feat: add PDF processing resumption via filename tracking, centralise DefaultPaths constants and add VLM paper citation

Author: aritraroy24

CHANGELOG.md

@@ -1,3 +1,13 @@
+# Unreleased
+
+### Added
+
+- Added `is_track_pdfs` and `track_pdfs_report_path` to `process_articles()` for local PDF workflows. When enabled (default), each processed PDF is recorded as a `filename<TAB>doi` entry in `logs/{keyword}_pdf_processed_dois.txt`, allowing re-runs to skip already-processed PDFs before any conversion or API calls. Falls back to scanning the output CSV when the tracking file does not yet exist.
+
+- Centralised non-keyword default file paths (`results/failed_automated_articles.txt`, `agentic_evaluation_result.json`, `detailed_evaluation.json`) as class-level constants on `DefaultPaths` so they can be changed in one place.
+
+---
+
 # 2026.05.19
 
 ### Added

CITATION.cff

@@ -50,6 +50,28 @@ preferred-citation:
   title: "ComProScanner: a multi-agent based framework for composition-property structured data extraction from scientific literature"
   type: article
   url: "https://doi.org/10.1039/D5DD00521C"
+references:
+  - authors:
+      - family-names: Roy
+        given-names: Aritra
+        orcid: "https://orcid.org/0000-0003-0243-9124"
+      - family-names: Grisan
+        given-names: Enrico
+        orcid: "https://orcid.org/0000-0002-7365-5652"
+      - family-names: Gattinoni
+        given-names: Chiara
+        orcid: "https://orcid.org/0000-0002-3376-6374"
+      - family-names: Buckeridge
+        given-names: John
+        orcid: "https://orcid.org/0000-0002-2537-5082"
+    identifiers:
+      - type: other
+        value: "arXiv:2606.00065"
+        description: "arXiv preprint"
+    title: "Beyond Text and Tables: Vision-Language Model Integration in ComProScanner for Extracting Materials Data from Scientific Figures with High Accuracy"
+    type: article
+    year: 2026
+    url: "https://arxiv.org/abs/2606.00065"
 repository-code: "https://github.com/slimeslab/ComProScanner"
 license: MIT
 title: "ComProScanner: A multi-agent based framework for composition-property structured data extraction from scientific literature"

README.md

@@ -2,7 +2,7 @@
   <img src="https://raw.githubusercontent.com/aritraroy24/ComProScanner/refs/heads/main/assets/comproscanner_logo.png" alt="ComProScanner Logo" width="500"/>
 </p>
 
