kof-parser

A JavaScript/TypeScript parser for Norwegian KOF files. Usable as an ES module or via script injection. Includes demo scripts and sample KOF files.

Features

• Parse Norwegian KOF files
• Use as ES module or via <script src=...>
• TypeScript and JavaScript support
• Demo scripts and sample files included

Additional features added during development

• Export parsed geometries (Point / LineString / Polygon) to GeoJSON.
• Optionally produce reprojections to EPSG:4326 (lat/lon) for demo files that contain an EPSG hint (currently supports EPSG:25832 and a best-effort EPSG:5110 definition) using proj4.
• Geometry properties include name (PPPPPPPPPP) and fcode (KKKKKKKK). Lines/polygons inherit these from their first point when not explicitly set.
• Tests generate human-readable .log output per demo file (under test/mocha/logs/) and a test/geojson/ folder containing produced GeoJSON files (both are gitignored by default).
• Parser emits structured warnings with line numbers and diagnostic strategies to aid debugging.

Example KOF Files

Example KOF files for testing parsing functionality are located in:

demo/kof_files

KOF File Format (Column Description)

The positioning of the data content in a KOF file is given by the following header description string (which tells us in which column ranges the data is located):

-05 PPPPPPPPPP KKKKKKKK XXXXXXXX.XXX YYYYYYY.YYY ZZZZ.ZZZ

Where:

• 05 is the row code (required)
• PPPPPPPPPP is the observation/point name (optional) (the parser will use an ID system per object and per observation, so the point name can be optional)
• KKKKKKKK is the code of the observation/point (optional) (defaults to empty string if blank)
• XXXXXXXX.XXX is northing or latitude (required if the row contains an observation; not required for start/close codes like 91, 99, 96, etc.)
• YYYYYYY.YYY is easting or longitude (required if the row contains an observation; not required for start/close codes like 91, 99, 96, etc.)
• ZZZZ.ZZZ is elevation (optional; set to -500 if not found)
• Other optional parameters after ZZZZ.ZZZ are allowed and will be parsed if present.

Geometry Start/Stop in KOF Files

KOF files use special codes to indicate the start and end of geometries:

• Points: Each point is a single row with code 05.
• Lines:
  • Start with a row code 91 (e.g. 09_91), followed by one or more 05 rows for the line's vertices.
  • End with a row code 99 (e.g. 09_99).
• Polygons:
  • Start with a row code 91 (e.g. 09_91), followed by one or more 05 rows for the polygon's vertices.
  • Closed with a row code 96 (e.g. 09_96).
  • Multiple polygons may be present in a file, each with their own start/close codes.

Example KOF File Snippet

From an actual KOF file with header and data:

-05 PPPPPPPPPP KKKKKKKK XXXXXXXX.XXX YYYYYYY.YYY ZZZZ.ZZZ
 05 LYKT-04    8751      6540290.081  314103.268    7.934

Usage

ES Module

import { kof } from 'kof-parser';
const result = kof.parse(kofString);

Script Injection

<script src="dist/kof-parser.umd.js"></script>
<script>
  const result = kof.parse(kofString);
</script>

Quick API examples

Node / CommonJS (parsing and GeoJSON export):

const { KOF } = require('kof-parser');
const fs = require('fs');

const content = fs.readFileSync('demo/kof_files/example.kof', 'utf8');
const k = new KOF('example.kof', content, { sourceCrs: 'EPSG:25832' });
k.parse();
// Export GeoJSON (no reprojection)
const gj = k.toGeoJSON();
console.log(gj.type, gj.features.length);

// Reproject to EPSG:4326 (if proj4 available)
const gj4326 = k.reproject('EPSG:4326');
console.log('Reprojected:', gj4326.features.length);

ES module example:

import { KOF } from 'kof-parser';
const content = await fetch('/demo/kof_files/example.kof').then(r => r.text());
const k = new KOF('example.kof', content, { sourceCrs: 'EPSG:25832' });
k.parse();
const geojson = k.toGeoJSON({ crs: { from: 'EPSG:25832', to: 'EPSG:4326' } });
console.log(geojson);

Notes and development status

• The parser prefers fixed-column parsing when a header line (-05 PPPPPPPPPP KKKKKKKK ...) is present; for files without a header it falls back to heuristic token parsing.
• The project includes a comprehensive Mocha test-suite under test/mocha/ that covers attributes, encodings, grouping, KOF codes and demo GeoJSON exports. Running npm test will clean old logs, run tests, and (for demo files) produce GeoJSON and reprojection outputs under test/geojson/.
• Reprojection: proj4 is used during testing to convert demo files containing epsg25832 or epsg5110 in their filenames to EPSG:4326; the converted files are written with epsg4326 in their filenames.
• EPSG:5110: the repo registers a reasonable NTM10-like proj4 definition as a best-effort. Replace this definition with an authoritative proj4 string if you have one.

Build / CI note

To ensure clean, reproducible builds and to avoid TypeScript write-errors when a dist/ folder is present in a working tree, the project now performs an explicit clean before building:

• Locally: run npm run clean (this removes the dist/ directory) or npm run build which will run the clean step automatically.
• CI: the GitHub Actions workflow runs a git clean -fdx step before building so the runner has a fresh workspace and tsc can write output into dist/ without conflicts.

This keeps local and CI builds consistent and prevents "TS5055: Cannot write file ... because it would overwrite input file" errors when dist/ is present in the repo.

License

This project is released under the MIT License; see the LICENSE file in the repository for full terms.

Release v0.1.0

• Added explicit LICENSE (MIT).
• Cleaned package metadata (removed accidental core-module devDependencies).
• Tests and demo exports generate .log and test/geojson/ artifacts (gitignored).

Publishing notes

• The package is configured to publish dist/ and types/ via the files field in package.json. Before publishing, run a fresh build (npm run build) and validate the bundle with npm pack.

License

MIT