Fix misc bugs

convert_single.py

@@ -3,8 +3,6 @@ import time
 import pypdfium2 # Needs to be at the top to avoid warnings
 import os
 
-from marker.settings import settings
-
 os.environ["PYTORCH_ENABLE_MPS_FALLBACK"] = "1" # For some reason, transformers decided to use .isin for a simple op, which is not supported on MPS
 
 import argparse

marker/convert.py

@@ -1,7 +1,4 @@
 import warnings
-
-from marker.cleaners.toc import compute_toc
-
 warnings.filterwarnings("ignore", category=UserWarning) # Filter torch pytree user warnings
 
 import os
@@ -31,6 +28,7 @@ from marker.postprocessors.markdown import merge_spans, merge_lines, get_full_te
 from marker.cleaners.text import cleanup_text
 from marker.images.extract import extract_images
 from marker.images.save import images_to_dict
+from marker.cleaners.toc import compute_toc
 
 from typing import List, Dict, Tuple, Optional
 from marker.settings import settings

marker/equations/equations.py

@@ -6,7 +6,7 @@ from marker.equations.inference import get_total_texify_tokens, get_latex_batche
 from marker.pdf.images import render_bbox_image
 from marker.schema.bbox import rescale_bbox
 from marker.schema.page import Page
-from marker.schema.block import Line, Span, Block, bbox_from_lines, split_block_lines, find_insert_block
+from marker.schema.block import Line, Span, Block, split_block_lines, find_insert_block
 from marker.settings import settings
 
 

poetry.lock

@@ -2510,13 +2510,13 @@ testing = ["docopt", "pytest"]
 
 [[package]]
 name = "pdftext"
-version = "0.3.15"
+version = "0.3.16"
 description = "Extract structured text from pdfs quickly"
 optional = false
 python-versions = "<4.0,>=3.10"
 files = [
-    {file = "pdftext-0.3.15-py3-none-any.whl", hash = "sha256:3151abacd5c2cfed9975d090333b543151e14de439ebe0d228b935328d512f3d"},
-    {file = "pdftext-0.3.15.tar.gz", hash = "sha256:3c6d55781c1adfd263cdc05c39cbea2c40d4e626439ab24078f860eca65c2e6c"},
+    {file = "pdftext-0.3.16-py3-none-any.whl", hash = "sha256:c541ee95496b51325fcdeebf9390a45cd2a1f511aac21b021bb45cca4634b6b7"},
+    {file = "pdftext-0.3.16.tar.gz", hash = "sha256:a30f4fddafe1ad67fb45b18689942a78c6263ef4a53888bd107cc25c27974a9d"},
 ]
 
 [package.dependencies]
@@ -5075,4 +5075,4 @@ propcache = ">=0.2.0"
 [metadata]
 lock-version = "2.0"
 python-versions = "^3.10"
-content-hash = "4983d14f11f46193fc13b3256d1f7ed8d0877ad579cb8b30fbab8e87d3febb22"
+content-hash = "fb9aaf82b97c40f5f850b3143558d44e7f6a501eb34e40bdfdc5d3a0aa58fecf"

pyproject.toml

@@ -35,7 +35,7 @@ rapidfuzz = "^3.8.1"
 surya-ocr = "^0.6.3"
 filetype = "^1.2.0"
 regex = "^2024.4.28"
-pdftext = "^0.3.15"
+pdftext = "^0.3.16"
 tabled-pdf = "^0.1.0"
 
 [tool.poetry.group.dev.dependencies]

scripts/verify_benchmark_scores.py

@@ -9,7 +9,7 @@ def verify_scores(file_path):
     multicolcnn_score = data["marker"]["files"]["multicolcnn.pdf"]["score"]
     switch_trans_score = data["marker"]["files"]["switch_trans.pdf"]["score"]
 
-    if multicolcnn_score <= 0.35 or switch_trans_score <= 0.39:
+    if multicolcnn_score <= 0.37 or switch_trans_score <= 0.4:
         raise ValueError("One or more scores are below the required threshold of 0.4")