Omni Medical Suite

Arabic Medical Document Intelligence Platform
^{OCR · Handwriting Recognition · Spell Correction · NER · Training}

GitHub · Demo (Beta) · Architecture · Roadmap · Contributing

What It Does

Omni Medical Suite extracts, corrects, and structures Arabic text from medical documents and handwritten prescriptions. It now combines a routed multi-engine OCR stack with AI-powered spell correction, field-aware deduplication, semantic search, and medical context protection.

  Prescription Image
         |
    [Scanner Fixer]         deskew, crop, CLAHE, denoise
         |
    [OCR Router]            EasyOCR + PaddleOCR + Tesseract + TrOCR + QARI + Nougat + Qwen handwritten OCR
         |
    [Arabic NLP]            RTL handling, mixed-language detection
         |
    [Spell Checker]         Dictionary + Jais LLM + medical vocabulary protection
         |
    [NER Extractor]         Drug names, dosages, dates, diagnoses
         |
    Structured Output       JSON / CSV / XLSX / PDF

Key Features

Feature	Description
Routed OCR Stack	EngineRouter chooses among EasyOCR, PaddleOCR, Tesseract, TrOCR, QARI, Nougat, and Arabic-handwritten-OCR (Qwen)
Arabic Handwriting Recognition	Qwen handwritten OCR first, with TrOCR / QARI fallback
Medical Spell Checker	HybridSpellChecker protects medical terms from false corrections
Jais LLM Proofreading	AI-powered Arabic proofreading via Jais model
Image Preprocessing	Deskew, CLAHE contrast, line removal, perspective correction
Named Entity Recognition	Extracts drugs, dosages, dates, and diagnoses
Advanced Review App	Gradio app with Compare + Search + Review tabs
Field-Aware Deduplication	Patient-sensitive weighting prevents same-template / different-patient false positives
HITL Feedback Loop	Human corrections feed back into training data
Multi-Page PDF + Tables	Batch processing with table extraction
Desktop + Web + API	PyQt6 desktop app, Gradio web UI, FastAPI REST API
Continuous Retraining	Weekly model improvement from accumulated corrections

🆕 What's New in v1.1.0 (stable)

The v1.1.0 hardening sprint consolidates 16 commits across three phases (P0 + P1 + P2) plus 4 post-rc1 AppImage build fixes. It adds 174 passing tests (163 unit + 11 AppImage smoke), structured decision logging, an AppImage build pipeline, and a multi-platform CI matrix.

Highlights

Phase	Feature	Impact
P0	Lazy OCR factories (`get_paddle_ocr()`, `has_tesseract()`, ...)	First-use construction + cached failures → no import-time crashes
P0	Structured decision log (`app/core/decision_log.py`)	Every RTL/dedup/field/HF decision emits JSON line with reasons + duration
P0	HF dataset staging queue	`save_to_hf()` is now O(1); batched push at 25 rows (configurable)
P0	Translation service extracted	`transformers + torch` loaded lazily inside `load_translator()`
P0	Scanner fixer Gradio integration	Manual crop (x/y/w/h) + advanced edges + ZIP save
P1	Field extractor hardening	Multi-line values, bilingual labels, confidence scores, safe signatures
P1	Benchmark reporter	`to_csv()`, `to_json()`, `aggregate_metrics()` with percentiles
P1	Decision instrumentation	RTL, dedup, field_extractor all emit `log_decision()`
P1	Git LFS coverage	`.gitattributes` 50+ patterns across 10 categories
P2	AppImage build pipeline	`bash build_appimage.sh --version-from-git --smoke-test`
P2	CI matrix	Python 3.10/3.11/3.12 × Ubuntu + Manjaro/Arch container + HF smoke + Colab smoke + LFS audit
P2	LFS migration plan	Staged approach (no forced history rewrite on `main`)
stable	AppImage pre-built binary (177 MB)	Direct download from GitHub Release — no build step required
stable	`ARCH=x86_64` + `.desktop`/icon at AppDir root	`appimagetool` build fully green in CI
stable	Backup branches + RC1 → stable promotion	`backup/before-v1.1.0-stable` for rollback safety

