COMMENT
LOW⚡	olmocr/train/dataloader.py	662	# Check if there was an error	COMMENT
LOW⚡	olmocr/train/dataloader.py	683	# Check if there are any tables in the text	COMMENT
LOW⚡	olmocr/train/dataloader.py	698	# Check if cell content contains <br> tags (any variation)	COMMENT
LOW⚡	olmocr/train/dataloader.py	713	# Check if there are any <table> tags at all	COMMENT
LOW	olmocr/train/dataloader.py	346	# Check if image exists	COMMENT
LOW	olmocr/train/dataloader.py	350	# Check if page_data exists (we need to update it)	COMMENT
84 more matches not shown…

Cross-Language Confusion34 hits · 194 pts

Severity	File	Line	Snippet	Context
HIGH	tests/test_mine_html_templates_markdown.py	444	<div class="line-text">unknown and undefined, so the present study aimed to investigate characteristics of the DHLA-</di	CODE
HIGH	olmocr/pipeline.py	181	r"---\nprimary_language: (?:[a-z]{2}\|null)\nis_rotation_valid: (?:True\|False\|true\|false)\nrotation_corre	CODE
HIGH	olmocr/bench/review_app_latex.py	442	let pdfDoc = null;	CODE
HIGH	olmocr/bench/review_app_latex.py	444	let canvas = null;	CODE
HIGH	olmocr/bench/katex/render.py	290	return null;	CODE
HIGH	olmocr/bench/katex/render.py	329	list.push({	CODE
HIGH	olmocr/bench/miners/mine_blank_pages_gpt.py	79	Optional[bool]: True if natural_text is null, False otherwise, None if detection fails	STRING
HIGH	olmocr/bench/miners/mine_blank_pages_gpt.py	79	Optional[bool]: True if natural_text is null, False otherwise, None if detection fails	STRING
HIGH⚡	olmocr/bench/miners/mine_blank_pages_gpt.py	119	# Check if natural_text is null	COMMENT
HIGH	olmocr/synth/cutoff_detection.py	61	if (child.nodeType === Node.TEXT_NODE && child.textContent.trim()) {	CODE
HIGH	olmocr/synth/cutoff_detection.py	68	leaves.push(el);	CODE
HIGH	olmocr/synth/cutoff_detection.py	83	if (rect.width <= 0 \|\| rect.height <= 0) continue;	CODE
HIGH	olmocr/synth/cutoff_detection.py	109	let clippingAncestorTag = null;	CODE
HIGH	olmocr/synth/cutoff_detection.py	123	if (clipsX \|\| clipsY) {	CODE
HIGH	olmocr/synth/cutoff_detection.py	142	(visLeft !== prevLeft \|\| visTop !== prevTop \|\|	CODE
HIGH	olmocr/synth/cutoff_detection.py	143	visRight !== prevRight \|\| visBottom !== prevBottom)) {	CODE
HIGH	olmocr/synth/cutoff_detection.py	166	results.push({	CODE
HIGH	olmocr/synth/cutoff_detection.py	200	if (child.nodeType === Node.TEXT_NODE && child.textContent.trim()) {	CODE
HIGH	olmocr/synth/cutoff_detection.py	205	if (hasDirectText) leaves.push(el);	CODE
HIGH⚡	olmocr/synth/cutoff_detection.py	221	if (!bgColor \|\| bgColor === 'rgba(0, 0, 0, 0)' \|\| bgColor === 'transparent') return 0;	CODE
HIGH⚡	olmocr/synth/cutoff_detection.py	231	if (rect.width <= 0 \|\| rect.height <= 0) continue;	CODE
HIGH⚡	olmocr/synth/cutoff_detection.py	238	if (!directText \|\| directText.length < 3) continue;	CODE
HIGH⚡	olmocr/synth/cutoff_detection.py	246	let blockerTag = null;	CODE
HIGH⚡	olmocr/synth/cutoff_detection.py	252	if (px < 0 \|\| py < 0) continue;	CODE
HIGH⚡	olmocr/synth/cutoff_detection.py	256	if (topEl === el \|\| isRelated(topEl, el)) continue;	CODE
HIGH	olmocr/synth/cutoff_detection.py	269	results.push({	CODE
HIGH⚡	scripts/scan_dolmadocs.py	1160	const primaryOption = activePrimaryButton ? activePrimaryButton.getAttribute('data-value') : null;	CODE
HIGH⚡	scripts/scan_dolmadocs.py	1165	publicPiiOptions.push(checkbox.getAttribute('data-value'));	CODE
HIGH⚡	scripts/scan_dolmadocs.py	1171	privatePiiOptions.push(checkbox.getAttribute('data-value'));	CODE
HIGH	scripts/scan_dolmadocs.py	1058	status.onclick = null;	CODE
HIGH	scripts/scan_dolmadocs.py	1343	if (data.publicPiiOptions && data.publicPiiOptions.length > 0) {	CODE
HIGH	scripts/scan_dolmadocs.py	1357	if (data.privatePiiOptions && data.privatePiiOptions.length > 0) {	CODE
HIGH	scripts/scan_dolmadocs.py	1528	if (event.target && event.target.id === 'start-button') {	CODE
HIGH	scripts/scan_dolmadocs.py	1539	if (event.target && event.target.id === 'view-instructions-button') {	CODE

