Named Entity Recognition with Docker

A Docker-powered service for named entity extraction from text or PDF files.

---

This repository provides a Docker-powered service for Named Entity Recognition (NER), enabling the extraction of specific entities from text or PDF files. The service enables extraction of various entities with the help of pdf-document-layout-analysis, a service that segments documents with high accuracy.

Project Links:

• GitHub: NER-in-docker
• HuggingFace: NER-in-docker

Quick Start

Clone the service:

git clone https://github.com/huridocs/NER-in-docker
cd NER-in-docker

Run the service:

• With GPU support:

  make start
  

• Without GPU support:

  make start_no_gpu
  

The service includes both:

• REST API: Available at http://localhost:5070
• Gradio UI: Available at http://localhost:7860

---

Gradio Web UI

The project includes a user-friendly web interface built with Gradio, accessible at http://localhost:7860 when the service is running.

Features

Text Extraction Tab

• Enter any text to extract named entities
• Supports multiple languages (English, Spanish, German, French)
• Visual highlighting of entities with color-coded types
• JSON response viewer for detailed analysis

PDF Extraction Tab

• Upload PDF documents for entity extraction
• Download annotated PDF with highlighted entities
• Visual display of extracted entities grouped by type
• Fast mode option for quicker processing
• JSON response viewer

Entity Types and Colors

• Person (PER): Blue
• Organization (ORG): Red
• Location (LOC): Green
• Date (DAT): Orange
• Law (LAW): Purple
• Document Code (DOC): Teal
• Reference (REF): Dark Orange

---

API Usage

The service exposes a FastAPI REST API. By default, it runs on http://localhost:8000.

Endpoints

1. / (POST)

Extract named entities from text or PDF.

Parameters:

• namespace (str, optional): Namespace for storing/retrieving entities in SQLite.
• identifier (str, optional): Source identifier for the text.
• text (str, optional): Text to analyze (if no file is provided).
• file (PDF, optional): PDF file to analyze (multipart/form-data).
• fast (bool, optional): Use fast PDF segmentation (default: False).

Example (Text):

curl -X POST http://localhost:8000/ \
  -F "text=Your text here" \
  -F "namespace=my_namespace"

Example (PDF):

curl -X POST http://localhost:8000/ \
  -F "file=@/path/to/file.pdf" \
  -F "namespace=my_namespace" \
  -F "fast=true"

2. /visualize (POST)

Extract named entities from a PDF and return an annotated PDF with visual highlights.

Parameters:

• file (PDF, required): PDF file to analyze (multipart/form-data).
• fast (bool, optional): Use fast PDF segmentation (default: False).

Returns: An annotated PDF file with colored rectangles highlighting each detected entity. Each entity type has a distinct color:

• Person (PER): Blue
• Organization (ORG): Red
• Location (LOC): Green
• Date (DAT): Orange
• Law (LAW): Purple
• Document Code (DOC): Teal
• Reference (REF): Dark Orange

Example:

curl -X POST http://localhost:8000/visualize \
  -F "file=@/path/to/file.pdf" \
  -F "fast=true" \
  --output annotated_file.pdf

3. /delete_namespace (POST)

Delete all entities for a given namespace.

Parameters:

• namespace (str, required): Namespace to delete.

Example:

curl -X POST http://localhost:8000/delete_namespace \
  -F "namespace=my_namespace"

4. / (GET)

Returns Python version info (for health check).

---

Response Structure

The main endpoint (/, POST) returns a JSON object:

{
  "entities": [
    {
      "text": "Entity text",
      "type": "EntityType",
      "source_id": "Source identifier",
      "character_start": 123,
      "character_end": 130,
      "relevance_percentage": 98.5,
      "group_name": "GroupName",
      "segment": {
        "page_number": 1,
        "segment_number": 0,
        ...
      }
    },
    ...
  ],
  "groups": [
    {
      "name": "GroupName",
      "entities": [
        {
          "text": "Entity text",
          "index": 0,
          ...
        },
        ...
      ]
    },
    ...
  ]
}

• entities: List of extracted named entities with metadata.
• groups: List of entity groups, each containing related entities.

---

Notes

• If namespace is provided, entities are stored and reused for reference extraction.
• If both text and file are provided, only the file is processed.
• The service supports both text and PDF input.

For more details, see the source code and API models in src/drivers/rest/response_entities/NamedEntitiesResponse.py.