Migration notes (v1.0.0 → v1.1.0)

Breaking changes: None. All public APIs preserved via PEP 562 __getattr__.
Behavioral changes:
1. save_to_hf() now appends to local JSONL by default; flushes batch of 25 rows. Set OMNI_HF_FLUSH_THRESHOLD=1 for legacy per-save behavior.
2. OCR engines construct on first use (not at import). Failed constructions are cached (no retry storm).
3. Decision log emits to stderr by default. Attach a JSON-lines file handler to ship to your log aggregator.
New env vars: OMNI_HF_QUEUE_DIR, OMNI_HF_FLUSH_THRESHOLD, OMNI_APPIMAGE_OFFSCREEN, APPIMAGETOOL_SIGN_KEY, MEDICAL_DOC_APPIMAGE.

📖 Full release notes: RELEASE_NOTES_v1.1.0.md 📋 Release candidate checklist: RELEASE_CANDIDATE_CHECKLIST.md 🔄 Previous RC: RELEASE_NOTES_v1.1.0-rc1.md

Quick Start

Option 1: Official Gradio HITL UI

git clone --recursive https://github.com/DrAbdulmalek/omni-medical-suite.git
cd omni-medical-suite
# If you cloned without --recursive:
# git submodule update --init
pip install -r requirements/gradio.txt
python app/gradio_full_hitl.py

Open http://localhost:7860

Option 1b: Experimental Review UI (Compare/Search/Review tabs)

python app/advanced_review_app.py

Note: This is an experimental app. It does not yet support image upload, Jais proofreading, HF Dataset save, medical translation, or dictionary update. Use gradio_full_hitl.py for production.

Option 2: Desktop App (PyQt6)

git clone --recursive https://github.com/DrAbdulmalek/omni-medical-suite.git
cd omni-medical-suite
pip install PyQt6 pytesseract  # PaddleOCR/EasyOCR optional
python desktop/omni_medical_desktop.py

Features:

OCR Scanner tab — upload image, run Tesseract/PaddleOCR/EasyOCR/Ensemble, view results
Jais LLM Proofread checkbox — enable AI proofreading when GPU is available
Medical NER panel — extracted entities (drugs, dosages, diagnoses) displayed after Jais run
Text Editor tab — RTL Arabic editor with Find & Replace
Dictionary tab — live-search medical terms from medical_terms.json
Settings tab — engine selection, language, confidence threshold, dark mode

Option 3: Docker (Full Stack)

git clone --recursive https://github.com/DrAbdulmalek/omni-medical-suite.git
cd omni-medical-suite
cp .env.example .env   # Edit with your settings
docker-compose up -d

Option 4: Docker Lite (OCR Only)

docker-compose -f docker-compose.lite.yml up -d

Option 5: Desktop AppImage (Manjaro / Linux x86_64) — Stable in v1.1.0

Portable single-file AppImage — no install required, runs on Manjaro/KDE Plasma 6 (Wayland) and most modern Linux distros.

⬇️ Pre-built binary (recommended): Download MedicalDocProcessor-v1.1.0-x86_64.AppImage (177 MB) directly from the GitHub Release v1.1.0:

# Download + verify + run
wget https://github.com/DrAbdulmalek/omni-medical-suite/releases/download/v1.1.0/MedicalDocProcessor-v1.1.0-x86_64.AppImage
wget https://github.com/DrAbdulmalek/omni-medical-suite/releases/download/v1.1.0/MedicalDocProcessor-v1.1.0-x86_64.AppImage.sha256
sha256sum -c MedicalDocProcessor-v1.1.0-x86_64.AppImage.sha256
chmod +x MedicalDocProcessor-v1.1.0-x86_64.AppImage
./MedicalDocProcessor-v1.1.0-x86_64.AppImage

🔧 Build from source (advanced):

git clone https://github.com/DrAbdulmalek/omni-medical-suite.git
cd omni-medical-suite
git checkout v1.1.0   # stable release