Deep Nesting178 hits · 153 pts

Severity	File	Line	Context
LOW	tests/test_s3_work_queue.py	224	CODE
LOW	tests/test_integration.py	11	CODE
LOW	olmocr/s3_utils.py	36	CODE
LOW	olmocr/s3_utils.py	199	CODE
LOW	olmocr/s3_utils.py	250	CODE
LOW	olmocr/s3_utils.py	378	CODE
LOW	olmocr/pipeline.py	106	CODE
LOW	olmocr/pipeline.py	281	CODE
LOW	olmocr/pipeline.py	382	CODE
LOW	olmocr/pipeline.py	485	CODE
LOW	olmocr/pipeline.py	698	CODE
LOW	olmocr/pipeline.py	956	CODE
LOW	olmocr/pipeline.py	1199	CODE
LOW	olmocr/bench/benchmark.py	162	CODE
LOW	olmocr/bench/convert.py	19	CODE
LOW	olmocr/bench/convert.py	103	CODE
LOW	olmocr/bench/review_app.py	38	CODE
LOW	olmocr/bench/review_app.py	219	CODE
LOW	olmocr/bench/review_app_latex.py	36	CODE
LOW	olmocr/bench/review_app_latex.py	217	CODE
LOW	olmocr/bench/utils.py	6	CODE
LOW	olmocr/bench/tests.py	766	CODE
LOW	olmocr/bench/tests.py	253	CODE
LOW	olmocr/bench/tests.py	376	CODE
LOW	olmocr/bench/tests.py	655	CODE
LOW	olmocr/bench/table_parsing.py	83	CODE
LOW	olmocr/bench/table_parsing.py	263	CODE
LOW	olmocr/bench/table_parsing.py	390	CODE
LOW	olmocr/bench/report.py	12	CODE
LOW	olmocr/bench/report.py	42	CODE
LOW	olmocr/bench/katex/render.py	414	CODE
LOW	olmocr/bench/katex/render.py	475	CODE
LOW	olmocr/bench/katex/render.py	517	CODE
LOW	olmocr/bench/scripts/workspace_to_bench.py	25	CODE
LOW	olmocr/bench/scripts/difference_viewer.py	15	CODE
LOW	olmocr/bench/scripts/difference_viewer.py	64	CODE
LOW	olmocr/bench/scripts/url_matcher.py	10	CODE
LOW	olmocr/bench/scripts/url_matcher.py	29	CODE
LOW	olmocr/bench/scripts/rotate_pdfs.py	10	CODE
LOW	olmocr/bench/miners/mine_reading_order.py	330	CODE
LOW	olmocr/bench/miners/mine_reading_order.py	426	CODE
LOW	olmocr/bench/miners/mine_reading_order.py	483	CODE
LOW	olmocr/bench/miners/mine_tables_gpt.py	189	CODE
LOW	olmocr/bench/miners/mine_diffs.py	90	CODE
LOW	olmocr/bench/miners/mine_multi_column.py	181	CODE
LOW	olmocr/bench/miners/mine_multi_column.py	330	CODE
LOW	olmocr/bench/miners/mine_blank_pages_gpt.py	204	CODE
LOW	olmocr/bench/miners/mine_tables_gemini.py	205	CODE
LOW	olmocr/bench/miners/mine_tables_gemini.py	398	CODE
LOW	olmocr/bench/miners/mine_tables_gemini.py	445	CODE
LOW	olmocr/bench/miners/mine_old_scan_pdf.py	9	CODE
LOW	olmocr/bench/miners/mine_multilingual_gpt.py	201	CODE
LOW	olmocr/bench/miners/mine_long_tiny_text.py	151	CODE
LOW	olmocr/bench/miners/cleanup_data.py	10	CODE
LOW	olmocr/bench/miners/cleanup_data.py	89	CODE
LOW	olmocr/bench/miners/delete_rejected.py	10	CODE
LOW	olmocr/bench/miners/delete_rejected.py	53	CODE
LOW	olmocr/bench/miners/mine_math.py	84	CODE
LOW	olmocr/bench/miners/check_old_scans_math.py	86	CODE
LOW	olmocr/bench/miners/mine_footnotes_gpt.py	211	CODE
118 more matches not shown…

