Submission Checker

A command-line tool for academic conference submissions that automatically validates PDFs for compliance with conference policies.

Features

• Page limit checking – Warn when the number of pages exceeds a configurable limit.
• References position – Detect if references start after the allowed page.
• Content validation – Flag figures/tables on pages that should contain references only.
• Appendix outside main pages – Identify appendix sections that appear after the configured main-page limit.
• Style conformance – Verify conformance to ACM or IEEE citation style.
• References format check – Detect non-standard references styles such as bullet points, author-year entries, or plain numbering without brackets.
• Anonymity checks – Detect non-anonymous emails mentioned on page 1.
• Suspicious wording – Identify potentially revealing phrases like "our previous paper [3]".
• Metadata inspection – Inspect PDF metadata for possible author information that could reveal identity.

Options

• --file <path>: Path to a single PDF file to check
• --folder <path>: Path to folder containing PDFs to check (recursive)
• --max-pages <int>: Maximum total pages allowed (main text + references)
• --min-pages <int>: Minimum total pages required (main text + references)
• --main-pages <int>: Maximum pages for main text (default: 10)
• --style <acm|ieee>: Expected citation style for validation
• --timeout <int>: Maximum seconds for PDF text extraction (default: 10)
• --csv <path>: Output CSV report file (requires --folder)

Check Details

The tool performs the following checks on each PDF. All checks require successful text extraction from the PDF.

1. Page Limit Check

• Logic: Counts the total number of pages in the PDF. If the count exceeds the specified --max-pages, a warning is issued. If the count is less than the specified --min-pages, a warning is issued.
• Configuration: --max-pages (integer, optional), --min-pages (integer, optional). --main-pages (integer, default 10) specifies the limit for main text.
• Example: --max-pages 12 --main-pages 10 warns if PDF has more than 12 pages total.
• Example (min): --min-pages 8 warns if PDF has fewer than 8 pages total.

2. References Placement Check

• Logic: Scans each page for a line starting with "reference" or "references" (case-insensitive). Warns if references start after the total page limit or if they start after main text limit +1 (implying main text exceeds limit). If no references found, warns if total pages exceed main text limit.
• Configuration: Requires --max-pages and uses --main-pages (default 10).
• Note: Ensures main text does not exceed limit and references are properly placed.

3. Figures/Tables After References Check

• Logic: After locating the references section, scans subsequent pages for figure/table markers and lists page numbers where found.
• Configuration: None (automatic if references are found).
• Note: Uses word boundaries to avoid false positives.

4. Appendix Outside Main Pages Check

• Logic: Scans pages after --main-pages for appendix section headers such as Appendix, Appendix A, or Appendices.
• Configuration: --main-pages (default 10).
• Output: Emits a warning with page numbers where appendix headers are detected outside the main-page window.

5. Style Detection Check

• Logic: Searches the first two pages for style-specific keywords:
  • ACM: "acm" or "association for computing machinery"
  • IEEE: "ieee" or "institute of electrical and electronics engineers"
• Configuration: --style acm or --style ieee (optional). If specified, warns on mismatch. If not, reports detected style.
• Note: Only reports if ACM or IEEE is detected.

6. Email Detection Check

• Logic: Uses regex to search for email patterns on the first page.
• Configuration: None (always checked).
• Regex: [A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}

7. Non-Standard References Format Check

• Logic: Scans reference entries and validates expected bracketed numeric format such as [1] Author, Title, Venue, Year.
• Detects:
  • Bullet-point references (for example, - Author, Title...)
  • Plain numbering without brackets (for example, 1. Author, Title...)
  • Author-year style entries (for example, Smith et al. (2020) ...)
• Output: Emits a warning including detected non-standard format categories.

8. Suspicious Wording Check

• Logic: Searches the entire document for predefined phrases (case-insensitive).
• Configuration: Hardcoded phrases: "our previous paper", "in our previous work".
• Note: Warns for each matching phrase found.

9. Metadata Check

• Logic: Extracts PDF metadata (e.g., author, title) and checks if any fields contain text.
• Configuration: None (always checked).
• Note: Metadata often includes identifying information like author names.

Installation

cd submission-checker
pip install -e .
pip install pypdf

Or use requirements.txt:

pip install -r requirements.txt
pip install -e .

Usage

Check a Single PDF

submission-checker --file paper.pdf --max-pages 12 --main-pages 10 --style ieee

This checks a single PDF with:

• Maximum total pages: 12 (main text + references)
• Maximum main text pages: 10 (references must start after page 10)
• Expected style: IEEE

Output:

Checking file: paper.pdf
Warnings:
 - Number of pages (13) exceeds limit (12).
 - Non-anonymous email detected on page 1.
 - Suspicious wording detected: 'our previous paper'.

Exit code: 0 if no issues, 1 if warnings found.