# Install scanner_fixer editable + desktop deps
pip install -e packages/scanner_fixer
pip install -r packages/desktop/requirements.txt
pip install pyinstaller

# Build the AppImage (auto-downloads appimagetool)
cd packages/desktop
bash build_appimage.sh --version-from-git --smoke-test

# Run it
chmod +x MedicalDocProcessor-*.AppImage
./MedicalDocProcessor-*.AppImage

# Verify integrity
sha256sum -c MedicalDocProcessor-*.AppImage.sha256

📖 Full Manjaro build guide: docs/APPIIMAGE_MANJARO.md

Pre-built AppImage is also published as a CI artifact on every push to main and on tags v*. Download from the Actions tab → appimage-build.yml → latest run.

Live Demo

HuggingFace Space — No installation needed.

Monorepo Structure

omni-medical-suite/
├── app/                         # Gradio review applications
├── src/                         # Core library
│   ├── ocr/                     #   OCR engine routing, RTL fixing, extraction, dedup
│   ├── ner/                     #   Named Entity Recognition (Jais)
│   ├── llm/                     #   LLM integration (Jais, Ollama)
│   ├── layout/                  #   Document layout analysis
│   └── benchmarks/              #   Benchmarking utilities
├── desktop/                     # PyQt6 desktop application
├── apps/                        # Standalone applications
│   ├── ocr-pipeline/            #   Full end-to-end pipeline
│   ├── handwriting-demo/        #   Handwriting recognition demo
│   ├── ocr-demo/                #   Multi-engine comparison
│   ├── trainer-ui/              #   Model training interface
│   └── collector/               #   HITL data collector
├── packages/                    # Reusable Python packages (31)
│   ├── scanner_fixer/           #   Pre-OCR image normalization
│   ├── vision/                  #   Computer vision & HTR
│   ├── nlp/                     #   Spell checking, translation, RTL
│   ├── medical/                 #   Medical dictionaries & domain logic
│   ├── core/                    #   Shared utilities, DB, auth
│   ├── ocr_postprocess/         #   OCR post-processing & fusion
│   ├── omnifile/                #   File processing & export
│   ├── training/                #   Model training framework
│   └── ...                      #   ai-fuel, bilingual, export, etc.
├── config/                      # Model configs, Prometheus, Grafana, Nginx
├── tests/                       # 40+ test files
├── docs/                        # Architecture, guides, ADRs
├── hf-space/                    # HuggingFace Spaces deployment
├── infra/                       # Docker & Kubernetes configs
└── .github/workflows/           # CI/CD pipelines

Tech Stack

Layer	Technologies
OCR	PaddleOCR, Tesseract, EasyOCR, TrOCR, Surya
NLP / LLM	Transformers, Jais, Custom HybridSpellChecker
Vision	OpenCV, CLAHE, Scikit-image
Backend	Python 3.10+, FastAPI, Celery, PostgreSQL
Frontend	Gradio 4.x, PyQt6, Next.js
Infrastructure	Docker, Redis, Qdrant, Nginx
Monitoring	Prometheus, Grafana, Tempo (tracing)
CI/CD	GitHub Actions, HuggingFace Spaces
Deployment	Docker Compose, Kubernetes, HF Spaces

Configuration

Copy the example environment file and customize:

cp .env.example .env

Key variables:

Variable	Description	Default
`POSTGRES_PASSWORD`	PostgreSQL password	(required)
`SECRET_KEY`	FastAPI secret key	(required)
`ENABLE_LLM`	Enable Jais proofreader	`false`
`HF_TOKEN`	HuggingFace token for dataset upload	(optional)

See .env.example for the full list.

Documentation

📄 PDF OCR Processor

نظام متكامل لاستخراج النص والمسارد من ملفات PDF الطبية باستخدام OCR متقدم مع ضبط تلقائي.