Cross-File Repetition26 hits · 130 pts

Severity	File	Snippet	Context
HIGH	olmocr/bench/miners/mine_reading_order.py	download a pdf file from s3. args: s3_path: the s3 path (s3://bucket/path/to/file.pdf) local_path: the local path to sav	STRING
HIGH	olmocr/bench/miners/mine_tables_gpt.py	download a pdf file from s3. args: s3_path: the s3 path (s3://bucket/path/to/file.pdf) local_path: the local path to sav	STRING
HIGH	olmocr/bench/miners/mine_blank_pages_gpt.py	download a pdf file from s3. args: s3_path: the s3 path (s3://bucket/path/to/file.pdf) local_path: the local path to sav	STRING
HIGH	olmocr/bench/miners/mine_tables_gemini.py	download a pdf file from s3. args: s3_path: the s3 path (s3://bucket/path/to/file.pdf) local_path: the local path to sav	STRING
HIGH	olmocr/bench/miners/mine_multilingual_gpt.py	download a pdf file from s3. args: s3_path: the s3 path (s3://bucket/path/to/file.pdf) local_path: the local path to sav	STRING
HIGH	olmocr/bench/miners/mine_headers_footers.py	download a pdf file from s3. args: s3_path: the s3 path (s3://bucket/path/to/file.pdf) local_path: the local path to sav	STRING
HIGH	olmocr/bench/miners/mine_footnotes_gpt.py	download a pdf file from s3. args: s3_path: the s3 path (s3://bucket/path/to/file.pdf) local_path: the local path to sav	STRING
HIGH	olmocr/bench/miners/mine_length_gpt_simple.py	download a pdf file from s3. args: s3_path: the s3 path (s3://bucket/path/to/file.pdf) local_path: the local path to sav	STRING
HIGH	olmocr/bench/miners/mine_tables_gpt_simple.py	download a pdf file from s3. args: s3_path: the s3 path (s3://bucket/path/to/file.pdf) local_path: the local path to sav	STRING
HIGH	olmocr/bench/miners/mine_reading_order.py	extract a specific page from a pdf and save it as a new pdf. args: input_path: path to the input pdf output_path: path t	STRING
HIGH	olmocr/bench/miners/mine_tables_gpt.py	extract a specific page from a pdf and save it as a new pdf. args: input_path: path to the input pdf output_path: path t	STRING
HIGH	olmocr/bench/miners/mine_tables_gemini.py	extract a specific page from a pdf and save it as a new pdf. args: input_path: path to the input pdf output_path: path t	STRING
HIGH	olmocr/bench/miners/mine_long_tiny_text.py	extract a specific page from a pdf and save it as a new pdf. args: input_path: path to the input pdf output_path: path t	STRING
HIGH	olmocr/bench/miners/mine_headers_footers.py	extract a specific page from a pdf and save it as a new pdf. args: input_path: path to the input pdf output_path: path t	STRING
HIGH	olmocr/bench/miners/mine_old_scans_math.py	extract a specific page from a pdf and save it as a new pdf. args: input_path: path to the input pdf output_path: path t	STRING
HIGH	olmocr/bench/miners/mine_blank_pages_gpt.py	process a single pdf from s3. args: s3_path: s3 path to the pdf temp_dir: directory for temporary files output_dir: dire	STRING
HIGH	olmocr/bench/miners/mine_multilingual_gpt.py	process a single pdf from s3. args: s3_path: s3 path to the pdf temp_dir: directory for temporary files output_dir: dire	STRING
HIGH	olmocr/bench/miners/mine_tables_gpt_simple.py	process a single pdf from s3. args: s3_path: s3 path to the pdf temp_dir: directory for temporary files output_dir: dire	STRING
HIGH	olmocr/bench/miners/mine_footnotes_gpt.py	process a single pdf from s3. args: s3_path: s3 path to the pdf temp_dir: directory for temporary files output_dir: dire	STRING
HIGH	olmocr/bench/miners/mine_length_gpt_simple.py	process a single pdf from s3. args: s3_path: s3 path to the pdf temp_dir: directory for temporary files output_dir: dire	STRING
HIGH	olmocr/bench/runners/run_transformers.py	convert page of a pdf file to markdown by calling a request running against an openai compatible server. you can use thi	STRING
HIGH	olmocr/bench/runners/run_nanonetsocr_2.py	convert page of a pdf file to markdown by calling a request running against an openai compatible server. you can use thi	STRING
HIGH	olmocr/bench/runners/run_server.py	convert page of a pdf file to markdown by calling a request running against an openai compatible server. you can use thi	STRING
HIGH	scripts/pii/rich_tagging_pipeline.py	you are a document analyzer that identifies personally identifiable information (pii) in documents. your task is to anal	STRING
HIGH	scripts/pii/autoscan_dolmadocs.py	you are a document analyzer that identifies personally identifiable information (pii) in documents. your task is to anal	STRING
HIGH	scripts/pii/chatgpt_tag_dolmadocs_v1.py	you are a document analyzer that identifies personally identifiable information (pii) in documents. your task is to anal	STRING

