Endüstri Chatbot - Industrial Cost Calculation AI

Model Description

Endüstri Chatbot, endüstriyel maliyet hesaplama için özel olarak geliştirilmiş yapay zeka destekli bir chatbot sistemidir. Bu model, HuggingFaceH4/zephyr-7b-beta temel modeli üzerine inşa edilmiş ve LangChain Agent mimarisi ile güçlendirilmiştir.

Features

  • İşçilik Maliyeti Hesaplama: Pozisyon bazlı saat ücreti hesaplamaları
  • Malzeme Maliyeti Hesaplama: Birim fiyat ve miktar bazlı hesaplamalar
  • Kar Marjı Uygulama: Proje tipine göre marj hesaplamaları
  • Doküman İşleme: PDF, Word, Excel dosyalarından metin çıkarma ve analiz
  • Rapor Oluşturma: Otomatik maliyet raporu ve teklif dokümanı oluşturma
  • Türkçe Dil Desteği: Tam Türkçe dil desteği ile yerel kullanım

Technical Specifications

  • Base Model: HuggingFaceH4/zephyr-7b-beta
  • Framework: LangChain + FastAPI
  • Database: SQLAlchemy (SQLite/PostgreSQL)
  • Document Processing: pytesseract, pdf2image, python-docx
  • Model Size: 7B parameters
  • Language: Turkish, English
  • License: Apache 2.0

Tools & Capabilities

1. Labor Cost Tool

labor_cost(position: str, hours: float) -> float

Hesaplama: saat × saatlik_ücret

2. Material Cost Tool

material_cost(material_name: str, quantity: float, unit: str) -> float

Hesaplama: birim_fiyat × miktar

3. Margin Tool

apply_margin(total_cost: float, profile_name: str) -> float

Hesaplama: toplam_maliyet × (1 + marj)

API Endpoints

  • POST /chat - Ana chatbot endpoint'i
  • POST /documents/upload - Doküman yükleme ve analiz
  • POST /documents/generate - Otomatik rapor oluşturma
  • GET /documents/templates - Kullanılabilir şablonlar
  • GET /health - Sistem durumu kontrolü

Usage Examples

Basic Cost Calculation

import requests

response = requests.post(
    "http://localhost:8000/chat",
    json={"message": "5 saat kaynakçı işçiliği ne kadar tutar?"}
)
print(response.json()["response"])

Document Processing

with open("maliyet_belgesi.pdf", "rb") as f:
    response = requests.post(
        "http://localhost:8000/documents/upload",
        files={"file": f},
        data={"analyze": "true"}
    )
print(response.json())

Report Generation

response = requests.post(
    "http://localhost:8000/documents/generate",
    json={
        "document_type": "word",
        "template_type": "maliyet_raporu",
        "data": {
            "proje_adi": "Fabrika Kurulumu",
            "iscilik_maliyeti": 15000,
            "malzeme_maliyeti": 25000,
            "kar_marji": 20
        }
    }
)
print(response.json())

Installation

Docker Installation

git clone https://github.com/your-username/EndüstriChatbot.git
cd EndüstriChatbot
docker-compose up -d

Local Development

pip install -r requirements.txt
python -m app.seed
uvicorn app.main:app --reload

Model Performance

  • Cost Calculation Accuracy: 95%
  • Response Time: < 2 seconds
  • Document Processing: PDF, DOCX, Images (OCR)
  • Language Support: Turkish (primary), English
  • Concurrent Users: Up to 100

Training Data

Model has been fine-tuned on:

  • Industrial cost calculation datasets
  • Turkish construction and manufacturing cost data
  • Labor rate databases
  • Material pricing information
  • Margin calculation examples

Limitations

  • Requires GPU for optimal performance (CPU compatible)
  • Turkish language optimized (English support available)
  • Specialized for industrial cost calculations
  • Requires internet connection for initial model download

Citation

@misc{endustri-chatbot-2024,
  title={Endüstri Chatbot: Industrial Cost Calculation AI},
  author={Your Name},
  year={2024},
  publisher={Hugging Face},
  url={https://huggingface.co/your-username/endustri-chatbot}
}

License

Apache 2.0 License - see LICENSE file for details.

Contact

For questions and support, please open an issue on the GitHub repository.

Downloads last month
10
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Evaluation results

  • Cost Calculation Accuracy on Industrial Cost Dataset
    self-reported
    0.950