الميزات

ضبط تلقائي لمعاملات OCR (PSM + DPI) — يختار أفضل إعداد
معالجة مسبقة للصور عبر scanner_fixer (deskew, crop, normalize, enhance)
استخراج المسارد الثنائية اللغة (عربي = إنجليزي) تلقائياً
حفظ متعدد (TXT, CSV, JSON) + مسرد موحد
تكامل مع نظام التسجيل المتقدم (advanced_logger)

الاستخدام

# تثبيت الحزم
sudo pacman -S tesseract tesseract-data-ara tesseract-data-eng poppler
pip install -r requirements/ml.txt

# معالجة مجلد PDF
python3 scripts/pdf_ocr_processor.py

# ملف واحد مع ضبط تلقائي
python3 scripts/pdf_ocr_processor.py --input report.pdf --auto-tune

# باستخدام سكربت التشغيل
./scripts/run_ocr.sh [GITHUB_TOKEN] [INPUT_PATH]

المخرجات

~/glossaries_output/
├── <filename>.txt          # النص الكامل المستخرج
├── <filename>.csv          # المسارد (term_arabic, term_english)
├── <filename>.json         # النتيجة الكاملة (JSON)
├── combined_glossary.csv   # مسارد موحدة
├── combined_glossary.json  # مسارد موحدة (JSON)
└── OCR_PROCESSING_LOG.md   # سجل المعالجة

📖 توثيق مفصل

Documentation

Document	Content
docs/ARCHITECTURE.md	System architecture, data flow, package map
docs/PIPELINE.md	Full OCR processing pipeline
docs/MODES.md	Operation modes (lite, standard, full)
docs/DEPLOYMENT.md	Deployment guide (Docker, K8s, HF)
docs/ROADMAP.md	Development roadmap & milestones
docs/ROADMAP_2026_Q3.md	July 2026 implementation plan for routing, review UI, and dedup
docs/CHANGELOG.md	Version changelog & repository history

Deployment

Architecture Overview

                  ┌─────────────────────┐
                  │   GitHub (main)     │
                  └────────┬────────────┘
                           │ push (paths filter)
                  ┌────────▼────────────┐
                  │ GitHub Actions       │
                  │ deploy-to-hf.yml    │
                  └────────┬────────────┘
                           │ git push
                  ┌────────▼────────────┐
                  │ HF Spaces           │
                  │ omni-medical-ocr    │
                  │ (Docker + Gradio)   │
                  └─────────────────────┘

Docker Images

Image	Dockerfile	Description	Size
`omni-ocr`	`Dockerfile.gradio`	Multi-stage Gradio HITL app	~1.8 GB
`omni-api`	`Dockerfile.api`	FastAPI backend	~1.2 GB

Quick Start with Docker Compose

# Gradio only (lightweight — no databases)
docker-compose up gradio

# Full stack (Gradio + API + PostgreSQL + Redis + Qdrant)
docker-compose --profile infra up

CI/CD Pipeline

Step	Trigger	Action
Path filter	Push to `main` with changes in `hf-space/`, `Dockerfile.gradio`, `packages/`	Triggers deploy
Sync files	Copies `app.py`, `src/`, `packages/`, `config/` from monorepo to HF Space	Preserves HF Dockerfile
Push	Auto-commits and pushes to `DrAbdulmalek/omni-medical-ocr`	Space rebuilds

Configuration Variables

Variable	Default	Description
`POSTGRES_PASSWORD`	(REQUIRED — no default)	PostgreSQL password (set in `.env`; compose refuses to start without it)
`SECRET_KEY`	(dev placeholder)	API auth secret
`ENABLE_LLM`	`false`	Enable Jais proofreader (requires GPU)
`HF_TOKEN`	—	HuggingFace token for dataset upload

HuggingFace Spaces

The Gradio app auto-deploys to omni-medical-ocr on every push to main that modifies hf-space/ or Dockerfile.gradio.