Unused Imports60 hits · 59 pts

Severity	File	Line	Context
LOW	tests/test_mine_html_templates_markdown.py	1	CODE
LOW	tests/test_mine_html_templates_markdown.py	3	CODE
LOW	tests/test_mine_html_templates_markdown.py	3	CODE
LOW	tests/test_mine_html_templates_markdown.py	5	CODE
LOW	tests/test_mine_html_templates_markdown.py	7	CODE
LOW	tests/test_mine_html_templates_markdown.py	8	CODE
LOW	tests/test_mine_html_templates_markdown.py	8	CODE
LOW	tests/test_mine_html_templates_markdown.py	8	CODE
LOW	tests/test_dataloader.py	6	CODE
LOW	tests/test_mine_html_templates_mine_tests.py	3	CODE
LOW	tests/test_mine_html_templates_mine_tests.py	3	CODE
LOW	tests/test_mine_html_templates_mine_tests.py	5	CODE
LOW	tests/test_mine_html_templates_mine_tests.py	7	CODE
LOW	tests/test_mine_html_templates_mine_tests.py	8	CODE
LOW	tests/test_mine_html_templates_mine_tests.py	8	CODE
LOW	tests/test_filter.py	4	CODE
LOW	tests/test_cutoff_detection.py	6	CODE
LOW	tests/test_mine_html_templates_metadata.py	1	CODE
LOW	tests/test_mine_html_templates_metadata.py	3	CODE
LOW	tests/test_mine_html_templates_metadata.py	3	CODE
LOW	tests/test_mine_html_templates_metadata.py	5	CODE
LOW	tests/test_mine_html_templates_metadata.py	7	CODE
LOW	tests/test_mine_html_templates_metadata.py	8	CODE
LOW	tests/test_mine_html_templates_metadata.py	8	CODE
LOW	tests/test_mine_html_templates_metadata.py	8	CODE
LOW	tests/test_grpo.py	13	CODE
LOW	tests/test_anchor.py	16	CODE
LOW	tests/test_anchor.py	17	CODE
LOW	tests/test_anchor.py	69	CODE
LOW	tests/test_anchor.py	70	CODE
LOW	tests/test_mine_html_templates_formatting.py	3	CODE
LOW	tests/test_mine_html_templates_formatting.py	3	CODE
LOW	tests/test_mine_html_templates_formatting.py	5	CODE
LOW	tests/test_mine_html_templates_formatting.py	7	CODE
LOW	tests/test_mine_html_templates_formatting.py	8	CODE
LOW	tests/test_mine_html_templates_formatting.py	8	CODE
LOW	tests/test_table_parsing.py	3	CODE
LOW	tests/test_mine_html_templates_incomplete_content.py	5	CODE
LOW	tests/test_mine_html_templates_incomplete_content.py	5	CODE
LOW	tests/test_mine_html_templates_incomplete_content.py	7	CODE
LOW	tests/test_mine_html_templates_incomplete_content.py	9	CODE
LOW	tests/test_mine_html_templates_introduce_text_errors.py	4	CODE
LOW	tests/test_mine_html_templates_math_extraction.py	5	CODE
LOW	tests/test_mine_html_templates_math_extraction.py	7	CODE
LOW	tests/test_mine_html_templates_math_extraction.py	8	CODE
LOW	tests/test_mine_html_templates_math_extraction.py	8	CODE
LOW	tests/test_mine_html_templates_math_extraction.py	8	CODE
LOW	olmocr/__init__.py	1	CODE
LOW	olmocr/__init__.py	1	CODE
LOW	olmocr/bench/katex/__init__.py	1	CODE
LOW	olmocr/bench/katex/__init__.py	1	CODE
LOW	olmocr/prompts/__init__.py	1	CODE
LOW	olmocr/prompts/__init__.py	1	CODE
LOW	olmocr/prompts/__init__.py	1	CODE
LOW	olmocr/prompts/__init__.py	1	CODE
LOW	olmocr/prompts/__init__.py	1	CODE
LOW	olmocr/prompts/__init__.py	1	CODE
LOW	olmocr/prompts/__init__.py	1	CODE
LOW	olmocr/filter/__init__.py	1	CODE
LOW	olmocr/data/repackage_olmocrmix.py	12	CODE