-[![Python Version](https://img.shields.io/badge/python-3.12%20%7C%203.13-blue.svg?logo=python&logoColor=white)](https://www.python.org/downloads/) [![License: MIT](https://custom-icon-badges.demolab.com/badge/license-MIT-brown.svg?logo=law&logoColor=white)](https://opensource.org/licenses/MIT) [![PyPI](https://img.shields.io/pypi/v/comproscanner?logo=pypi&logoColor=white)](https://pypi.org/project/comproscanner/) [![Documentation](https://custom-icon-badges.demolab.com/badge/docs-latest-brightgreen.svg?logo=materialformkdocs&logoColor=white)](https://slimeslab.github.io/ComProScanner/) [![Coverage](https://img.shields.io/codecov/c/github/aritraroy24/ComProScanner?logo=codecov&logoColor=white&label=coverage&color=e62277)](https://codecov.io/gh/aritraroy24/ComProScanner) [![PyPI - Downloads](https://custom-icon-badges.demolab.com/pypi/dm/comproscanner?logo=download&logoColor=white&color=purple)](https://pypistats.org/packages/comproscanner) [![Ask DeepWiki](https://custom-icon-badges.demolab.com/badge/Ask%20DeepWiki-brightgreen.svg?logo=deepwikidevin&logoColor=white&labelColor=grey&color=5ab998)](https://deepwiki.com/slimeslab/ComProScanner) [![Digital Discovery](https://custom-icon-badges.demolab.com/badge/Digital_Discovery-10.1039/D5DD00521C-brightgreen.svg?logo=rsc&logoColor=white&color=c8c300)](https://doi.org/10.1039/D5DD00521C)
+[![Python Version](https://img.shields.io/badge/python-3.12%20%7C%203.13-blue.svg?logo=python&logoColor=white)](https://www.python.org/downloads/) [![License: MIT](https://custom-icon-badges.demolab.com/badge/license-MIT-brown.svg?logo=law&logoColor=white)](https://opensource.org/licenses/MIT) [![PyPI](https://img.shields.io/pypi/v/comproscanner?logo=pypi&logoColor=white)](https://pypi.org/project/comproscanner/) [![Documentation](https://custom-icon-badges.demolab.com/badge/docs-latest-brightgreen.svg?logo=materialformkdocs&logoColor=white)](https://slimeslab.github.io/ComProScanner/) [![Coverage](https://img.shields.io/codecov/c/github/aritraroy24/ComProScanner?logo=codecov&logoColor=white&label=coverage&color=e62277)](https://codecov.io/gh/aritraroy24/ComProScanner) [![PyPI - Downloads](https://custom-icon-badges.demolab.com/pypi/dm/comproscanner?logo=download&logoColor=white&color=purple)](https://pypistats.org/packages/comproscanner) [![Ask DeepWiki](https://custom-icon-badges.demolab.com/badge/Ask%20DeepWiki-brightgreen.svg?logo=deepwikidevin&logoColor=white&labelColor=grey&color=5ab998)](https://deepwiki.com/slimeslab/ComProScanner) [![Digital Discovery](https://custom-icon-badges.demolab.com/badge/Digital_Discovery-10.1039/D5DD00521C-brightgreen.svg?logo=rsc&logoColor=white&color=c8c300)](https://doi.org/10.1039/D5DD00521C) [![arXiv Preprint](https://custom-icon-badges.demolab.com/badge/arXiv-2606.00065-brightgreen.svg?logo=arxiv&logoColor=white&color=b22929)](https://arxiv.org/abs/2606.00065)
 
 # ComProScanner
 
@@ -129,20 +129,30 @@ The ComProScanner workflow consists of four main stages:
 
 ## Citation
 
-If you use ComProScanner in your research, please cite:
+If you use ComProScanner in your research, please cite the following papers:
 
 ```bibtex
-@Article{roy2026comproscanner,
-  title={ComProScanner: a multi-agent based framework for composition-property structured data extraction from scientific literature},
-  author={Roy, Aritra and Grisan, Enrico and Buckeridge, John and Gattinoni, Chiara},
-  journal={Digital Discovery},
-  volume={5},
-  number={4},
-  pages={1794--1808},
-  year={2026},
-  publisher={Royal Society of Chemistry},
-  doi  ="10.1039/D5DD00521C",
-  url  ="https://doi.org/10.1039/D5DD00521C"
+@article{roy2026comproscanner,
+      title={ComProScanner: a multi-agent based framework for composition-property structured data extraction from scientific literature},
+      author={Roy, Aritra and Grisan, Enrico and Buckeridge, John and Gattinoni, Chiara},
+      journal={Digital Discovery},
+      volume={5},
+      number={4},
+      pages={1794--1808},
+      year={2026},
+      publisher={Royal Society of Chemistry},
+      doi  ="10.1039/D5DD00521C",
+      url  ="https://doi.org/10.1039/D5DD00521C"
+}
+@misc{roy2026comproscanner_vlm,
+      title={Beyond Text and Tables: Vision-Language Model Integration in ComProScanner for Extracting Materials Data from Scientific Figures with High Accuracy}, 
+      author={Aritra Roy and Enrico Grisan and Chiara Gattinoni and John Buckeridge},
+      year={2026},
+      eprint={2606.00065},
+      archivePrefix={arXiv},
+      primaryClass={cs.IR},
+      doi={10.48550/arXiv.2606.00065},
+      url={https://arxiv.org/abs/2606.00065}, 
 }
 ```
 

