How gunalgupta/academic-plagiarism-checker---for-handwritten-submissions works
This is a Streamlit application for checking handwritten assignment submissions for plagiarism. A user uploads a ZIP archive and starts a scan; the app extracts the submissions and invokes the comparison workflow, which combines image-pattern similarity with OCR text embeddings and reports complete or potential matches. An optional Level 4 branch asks OpenAI to assess paraphrasing. The graph reflects the shown calls and README-described capabilities; it does not infer additional wiring for unsampled behavior.
Main entry points: model.py Pairwise scanner · streamlit.py Streamlit app · ocr.py OCR extraction · openai_model.py Paraphrase check
Components 13
Upload interface
- Submission ZIP
- Streamlit app
streamlit.py - Archive extraction
streamlit.py - Submission files
Plagiarism analysis
- Pairwise scanner
model.py - HOG patterns
udp.py - Text similarity
model.py - Match results
model.py
External services
- OCR extraction
ocr.py - OCR API
- Paraphrase check
openai_model.py - OpenAI service
Other components
- Instructor
Connections 14
- Instructor → Streamlit app: uploads and starts
README.md - Streamlit app → Submission ZIP: accepts ZIP
streamlit.py - Streamlit app → Archive extraction: validates and extracts
streamlit.py - Archive extraction → Submission files: produces files
streamlit.py - Streamlit app → Pairwise scanner: invokes scan
streamlit.py - Pairwise scanner → Submission files: reads files
model.py - Pairwise scanner → OCR extraction: extracts text
model.py - OCR extraction → OCR API: sends image
ocr.py - Pairwise scanner → HOG patterns: compares patterns
model.py - Pairwise scanner → Text similarity: scores text embeddings
model.py - Pairwise scanner → Paraphrase check: optionally checks paraphrases
model.py - Paraphrase check → OpenAI service: requests comparison
openai_model.py - Pairwise scanner → Match results: reports match pairs
model.py - Match results → Streamlit app: displays results
model.py