Verbosity Indicators31 hits · 50 pts

Severity	File	Line	Snippet	Context
LOW⚡	tests/test_olmocrmix.py	17	# Step 1: Repackage the sample dataset into parquet + tarballs	COMMENT
LOW	tests/test_olmocrmix.py	44	# Step 2: Repackage the sample dataset into parquet + tarballs	COMMENT
LOW⚡	olmocr/bench/scripts/url_matcher.py	35	# Step 1: Get all local JSONL files	COMMENT
LOW⚡	olmocr/bench/scripts/url_matcher.py	39	# Step 2: Extract URLs from local files	COMMENT
LOW⚡	olmocr/bench/scripts/url_matcher.py	58	# Step 3: Load Hugging Face dataset	COMMENT
LOW⚡	olmocr/bench/scripts/url_matcher.py	64	# Step 4: Extract URLs from Hugging Face dataset	COMMENT
LOW⚡	olmocr/bench/scripts/url_matcher.py	79	# Step 5: Find matches	COMMENT
LOW⚡	olmocr/bench/scripts/url_matcher.py	82	# Step 6: Group matches by local file with metadata	COMMENT
LOW⚡	olmocr/synth/mine_html_templates.py	529	# Step 3: Render the initial HTML to PDF and then back to PNG for comparison	COMMENT
LOW	olmocr/synth/mine_html_templates.py	390	# Step 0: Check that the orientation of the original document is right-side-up. If not, we will	COMMENT
LOW	olmocr/synth/mine_html_templates.py	436	# Step 1: Initial analysis and column detection	COMMENT
LOW	olmocr/synth/mine_html_templates.py	476	# Step 2: Initial HTML generation with detailed layout instructions	COMMENT
LOW	olmocr/synth/mine_html_templates.py	562	# Step 4: Refinement - Show both images to Claude and ask for corrections	COMMENT
LOW⚡	olmocr/synth/mine_html_templates.py	861	# Step 1: Process headers, footers, and page numbers for TextAbsenceTests	STRING
LOW⚡	olmocr/synth/mine_html_templates.py	1476	# Step 6: Generate FootnoteTests for footnotes on the page	STRING
LOW	olmocr/synth/mine_html_templates.py	943	# Step 2: Generate tests from tables using parse_html_tables	STRING
LOW	olmocr/synth/mine_html_templates.py	1054	# Step 3: Generate TextPresenceTests and OrderingTests from markdown content	STRING
LOW	olmocr/synth/mine_html_templates.py	1368	# Step 4: Generate Math tests for LaTeX equations from the markdown	STRING
LOW	olmocr/synth/mine_html_templates.py	1409	# Step 5: Generate FormatTests for headings, bold, and italic text	STRING
LOW⚡	olmocr/synth/rotate_html_templates.py	234	# Step 1: Copy entire directory structure	COMMENT
LOW⚡	olmocr/synth/rotate_html_templates.py	241	# Step 2: Find all PDFs in the destination bench_data/pdfs directory	COMMENT
LOW⚡	olmocr/synth/rotate_html_templates.py	262	# Step 3: Select PDFs to rotate based on percentage	COMMENT
LOW⚡	olmocr/synth/rotate_html_templates.py	271	# Step 4: Rotate selected PDFs and update corresponding markdown files	COMMENT
LOW	olmocr/data/prepare_olmocrmix.py	148	# Step 1: Download dataset using hugging face hub snapshot_download to destination/hugging_face folder	COMMENT
LOW	olmocr/data/prepare_olmocrmix.py	197	# Step 2: Create destination folder structure for processed markdown files	COMMENT
LOW⚡	olmocr/data/prepare_olmocrmix.py	216	# Step 3: Extract PDF tarballs	COMMENT
LOW⚡	olmocr/data/prepare_olmocrmix.py	265	# Step 4: Process parquet files	COMMENT
LOW⚡	scripts/check_contamination.py	197	# Step 1: Get all URLs from bench data	COMMENT
LOW⚡	scripts/check_contamination.py	202	# Step 2: Read metadata JSONL and process source URLs	COMMENT
LOW	scripts/check_contamination.py	216	# Step 3: Map URLs to hashes and query database	COMMENT
LOW	scripts/check_contamination.py	285	# Step 4: Check for contamination	COMMENT

