CoolFace
Apppublic

ijsasif/Module-1-Skill-Extractor

sourceHugging Facemitupdated 3mo agoView on Hugging Face
0likes
App README

TalentPulse - Module 1 Skill Extraction API

Backend-only Hugging Face Space for the TalentPulse Module 1 recruitment workflow.

What this Space does

  • —Accepts PDF, DOCX, and TXT CV files
  • —Extracts resume text from the uploaded file
  • —Runs GLiNER skill extraction
  • —Returns categorized skills with confidence scores for the frontend

Endpoints

MethodPathDescription
GET/Health check
GET/healthRuntime and model status
POST/extract-skillsUpload CV file using multipart/form-data with field file
POST/extract-textExtract categorized skills from raw text

Frontend usage

The recruitment page should call:

bash
POST /extract-skills?mode=gliner

The response contains:

  • —categorized: grouped skills by category
  • —skills: flat skill list
  • —total: total extracted skills
  • —meta: model and runtime info

Example response

json
{
  "filename": "resume.pdf",
  "extractor": "gliner",
  "mode": "gliner",
  "total": 4,
  "skills": [
    { "name": "python", "category": "programming language", "score": 0.95 },
    { "name": "sql", "category": "database", "score": 0.91 }
  ],
  "categorized": {
    "programming language": [
      { "name": "python", "score": 0.95 }
    ],
    "database": [
      { "name": "sql", "score": 0.91 }
    ]
  },
  "meta": {
    "text_length": 1820,
    "gliner_model": "urchade/gliner_medium-v2.1",
    "gliner_threshold": 0.35,
    "device": "cpu",
    "supported_extensions": [".docx", ".pdf", ".txt"]
  }
}