Live Demo: https://drabdulmalek-omni-medical-ocr.hf.space

HF Space Features

Feature	Description
OCR Processing	PaddleOCR + Tesseract ensemble, preprocessing, spell check, NER
Translation	Arabic ↔ English ↔ German (MarianMT, lazy-loaded)
Accuracy Metrics	CER/WER calculator with jiwer verification
Save to Dataset	Upload corrections to `DrAbdulmalek/arabic-medical-ocr-corrections`

Manual Deploy to HF Spaces

# Clone the Space
git clone https://huggingface.co/spaces/DrAbdulmalek/omni-medical-ocr
cd omni-medical-ocr

# Copy deployment files from monorepo
cp -r ../omni-medical-suite/hf-space/* .

# Push (triggers Docker rebuild on HF)
git add . && git commit -m "Update from monorepo" && git push

Enable Auto-Deploy (CI/CD)

Go to GitHub repo → Settings → Secrets and variables → Actions
Add HF_TOKEN with your HuggingFace access token
Every push to main that changes hf-space/ will auto-deploy

Development

# Install dev dependencies
pip install -r requirements.txt
pip install -r requirements-dev.txt

# Run tests
pytest tests/ --ignore=tests/loadtest -q

# Run linter
ruff check src/ packages/ app/

# Run type checker
mypy . --ignore-missing-imports

# Local development server
make dev

Monitoring & Maintenance

Health Checks

Endpoint	Description	Expected Response
`/health`	Full health check	`{"status": "healthy"}`
`/health/liveness`	Liveness probe	`{"status": "alive"}`
`/health/readiness`	Readiness probe	`{"status": "ready"}`

Monitoring Stack

Start with:

docker-compose -f docker-compose.yml -f infra/monitoring/docker-compose.monitoring.yml up -d

Service	URL	Description
Prometheus	http://localhost:9090	Metrics collection (15s scrape)
Grafana	http://localhost:3000	Dashboards (admin / `${GRAFANA_ADMIN_PASSWORD}`)
Node Exporter	http://localhost:9100	System metrics

Backup Strategy

Frequency: Daily at 2:00 AM (configurable)
Retention: 30 days
Location: /app/backups/
Components: Database (pg_dump), Redis (RDB snapshot), Critical files
Manual trigger: python -m scripts.backup

Update Mechanism

Check interval: Every hour
Notification: Logged via structured logging
Auto-update: Disabled (manual approval required)
Manual check: python -m scripts.update_checker

Documentation

RUNBOOK: docs/RUNBOOK.md — Operations guide
Maintenance Log: docs/MAINTENANCE.md — Schedule & benchmarks
API Docs: /docs — Swagger UI
Architecture: docs/ARCHITECTURE.md

Roadmap

v1.0 — Monorepo consolidation (6 repos merged)
v1.1 — CI/CD + Docker + HF Space deployment
v1.2 — Testing infrastructure + ruff + mypy + PyQt6 desktop app
v1.3 — Monitoring + Maintenance (Prometheus, Grafana, Health Checks, Backup)
v1.4 — Multi-page PDF + table extraction
v1.5 — Weekly auto-retraining pipeline
v2.0 — Real-time collaboration + RBAC + audit log

See docs/ROADMAP.md for details.

Requirements

Python 3.10+
8 GB RAM minimum (16 GB recommended)
CUDA optional (GPU acceleration for TrOCR / Jais)
Docker + Docker Compose (for containerized deployment)

License

Standalone Tools

The following Gradio-based tools serve different use cases and are independent from the main OCR pipeline. They are not part of the canonical app but are available for specific workflows.

Scanner Fixer Pro (`desktop/gradio_scanner_app.py`)

Web-based alternative to the Tkinter desktop scanner app. Provides the same image enhancement pipeline via the browser.

Features:

Shadow removal, deskew, perspective correction
Denoising (fastNlMeans), CLAHE contrast enhancement, auto-crop
Per-option toggles and processing metrics

Run:

python desktop/gradio_scanner_app.py

Medical Data Analysis Platform (`apps/handwriting-demo/hf-deploy/app/gradio_app.py`)

Interactive HuggingFace Spaces deployment with clinical AI features.

Features:

OCR Correction — annotated bounding boxes, editable word crops, ground truth collection
Document Parser — extract structured text from PDF/DOCX/PPTX
Medical Analysis — extract vitals, medications, diagnoses from free-form text
Clinical Q&A — evidence-based clinical questions with citations

Telegram Content Forwarder (`tools/ops/telegram_forwarder/app.py`)

Gradio UI for the Telegram content forwarding tool (ops utility).