Structural Annotation Overuse31 hits · 50 pts

Severity	File	Line	Snippet	Context
LOW⚡	tests/test_olmocrmix.py	17	# Step 1: Repackage the sample dataset into parquet + tarballs	COMMENT
LOW	tests/test_olmocrmix.py	44	# Step 2: Repackage the sample dataset into parquet + tarballs	COMMENT
LOW⚡	olmocr/bench/scripts/url_matcher.py	35	# Step 1: Get all local JSONL files	COMMENT
LOW⚡	olmocr/bench/scripts/url_matcher.py	39	# Step 2: Extract URLs from local files	COMMENT
LOW⚡	olmocr/bench/scripts/url_matcher.py	58	# Step 3: Load Hugging Face dataset	COMMENT
LOW⚡	olmocr/bench/scripts/url_matcher.py	64	# Step 4: Extract URLs from Hugging Face dataset	COMMENT
LOW⚡	olmocr/bench/scripts/url_matcher.py	79	# Step 5: Find matches	COMMENT
LOW⚡	olmocr/bench/scripts/url_matcher.py	82	# Step 6: Group matches by local file with metadata	COMMENT
LOW⚡	olmocr/synth/mine_html_templates.py	529	# Step 3: Render the initial HTML to PDF and then back to PNG for comparison	COMMENT
LOW	olmocr/synth/mine_html_templates.py	390	# Step 0: Check that the orientation of the original document is right-side-up. If not, we will	COMMENT
LOW	olmocr/synth/mine_html_templates.py	436	# Step 1: Initial analysis and column detection	COMMENT
LOW	olmocr/synth/mine_html_templates.py	476	# Step 2: Initial HTML generation with detailed layout instructions	COMMENT
LOW	olmocr/synth/mine_html_templates.py	562	# Step 4: Refinement - Show both images to Claude and ask for corrections	COMMENT
LOW⚡	olmocr/synth/mine_html_templates.py	861	# Step 1: Process headers, footers, and page numbers for TextAbsenceTests	STRING
LOW⚡	olmocr/synth/mine_html_templates.py	1476	# Step 6: Generate FootnoteTests for footnotes on the page	STRING
LOW	olmocr/synth/mine_html_templates.py	943	# Step 2: Generate tests from tables using parse_html_tables	STRING
LOW	olmocr/synth/mine_html_templates.py	1054	# Step 3: Generate TextPresenceTests and OrderingTests from markdown content	STRING
LOW	olmocr/synth/mine_html_templates.py	1368	# Step 4: Generate Math tests for LaTeX equations from the markdown	STRING
LOW	olmocr/synth/mine_html_templates.py	1409	# Step 5: Generate FormatTests for headings, bold, and italic text	STRING
LOW⚡	olmocr/synth/rotate_html_templates.py	234	# Step 1: Copy entire directory structure	COMMENT
LOW⚡	olmocr/synth/rotate_html_templates.py	241	# Step 2: Find all PDFs in the destination bench_data/pdfs directory	COMMENT
LOW⚡	olmocr/synth/rotate_html_templates.py	262	# Step 3: Select PDFs to rotate based on percentage	COMMENT
LOW⚡	olmocr/synth/rotate_html_templates.py	271	# Step 4: Rotate selected PDFs and update corresponding markdown files	COMMENT
LOW	olmocr/data/prepare_olmocrmix.py	148	# Step 1: Download dataset using hugging face hub snapshot_download to destination/hugging_face folder	COMMENT
LOW	olmocr/data/prepare_olmocrmix.py	197	# Step 2: Create destination folder structure for processed markdown files	COMMENT
LOW⚡	olmocr/data/prepare_olmocrmix.py	216	# Step 3: Extract PDF tarballs	COMMENT
LOW⚡	olmocr/data/prepare_olmocrmix.py	265	# Step 4: Process parquet files	COMMENT
LOW⚡	scripts/check_contamination.py	197	# Step 1: Get all URLs from bench data	COMMENT
LOW⚡	scripts/check_contamination.py	202	# Step 2: Read metadata JSONL and process source URLs	COMMENT
LOW	scripts/check_contamination.py	216	# Step 3: Map URLs to hashes and query database	COMMENT
LOW	scripts/check_contamination.py	285	# Step 4: Check for contamination	COMMENT