docs/about/changelog.md

@@ -1,3 +1,13 @@
+# Unreleased
+
+### Added
+
+- Added `is_track_pdfs` and `track_pdfs_report_path` to `process_articles()` for local PDF workflows. When enabled (default), each processed PDF is recorded as a `filename<TAB>doi` entry in `logs/{keyword}_pdf_processed_dois.txt`, allowing re-runs to skip already-processed PDFs before any conversion or API calls. Falls back to scanning the output CSV when the tracking file does not yet exist.
+
+- Centralised non-keyword default file paths (`results/failed_automated_articles.txt`, `agentic_evaluation_result.json`, `detailed_evaluation.json`) as class-level constants on `DefaultPaths` so they can be changed in one place.
+
+---
+
 # 2026.05.19
 
 ### Added
@@ -28,6 +38,10 @@
 
 - Added `save_failed_automated_report` and `failed_automated_report_path` to `process_articles()` for automated publisher sources (Elsevier, Springer Nature, IOP, Wiley), mirroring the existing PDF failure report. Failed articles are written as tab-separated `doi`, `publisher`, `reason` entries to `results/failed_automated_articles.txt` by default.
 
+- Added `is_track_pdfs` and `track_pdfs_report_path` to `process_articles()` for local PDF workflows. When enabled (default), each processed PDF is recorded as a `filename<TAB>doi` entry in `logs/{keyword}_pdf_processed_dois.txt`, allowing re-runs to skip already-processed PDFs before any conversion or API calls. Falls back to scanning the output CSV when the tracking file does not yet exist.
+
+- Centralised default file paths (`results/failed_automated_articles.txt`, `agentic_evaluation_result.json`, `detailed_evaluation.json`) as class-level constants on `DefaultPaths` so they can be changed in one place.
+
 - Added image-aware fallback in `DataExtractionFlow.identify_materials_data_presence()`:
 
   - The Materials Data Identifier still runs text RAG first.

docs/about/citation.md

@@ -3,16 +3,26 @@
 If you use ComProScanner in your research, please cite our related paper:
 
 ```bibtex
-@Article{roy2026comproscanner,
-  title={ComProScanner: a multi-agent based framework for composition-property structured data extraction from scientific literature},
-  author={Roy, Aritra and Grisan, Enrico and Buckeridge, John and Gattinoni, Chiara},
-  journal={Digital Discovery},
-  volume={5},
-  number={4},
-  pages={1794--1808},
-  year={2026},
-  publisher={Royal Society of Chemistry},
-  doi  ="10.1039/D5DD00521C",
-  url  ="https://doi.org/10.1039/D5DD00521C"
+@article{roy2026comproscanner,
+      title={ComProScanner: a multi-agent based framework for composition-property structured data extraction from scientific literature},
+      author={Roy, Aritra and Grisan, Enrico and Buckeridge, John and Gattinoni, Chiara},
+      journal={Digital Discovery},
+      volume={5},
+      number={4},
+      pages={1794--1808},
+      year={2026},
+      publisher={Royal Society of Chemistry},
+      doi  ="10.1039/D5DD00521C",
+      url  ="https://doi.org/10.1039/D5DD00521C"
+}
+@misc{roy2026comproscanner_vlm,
+      title={Beyond Text and Tables: Vision-Language Model Integration in ComProScanner for Extracting Materials Data from Scientific Figures with High Accuracy}, 
+      author={Aritra Roy and Enrico Grisan and Chiara Gattinoni and John Buckeridge},
+      year={2026},
+      eprint={2606.00065},
+      archivePrefix={arXiv},
+      primaryClass={cs.IR},
+      doi={10.48550/arXiv.2606.00065},
+      url={https://arxiv.org/abs/2606.00065}, 
 }
 ```

