Support loading multiple mapping files

Change-Id: I3c6caaa3c4c3434dfacfb842f0407250b5e980f0
5 files changed
tree: b292d00e8b2ba1201d6e0c4628f14994288626ac
  1. ast/
  2. cmd/
  3. config/
  4. mapper/
  5. matcher/
  6. parser/
  7. .gitignore
  8. go.mod
  9. go.sum
  10. LICENSE
  11. Makefile
  12. README.md
README.md

KoralPipe-TermMapper

A KorAP service using the KoralPipe mechanism to rewrite terms in queries and responses between different annotations.

Overview

KoralPipe-TermMapper is a tool for transforming linguistic annotations between different annotation schemes. It allows you to define mapping rules in YAML configuration files and apply these mappings to JSON-encoded linguistic annotations.

Installation

go get github.com/KorAP/KoralPipe-TermMapper

Usage

termmapper -c config.yaml -m extra-mapper1.yaml -m extra-mapper2.yaml

Command Line Options

  • --config or -c: YAML configuration file containing mapping directives and global settings (optional)
  • --mappings or -m: Individual YAML mapping files to load (can be used multiple times, optional)
  • --port or -p: Port to listen on (default: 8080)
  • --log-level or -l: Log level (debug, info, warn, error) (default: info)
  • --help or -h: Show help message

Note: At least one mapping source must be provided

Configuration

KoralPipe-TermMapper supports loading configuration from multiple sources:

  1. Main Configuration File (-c): Contains global settings (SDK, server endpoints) and optional mapping lists
  2. Individual Mapping Files (-m): Contains single mapping lists, can be specified multiple times

The main configuration provides global settings, and all mapping lists from both sources are combined. Duplicate mapping IDs across all sources will result in an error.

Configurations can contain global settings and mapping lists (used with the -c flag):

# Optional: Custom SDK endpoint for Kalamar plugin integration
sdk: "https://custom.example.com/js/korap-plugin.js"

# Optional: Custom server endpoint for Kalamar plugin integration  
server: "https://custom.example.com/"

# Optional: Mapping lists (same format as individual mapping files)
lists:
  - id: mapping-list-id
    foundryA: source-foundry
    layerA: source-layer
    foundryB: target-foundry
    layerB: target-layer
    mappings:
      - "[pattern1] <> [replacement1]"
      - "[pattern2] <> [replacement2]"

Map files contain a single mapping list (used with the -m flag):

id: mapping-list-id
foundryA: source-foundry
layerA: source-layer
foundryB: target-foundry
layerB: target-layer
mappings:
  - "[pattern1] <> [replacement1]"
  - "[pattern2] <> [replacement2]"

The sdk and server fields in the main configuration file are optional and override the default endpoints used for Kalamar plugin integration:

  • sdk: Custom SDK JavaScript file URL (default: https://korap.ids-mannheim.de/js/korap-plugin-latest.js)
  • server: Custom server endpoint URL (default: https://korap.ids-mannheim.de/)

These values are applied during configuration parsing and affect the HTML plugin page served at the root endpoint (/). When using only individual mapping files (-m flags), default values are used.

Mapping Rules

Each mapping rule consists of two patterns separated by <>. The patterns can be:

  • Simple terms: [key] or [foundry/layer=key] or [foundry/layer=key:value]
  • Complex terms with AND/OR relations: [term1 & term2] or [term1 | term2] or [term1 | (term2 & term3)]

API Endpoints

POST /:map/query

Transform a JSON object using the specified mapping list.

Parameters:

  • :map: ID of the mapping list to use
  • dir (query): Direction of transformation (atob or btoa, default: atob)
  • foundryA (query): Override default foundryA from mapping list
  • foundryB (query): Override default foundryB from mapping list
  • layerA (query): Override default layerA from mapping list
  • layerB (query): Override default layerB from mapping list

Request body: JSON object to transform

Example request:

POST /opennlp-mapper/query?dir=atob&foundryB=custom HTTP/1.1
Content-Type: application/json

{
  "@type": "koral:token",
  "wrap": {
    "@type": "koral:term",
    "foundry": "opennlp",
    "key": "PIDAT",
    "layer": "p",
    "match": "match:eq"
  }
}

Example response:

{
  "@type": "koral:token",
  "wrap": {
    "@type": "koral:termGroup",
    "operands": [
      {
        "@type": "koral:term",
        "foundry": "custom",
        "key": "PIDAT",
        "layer": "p",
        "match": "match:eq"
      },
      {
        "@type": "koral:term",
        "foundry": "custom",
        "key": "AdjType",
        "layer": "p",
        "match": "match:eq",
        "value": "Pdt"
      }
    ],
    "relation": "relation:and"
  }
}

GET /

Serves the Kalamar plugin integration page. This HTML page includes:

  • Plugin information and available mapping lists
  • JavaScript integration code for Kalamar
  • SDK and server endpoints configured via sdk and server configuration fields

The SDK script and server data-attribute in the HTML are determined by the configuration file's sdk and server values, with fallback to default endpoints if not specified.

GET /health

Health check endpoint that returns "OK" with HTTP 200 status.

Progress

  • [x] Mapping functionality
  • [x] Support for rewrites
  • [x] Web service
  • [x] JSON script for Kalamar integration
  • [x] Integration of multiple mapping files
  • [ ] Support for negation
  • [ ] Support multiple mappings (by having a check list)
  • [ ] Response rewriting

COPYRIGHT AND LICENSE

Copyright (C) 2025, IDS Mannheim
Author: Nils Diewald

TermMapper is free software published under the BSD-2 License.

Disclaimer: This software was developed (as an experiment) with major assistance by AI (mainly Claude 3.5-sonnet and Claude 4-sonnet).