Over-Commented Block23 hits · 21 pts

Severity	File	Line	Snippet	Context
LOW	olmocr/bench/scripts/convert_all.sh	221	# Run raw server benchmarks with generic server function	COMMENT
LOW	olmocr/bench/scripts/convert_all.sh	241		COMMENT
LOW	olmocr/synth/mine_html_templates.py	1281	# Run the test against the markdown content	COMMENT
LOW	olmocr/synth/mine_html_templates.py	1301		COMMENT
LOW	olmocr/synth/mine_html_templates.py	1321	# # Calculate similarity scores for each absent word	COMMENT
LOW	olmocr/synth/mine_html_templates.py	1341		COMMENT
LOW	…igs/v0.5.0/qwen3_vl_olmocrv5_1epoch_mix_1025_mini.yaml	61	# - name: processed_00_documents_eval_s2pdf	COMMENT
LOW	olmocr/data/prepare_national_archive_transcripts.py	1	# This script prepares transcriptions from the National Archives into a format usable by olmOCR	COMMENT
LOW	olmocr/data/prepare_national_archive_transcripts.py	21	# "objectFileSize": 9496446,	COMMENT
LOW	olmocr/data/prepare_national_archive_transcripts.py	41	# "userId": "dff3eed0-38e5-35fc-b7e7-d2d58b023262",	COMMENT
LOW	olmocr/data/prepare_national_archive_transcripts.py	61	# }	COMMENT
LOW	scripts/remove_paths_from_olmocrmix.py	1	# This code takes in a path to an olmocr-mix processed data folder	COMMENT
LOW	scripts/run_transformers_benchmark.sh	1	#!/bin/bash	COMMENT
LOW	scripts/run_server_benchmark.sh	1	#!/bin/bash	COMMENT
LOW	scripts/check_contamination.py	1	#!/usr/bin/env python3	COMMENT
LOW	scripts/run_dots_mocr_benchmark.sh	1	#!/bin/bash	COMMENT
LOW	scripts/run_api_benchmark.sh	1	#!/bin/bash	COMMENT
LOW	scripts/run_mineru_benchmark.sh	1	#!/bin/bash	COMMENT
LOW	scripts/run_benchmark.sh	1	#!/bin/bash	COMMENT
LOW	scripts/run_qianfan_benchmark.sh	1	#!/bin/bash	COMMENT
LOW	scripts/run_infrapartner_benchmark.sh	1	#!/bin/bash	COMMENT
LOW	scripts/run_chandra_benchmark.sh	1	#!/bin/bash	COMMENT
LOW	.github/workflows/main.yml	301	# with:	COMMENT

Docstring Block Structure4 hits · 20 pts

Severity	File	Line	Snippet	Context
HIGH	olmocr/image_utils.py	7	Convert one or multiple image files to PDF bytes. Args: image_files: A single image file path (str) or	STRING
HIGH	olmocr/bench/tests.py	767	Load a single test from a JSON line string or JSON object. Args: data: Either a JSON string to parse o	STRING
HIGH	olmocr/data/renderpdf.py	74	Returns the (width, height) of a PNG image given its base64-encoded data, without base64-decoding the entire da	STRING
HIGH	scripts/pii/pii_rule_comparison.py	1411	Generate HTML section with CDF plots for all numeric attributes. Args: numeric_attributes: Dictionary	STRING

Modern Structural Boilerplate18 hits · 18 pts

Severity	File	Line	Snippet	Context
LOW	olmocr/check.py	6	logger = logging.getLogger(__name__)	CODE
LOW	olmocr/s3_utils.py	22	logger = logging.getLogger(__name__)	CODE
LOW	olmocr/pipeline.py	55	logger = logging.getLogger(__name__)	CODE
LOW	olmocr/work_queue.py	23	logger = logging.getLogger(__name__)	CODE
LOW	olmocr/bench/runners/run_gemini.py	24	logger = logging.getLogger(__name__)	CODE
LOW	olmocr/bench/runners/run_chatgpt.py	28	logger = logging.getLogger(__name__)	CODE
LOW	olmocr/train/front_matter.py	18	logger = logging.getLogger(__name__)	CODE
LOW	olmocr/train/train.py	37	logger = logging.getLogger(__name__)	CODE
LOW	olmocr/train/grpo_train.py	46	logger = logging.getLogger(__name__)	CODE
LOW	olmocr/train/dataloader.py	39	logger = logging.getLogger(__name__)	CODE
LOW	olmocr/train/compare_vllm_checkpoint.py	28	logger = logging.getLogger(__name__)	CODE
LOW	olmocr/filter/filter.py	10	logger = logging.getLogger(__name__)	CODE
LOW	olmocr/data/prepare_workspace.py	29	logger = logging.getLogger(__name__)	CODE
LOW	scripts/pii/rich_tagging_pipeline.py	42	logger = logging.getLogger(__name__)	CODE
LOW	scripts/pii/pii_rule_comparison.py	59	logger = logging.getLogger(__name__)	CODE
LOW	scripts/pii/chatgpt_tag_dolmadocs_v1.py	32	logger = logging.getLogger(__name__)	CODE
LOW	scripts/pii/chatgpt_tag_dolmadocs_v2.py	32	logger = logging.getLogger(__name__)	CODE
LOW	scripts/pii/tagging_pipeline.py	42	logger = logging.getLogger(__name__)	CODE