docs/index.md

@@ -13,6 +13,7 @@
   <a href="https://pypistats.org/packages/comproscanner"><img src="https://custom-icon-badges.demolab.com/pypi/dm/comproscanner?logo=download&logoColor=white&color=purple" alt="Downloads"></a>
   <a href="https://deepwiki.com/slimeslab/ComProScanner"><img src="https://custom-icon-badges.demolab.com/badge/Ask%20DeepWiki-brightgreen.svg?logo=deepwikidevin&logoColor=white&labelColor=grey&color=5ab998" alt="Ask DeepWiki"></a>
   <a href="https://doi.org/10.1039/D5DD00521C"><img src="https://custom-icon-badges.demolab.com/badge/Digital_Discovery-10.1039/D5DD00521C-brightgreen.svg?logo=rsc&logoColor=white&color=c8c300" alt="Digital Discovery"></a>
+  <a href="https://arxiv.org/abs/2606.00065"><img src="https://custom-icon-badges.demolab.com/badge/arXiv-2606.00065-brightgreen.svg?logo=arxiv&logoColor=white&color=b22929" alt="arXiv Preprint"></a>
 </p>
 
 ## Welcome
@@ -155,15 +156,25 @@ Read the details of ComProScanner in the following Digital Discovery paper: [10.
 If you use ComProScanner in your research, please cite:
 
 ```bibtex
-@Article{roy2026comproscannermultiagentbasedframework,
-author ="Roy, Aritra and Grisan, Enrico and Buckeridge, John and Gattinoni, Chiara",
-title  ="ComProScanner: a multi-agent based framework for composition-property structured data extraction from scientific literature",
-journal  ="Digital Discovery",
-year  ="2026",
-pages  ="Accepted",
-publisher  ="RSC",
-doi  ="10.1039/D5DD00521C",
-url  ="https://doi.org/10.1039/D5DD00521C"
+@article{roy2026comproscannermultiagentbasedframework,
+      author ="Roy, Aritra and Grisan, Enrico and Buckeridge, John and Gattinoni, Chiara",
+      title  ="ComProScanner: a multi-agent based framework for composition-property structured data extraction from scientific literature",
+      journal  ="Digital Discovery",
+      year  ="2026",
+      pages  ="Accepted",
+      publisher  ="RSC",
+      doi  ="10.1039/D5DD00521C",
+      url  ="https://doi.org/10.1039/D5DD00521C"
+}
+@misc{roy2026comproscanner_vlm,
+      title={Beyond Text and Tables: Vision-Language Model Integration in ComProScanner for Extracting Materials Data from Scientific Figures with High Accuracy}, 
+      author={Aritra Roy and Enrico Grisan and Chiara Gattinoni and John Buckeridge},
+      year={2026},
+      eprint={2606.00065},
+      archivePrefix={arXiv},
+      primaryClass={cs.IR},
+      doi={10.48550/arXiv.2606.00065},
+      url={https://arxiv.org/abs/2606.00065}, 
 }
 ```
 

docs/news/posts/graphtool-vlm-benchmark.md

@@ -1,6 +1,6 @@
 ---
 date:
-  created: 2026-05-08
+  created: 2026-06-03
 authors:
   - aritraroy24
 cover_image: https://i.ibb.co/jPJf86Qg/comproscanner-vlm-integration.png

docs/usage/article-processing.md

@@ -132,6 +132,14 @@ For `source_list=["pdfs"]` processing only. If `True`, saves a text report for P
 
 For `source_list=["pdfs"]` processing only. Custom output path for the failed PDF filename report. If not provided, defaults to `{folder_path}/failed_pdf_filenames.txt`.
 
+#### :material-square-medium:`is_track_pdfs` _(bool)_
+
+For `source_list=["pdfs"]` processing only. If `True`, each successfully processed PDF is recorded in a plain-text tracking file as a tab-separated `filename<TAB>doi` entry so that re-runs skip already-processed PDFs before any conversion or API calls. Falls back to scanning the output CSV when the tracking file does not yet exist. Defaults to `True`.
+
+#### :material-square-medium:`track_pdfs_report_path` _(str)_
+
+For `source_list=["pdfs"]` processing only. Custom path for the PDF tracking file. If not provided, defaults to `logs/{keyword}_pdf_processed_dois.txt`.
+
 #### :material-square-medium:`save_failed_automated_report` _(bool)_
 