Scan a Folder of PDFs

Check all PDFs in a directory (recursive):

submission-checker --folder submissions --max-pages 12 --main-pages 10 --style ieee --timeout 30 --csv report.csv

This scans all PDFs in the submissions folder with the same page limits and style check, saves results to report.csv.

Output:

Checking file: paper1.pdf
Checking file: paper2.pdf
Checking file: paper3.pdf
Filename                                      Status     Issues
================================================== ===== ========
paper1.pdf                                    ✓ PASS     0
paper2.pdf                                    ✗ FAIL     2
  - Number of pages (10) exceeds limit (8).
  - Non-anonymous email detected on page 1.
paper3.pdf                                    ✗ FAIL     1
  - Could not extract text from PDF (possibly corrupted, encrypted, or slow to read).

================================================== ===== ========
Summary: 1 passed, 2 failed out of 3 files

For CSV output:

submission-checker --folder submissions --max-pages 8 --style acm --csv report.csv

This generates a CSV file with columns: Filename, Status, Issues.

Exit code: 0 if all passed, 1 if any failed.

Troubleshooting

• "Could not extract text from PDF": The PDF may be scanned (image-only), encrypted, or corrupted. Try re-saving as a text PDF or using OCR tools like ocrmypdf.
• Slow scans on network drives: Increase --timeout (e.g., --timeout 60) or copy files locally first.
• No warnings on expected issues: Ensure the PDF has extractable text. Test with pdftotext file.pdf - to verify.
• Import errors: Install dependencies with pip install -r requirements.txt.

Development

Run tests:

pytest

Build package:

pip install -e .

Filename                                 Status     Issues
======================================================================
paper1.pdf                               ✓ PASS     0
paper2.pdf                               ✗ FAIL     2
  - Number of pages (10) exceeds limit (8).
  - Non-anonymous email detected on page 1.
paper3.pdf                               ✓ PASS     0

======================================================================
Summary: 2 passed, 1 failed out of 3 files

Command-Line Options

• --file <path> – Check a single PDF file
• --folder <path> – Check all PDFs in a folder (recursive search for *.pdf)
• --max-pages <num> – Maximum total pages allowed (optional)
• --min-pages <num> – Minimum total pages required (optional)
• --main-pages <num> – Maximum main text pages (default 10)
• --style {acm,ieee} – Enforce a specific citation style (optional)
• --timeout <num> – Timeout in seconds for PDF text extraction (default: 10)
• --csv <path> – Output results to CSV (folder only)

Note: Provide either --file or --folder, not both.

Examples

# Single file, no style requirement
submission-checker --file paper.pdf --max-pages 10

Single file, ACM style required

submission-checker --file paper.pdf --max-pages 8 --style acm

Batch check IEEE submissions

submission-checker --folder submissions --max-pages 12 --style ieee

Check folder without page limit

submission-checker --folder papers

Running Tests

pytest tests/test_checker.py -q

With verbose output:

pytest tests/test_checker.py -v -s

How It Works

1. Text Extraction – Extracts text from each page of the PDF
2. Content Analysis – Scans for patterns, emails, suspicious phrases
3. Metadata Inspection – Checks PDF metadata for identifying information
4. Style Detection – Analyzes text for ACM/IEEE keywords
5. Reporting – Lists all issues found with line-by-line details

Contributing

We welcome contributions! For any bug reports, feature requests, or suggestions for general improvements to the tool, please raise them on our GitHub Issues page.

General Contributions

For bug fixes and general improvements to the tool (those that benefit all users):

1. Raise an issue first on our GitHub Issues page to discuss the change
2. Fork the repository and create a feature branch (e.g., fix/email-detection, feature/new-check)
3. Submit a pull request to the main branch with a description of your changes and a reference to the issue

For Conference Organizers

If your conference wants to use and customize this tool for your submission review process, follow these steps:

1. Create a Conference Branch

   • Create a new branch with the naming format: CONFERENCENAME-YEAR
   • Example: ICSE-2025, OSDI-2026, NeurIPS-2025

2. Customize for Your Conference

   • Within your conference branch, you can customize the tool to match your specific requirements:
     • Adjust page limits via configuration
     • Add/modify citation style requirements
     • Create custom validation checks specific to your conference
     • Update check thresholds and parameters
   • Make all changes within this branch without submitting pull requests to main

3. Benefits of This Approach

   • Your conference-specific customizations remain isolated and don't affect the main tool
   • You can maintain your branch independently without conflicts with other conferences
   • Easy to update from main when new general improvements are released
   • Clear separation between core tool improvements and conference-specific configurations

4. Updating from Main

   • To incorporate improvements from the main branch into your conference branch:

   git fetch origin
   git rebase origin/main

   • Resolve any conflicts specific to your conference customizations

For questions or if you'd like to establish a conference-specific branch, please open an issue on our GitHub Issues page.