Slop Phrases6 hits · 12 pts

Severity	File	Line	Snippet	Context
LOW	olmocr/pipeline.py	470	# But just make sure to close the socket on your way out	COMMENT
MEDIUM	olmocr/data/buildsilver.py	39	# Ask for logprobs, it's not any more expensive and you can use them later to help identify problematic responses	COMMENT
LOW	scripts/pii/rich_tagging_pipeline.py	301	# But just make sure to close the socket on your way out	STRING
LOW	scripts/pii/tagging_pipeline.py	227	# But just make sure to close the socket on your way out	COMMENT
MEDIUM	scripts/data/buildsilver.py	39	# Ask for logprobs, it's not any more expensive and you can use them later to help identify problematic responses	COMMENT
MEDIUM	.github/workflows/main.yml	298	# jakep: (you can try by setting load: true above, but you'll need a larger runner)	COMMENT

Example Usage Blocks6 hits · 10 pts

Severity	File	Line	Snippet	Context
LOW	olmocr/bench/runners/run_olmocr_pipeline.py	100	# Example usage	COMMENT
LOW	scripts/run_dots_mocr_benchmark.sh	5	# Usage:	COMMENT
LOW⚡	scripts/jsonl_to_markdown.py	40	# Example usage	COMMENT
LOW	scripts/run_qianfan_benchmark.sh	8	# Usage:	COMMENT
LOW	scripts/run_chandra_benchmark.sh	4	# Usage:	COMMENT
LOW	scripts/eval/scoreelo.py	288	# Example usage	COMMENT

Modern AI Meta-Vocabulary3 hits · 9 pts

Severity	File	Line	Snippet	Context
MEDIUM	olmocr/data/buildsilver.py	37	# Also, structured outputs let you cheat, because the order in which fields are in your schema, is the order in whic	COMMENT
MEDIUM	scripts/data/buildsilver.py	37	# Also, structured outputs let you cheat, because the order in which fields are in your schema, is the order in whic	COMMENT
MEDIUM	scripts/plots/plot_olmocr2_timeline.py	210	# "descriptor": "Fixing bug with blank page hallucinations",	COMMENT

Decorative Section Separators4 hits · 8 pts

Severity	File	Line	Snippet	Context
MEDIUM	scripts/pii/rich_tagging_pipeline.py	511	# ------------------------------------------------------------------	STRING
MEDIUM	scripts/pii/rich_tagging_pipeline.py	513	# ------------------------------------------------------------------	STRING
MEDIUM	scripts/pii/tagging_pipeline.py	346	# ------------------------------------------------------------------	COMMENT
MEDIUM	scripts/pii/tagging_pipeline.py	348	# ------------------------------------------------------------------	COMMENT

Synthetic Comment Markers1 hit · 8 pts

Severity	File	Line	Snippet	Context
HIGH	olmocr/data/prepare_national_archive_transcripts.py	246	# Skip if AI generated tags	COMMENT

Fake / Example Data2 hits · 3 pts

Severity	File	Line	Snippet	Context
LOW⚡	tests/test_tests.py	604	long_text = "Lorem ipsum dolor sit amet, " * 100	CODE
LOW⚡	tests/test_tests.py	604	long_text = "Lorem ipsum dolor sit amet, " * 100	CODE

Overly Generic Function Names1 hit · 2 pts

Severity	File	Line	Snippet	Context
LOW⚡	tests/test_mine_html_templates_introduce_text_errors.py	194	<script>function doSomething() { return value; }</script>	CODE

AI Structural Patterns1 hit · 1 pts

Severity	File	Line	Snippet	Context
LOW	olmocr/bench/runners/run_dotsocr.py	147		CODE

Analysis Overview

What These Metrics Mean

Score History

Severity Breakdown

Directory Score Breakdown

Pattern Findings