Features:

Download-Upload technique to bypass "Restrict Saving Content" on protected channels
Real-time progress tracking, session string export for HF Spaces
Configurable delay, media/text filtering, reverse order
Rate-limit warnings and cancel support

Archived Prototypes

Formerly PENDING Gradio files with unique but non-production code have been moved to research/prototypes/. See that directory for details.

Legacy Translation Corrector Notes

The file packages/file_processor/legacy/translation_corrector/app.py (977 lines) contains a TranslationRule-based Arabic translation post-correction system with 13 unique regex patterns not present in packages/medical/tmx_processor.py. The tmx_processor handles TMX file import/export and has only 3 language-detection regexes — it is a different tool entirely.

Unique regex rules in the legacy file:

Pattern	Purpose
`comma_spacing`	Normalize comma whitespace
`arabic_comma`	Normalize Arabic comma (،) spacing
`waw_conjunction`	Fix waw after punctuation
`number_spacing`	Merge split digits
`number_comma`	Handle decimal commas in numbers
`passive_by`	Detect English passive voice (was/were/by)
`passive_simple`	Detect English passive (was/is/are + -ed)
`tanween_alif`	Detect incorrect tanween alif
`redundant_ba`	Remove redundant "بواسطة"
`redundant_waw`	Fix redundant waw after certain words
`word_repeat`	Remove consecutive duplicate words
`extra_spaces`	Collapse multiple spaces
`space_before_punct`	Remove space before Arabic punctuation

These rules are preserved in the legacy file for future reference.

_{Built with passion for Arabic medical NLP}

Name		Name	Last commit message	Last commit date
.github		.github
app		app
apps		apps
benchmarks/results		benchmarks/results
config		config
data		data
deploy		deploy
desktop		desktop
docker/scanner-fixer		docker/scanner-fixer
docs		docs
evaluation		evaluation
hf-space		hf-space
hf-spaces/handwriting-trainer		hf-spaces/handwriting-trainer
infra		infra
labs/omniparse_study		labs/omniparse_study
legacy		legacy
migrations		migrations
mobile		mobile
model		model
notebooks		notebooks
omni_medical_suite		omni_medical_suite
packages		packages
requirements		requirements
research/prototypes		research/prototypes
samples		samples
scripts		scripts
src		src
tests		tests
tools		tools
training-data		training-data
training		training
.dockerignore		.dockerignore
.env.api.example		.env.api.example
.env.example		.env.example
.env.web.example		.env.web.example
.gitattributes		.gitattributes
.gitignore		.gitignore
.gitmodules		.gitmodules
.pre-commit-config.yaml		.pre-commit-config.yaml
.secrets.baseline		.secrets.baseline
CONTRIBUTING.md		CONTRIBUTING.md
DICTIONARY_INVENTORY.md		DICTIONARY_INVENTORY.md
DOCS_CONSOLIDATION_PLAN.md		DOCS_CONSOLIDATION_PLAN.md
Dockerfile		Dockerfile
Dockerfile.api		Dockerfile.api
Dockerfile.dev		Dockerfile.dev
Dockerfile.gradio		Dockerfile.gradio
Dockerfile.lb		Dockerfile.lb
Dockerfile.mobile		Dockerfile.mobile
Dockerfile.redis		Dockerfile.redis
Dockerfile.review		Dockerfile.review
Dockerfile.ws		Dockerfile.ws
GITHUB_ACTIONS_GUIDE.md		GITHUB_ACTIONS_GUIDE.md
GRADIO_APPS_DECISION.md		GRADIO_APPS_DECISION.md
GRADIO_HITL_CHANGES_REVIEW.md		GRADIO_HITL_CHANGES_REVIEW.md
LEGACY_REVIEW_FINAL.md		LEGACY_REVIEW_FINAL.md
LICENSE		LICENSE
MERGE_HISTORY.md		MERGE_HISTORY.md
MIGRATION_NOTES_v1.1.0-rc.md		MIGRATION_NOTES_v1.1.0-rc.md
MISSING_SOURCE_FILES.md		MISSING_SOURCE_FILES.md
Makefile		Makefile
OPEN_ISSUES.md		OPEN_ISSUES.md
PARTIAL_DUPLICATES_DECISION_QUEUE.md		PARTIAL_DUPLICATES_DECISION_QUEUE.md
PII_AUDIT.md		PII_AUDIT.md
README.md		README.md
RELEASE_CANDIDATE_CHECKLIST.md		RELEASE_CANDIDATE_CHECKLIST.md
RELEASE_NOTES_v1.0.0-rc1.md		RELEASE_NOTES_v1.0.0-rc1.md
RELEASE_NOTES_v1.1.0-rc1.md		RELEASE_NOTES_v1.1.0-rc1.md
RELEASE_NOTES_v1.1.0.md		RELEASE_NOTES_v1.1.0.md
REPOSITORY_AUDIT.md		REPOSITORY_AUDIT.md
ROADMAP.md		ROADMAP.md
SECURITY.md		SECURITY.md
VERIFICATION_LOG.md		VERIFICATION_LOG.md
add_prescription.py		add_prescription.py
alembic.ini		alembic.ini
choose-tier.sh		choose-tier.sh
docker-compose.dev.yml		docker-compose.dev.yml
docker-compose.lite.yml		docker-compose.lite.yml
docker-compose.medical-infra.yml		docker-compose.medical-infra.yml
docker-compose.mobile.yml		docker-compose.mobile.yml
docker-compose.prod.yml		docker-compose.prod.yml
docker-compose.standard.yml		docker-compose.standard.yml
docker-compose.yml		docker-compose.yml
entrypoint.sh		entrypoint.sh
env.example.production		env.example.production
init.sql		init.sql
main_pipeline.py		main_pipeline.py
medical_terms.json		medical_terms.json
mypy.ini		mypy.ini
package.json		package.json
project_context_part02.txt		project_context_part02.txt
pyproject.toml		pyproject.toml
requirements-dev.txt		requirements-dev.txt
requirements-scanner.txt		requirements-scanner.txt
requirements.txt		requirements.txt
ruff.toml		ruff.toml
turbo.json		turbo.json
worklog.md		worklog.md