 For automated publisher sources (`elsevier`, `springer`, `iop`, `wiley`). If `True`, appends a tab-separated record for every article that could not be downloaded or parsed to the failure report. Each line contains three fields: `doi`, `publisher`, and a short reason code:
@@ -150,7 +158,7 @@ Custom output path for the automated failure report. If not provided, defaults t
 
 !!! info "Default Values"
 
-    :material-square-small:**`source_list`** = ["elsevier", "wiley", "iop", "springer"]<br>:material-square-small:**`folder_path`** = None<br>:material-square-small:**`doi_list`** = None<br>:material-square-small:**`is_sql_db`** = False<br>:material-square-small:**`is_save_xml`** = False<br>:material-square-small:**`is_save_pdf`** = False<br>:material-square-small:**`rag_db_path`** = "db"<br>:material-square-small:**`chunk_size`** = 1000<br>:material-square-small:**`chunk_overlap`** = 25<br>:material-square-small:**`embedding_model`** = "huggingface:thellert/physbert_cased"<br>:material-square-small:**`main_figure_keywords`** = `property_keywords`<br>:material-square-small:**`additional_figure_keywords`** = None<br>:material-square-small:**`save_failed_pdf_report`** = True<br>:material-square-small:**`failed_pdf_report_path`** = None (auto: `{folder_path}/failed_pdf_filenames.txt`)<br>:material-square-small:**`save_failed_automated_report`** = True<br>:material-square-small:**`failed_automated_report_path`** = None (auto: `results/failed_automated_articles.txt`)
+    :material-square-small:**`source_list`** = ["elsevier", "wiley", "iop", "springer"]<br>:material-square-small:**`folder_path`** = None<br>:material-square-small:**`doi_list`** = None<br>:material-square-small:**`is_sql_db`** = False<br>:material-square-small:**`is_save_xml`** = False<br>:material-square-small:**`is_save_pdf`** = False<br>:material-square-small:**`rag_db_path`** = "db"<br>:material-square-small:**`chunk_size`** = 1000<br>:material-square-small:**`chunk_overlap`** = 25<br>:material-square-small:**`embedding_model`** = "huggingface:thellert/physbert_cased"<br>:material-square-small:**`main_figure_keywords`** = `property_keywords`<br>:material-square-small:**`additional_figure_keywords`** = None<br>:material-square-small:**`save_failed_pdf_report`** = True<br>:material-square-small:**`failed_pdf_report_path`** = None (auto: `{folder_path}/failed_pdf_filenames.txt`)<br>:material-square-small:**`is_track_pdfs`** = True<br>:material-square-small:**`track_pdfs_report_path`** = None (auto: `logs/{keyword}_pdf_processed_dois.txt`)<br>:material-square-small:**`save_failed_automated_report`** = True<br>:material-square-small:**`failed_automated_report_path`** = None (auto: `results/failed_automated_articles.txt`)
 
 ## Processing Workflow
 

src/comproscanner/__init__.py

@@ -16,6 +16,7 @@
 
 import sys
 import os
+from .utils.configs.paths_config import DefaultPaths
 
 # Check if the script is running under pytest
 _is_testing = "pytest" in sys.modules or "PYTEST_CURRENT_TEST" in os.environ
@@ -221,7 +222,7 @@ def evaluate_semantic(
     def evaluate_agentic(
         ground_truth_file=None,
         test_data_file=None,
-        output_file="agentic_evaluation_result.json",
+        output_file=DefaultPaths.AGENTIC_EVALUATION_RESULT_FILENAME,
         extraction_agent_model_name="gpt-4o-mini",
         is_synthesis_evaluation=True,
         weights=None,

src/comproscanner/article_processors/elsevier_processor.py

@@ -161,7 +161,7 @@ def __init__(
         )
         self.save_failed_automated_report = save_failed_automated_report
         self.failed_automated_report_path = (
-            failed_automated_report_path or "results/failed_automated_articles.txt"
+            failed_automated_report_path or self.all_paths.FAILED_AUTOMATED_ARTICLES_FILENAME
         )
         self.failed_automated_count = 0