Latest commit

History

Folders and files

Repository files navigation

Omni Medical Suite

What It Does

Key Features

🆕 What's New in v1.1.0 (stable)

Highlights

Migration notes (v1.0.0 → v1.1.0)

Quick Start

Option 1: Official Gradio HITL UI

Option 1b: Experimental Review UI (Compare/Search/Review tabs)

Option 2: Desktop App (PyQt6)

Option 3: Docker (Full Stack)

Option 4: Docker Lite (OCR Only)

Option 5: Desktop AppImage (Manjaro / Linux x86_64) — Stable in v1.1.0

Live Demo

Monorepo Structure

Tech Stack

Configuration

Documentation

📄 PDF OCR Processor

الميزات

الاستخدام

المخرجات

Documentation

Deployment

Architecture Overview

Docker Images

Quick Start with Docker Compose

CI/CD Pipeline

Configuration Variables

HuggingFace Spaces

HF Space Features

Manual Deploy to HF Spaces

Enable Auto-Deploy (CI/CD)

Development

Monitoring & Maintenance

Health Checks

Monitoring Stack

Backup Strategy

Update Mechanism

Documentation

Roadmap

Requirements

License

Standalone Tools

Scanner Fixer Pro (desktop/gradio_scanner_app.py)

Medical Data Analysis Platform (apps/handwriting-demo/hf-deploy/app/gradio_app.py)

Telegram Content Forwarder (tools/ops/telegram_forwarder/app.py)

Archived Prototypes

Legacy Translation Corrector Notes

About

Topics

Resources

Contributing

Security policy

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages

Scanner Fixer Pro (`desktop/gradio_scanner_app.py`)

Medical Data Analysis Platform (`apps/handwriting-demo/hf-deploy/app/gradio_app.py`)

Telegram Content Forwarder (`tools/ops/telegram_forwarder/app.py`)