dawarich/app/services/kml/importer.rb
2025-11-14 18:04:46 +01:00

234 lines
6.6 KiB
Ruby

# frozen_string_literal: true
require 'rexml/document'
class Kml::Importer
include Imports::Broadcaster
include Imports::FileLoader
attr_reader :import, :user_id, :file_path
def initialize(import, user_id, file_path = nil)
@import = import
@user_id = user_id
@file_path = file_path
end
def call
file_content = load_file_content
doc = REXML::Document.new(file_content)
points_data = []
# Process all Placemarks which can contain various geometry types
REXML::XPath.each(doc, '//Placemark') do |placemark|
points_data.concat(parse_placemark(placemark))
end
# Process gx:Track elements (Google Earth extensions for GPS tracks)
REXML::XPath.each(doc, '//gx:Track') do |track|
points_data.concat(parse_gx_track(track))
end
points_data.compact!
return if points_data.empty?
# Process in batches to avoid memory issues with large files
points_data.each_slice(1000) do |batch|
bulk_insert_points(batch)
end
end
private
def parse_placemark(placemark)
points = []
timestamp = extract_timestamp(placemark)
# Handle Point geometry
point_node = REXML::XPath.first(placemark, './/Point/coordinates')
if point_node
coords = parse_coordinates(point_node.text)
points << build_point(coords.first, timestamp, placemark) if coords.any?
end
# Handle LineString geometry (tracks/routes)
linestring_node = REXML::XPath.first(placemark, './/LineString/coordinates')
if linestring_node
coords = parse_coordinates(linestring_node.text)
coords.each do |coord|
points << build_point(coord, timestamp, placemark)
end
end
# Handle MultiGeometry (can contain multiple Points, LineStrings, etc.)
REXML::XPath.each(placemark, './/MultiGeometry//coordinates') do |coords_node|
coords = parse_coordinates(coords_node.text)
coords.each do |coord|
points << build_point(coord, timestamp, placemark)
end
end
points.compact
end
def parse_gx_track(track)
# Google Earth Track extension with coordinated when/coord pairs
points = []
timestamps = []
REXML::XPath.each(track, './/when') do |when_node|
timestamps << when_node.text.strip
end
coordinates = []
REXML::XPath.each(track, './/gx:coord') do |coord_node|
coordinates << coord_node.text.strip
end
# Match timestamps with coordinates
[timestamps.size, coordinates.size].min.times do |i|
begin
time = Time.parse(timestamps[i]).to_i
coord_parts = coordinates[i].split(/\s+/)
next if coord_parts.size < 2
lng, lat, alt = coord_parts.map(&:to_f)
points << {
lonlat: "POINT(#{lng} #{lat})",
altitude: alt&.to_i || 0,
timestamp: time,
import_id: import.id,
velocity: 0.0,
raw_data: { source: 'gx_track', index: i },
user_id: user_id,
created_at: Time.current,
updated_at: Time.current
}
rescue StandardError => e
Rails.logger.warn("Failed to parse gx:Track point at index #{i}: #{e.message}")
next
end
end
points
end
def parse_coordinates(coord_text)
# KML coordinates format: "longitude,latitude[,altitude] ..."
# Multiple coordinates separated by whitespace
return [] if coord_text.blank?
coord_text.strip.split(/\s+/).map do |coord_str|
parts = coord_str.split(',')
next if parts.size < 2
{
lng: parts[0].to_f,
lat: parts[1].to_f,
alt: parts[2]&.to_f || 0.0
}
end.compact
end
def extract_timestamp(placemark)
# Try TimeStamp first
timestamp_node = REXML::XPath.first(placemark, './/TimeStamp/when')
return Time.parse(timestamp_node.text).to_i if timestamp_node
# Try TimeSpan begin
timespan_begin = REXML::XPath.first(placemark, './/TimeSpan/begin')
return Time.parse(timespan_begin.text).to_i if timespan_begin
# Try TimeSpan end as fallback
timespan_end = REXML::XPath.first(placemark, './/TimeSpan/end')
return Time.parse(timespan_end.text).to_i if timespan_end
# Default to import creation time if no timestamp found
import.created_at.to_i
rescue StandardError => e
Rails.logger.warn("Failed to parse timestamp: #{e.message}")
import.created_at.to_i
end
def build_point(coord, timestamp, placemark)
return if coord[:lat].blank? || coord[:lng].blank?
{
lonlat: "POINT(#{coord[:lng]} #{coord[:lat]})",
altitude: coord[:alt].to_i,
timestamp: timestamp,
import_id: import.id,
velocity: extract_velocity(placemark),
raw_data: extract_extended_data(placemark),
user_id: user_id,
created_at: Time.current,
updated_at: Time.current
}
end
def extract_velocity(placemark)
# Try to extract speed from ExtendedData
speed_node = REXML::XPath.first(placemark, ".//Data[@name='speed']/value") ||
REXML::XPath.first(placemark, ".//Data[@name='Speed']/value") ||
REXML::XPath.first(placemark, ".//Data[@name='velocity']/value")
return speed_node.text.to_f.round(1) if speed_node
0.0
rescue StandardError
0.0
end
def extract_extended_data(placemark)
data = {}
# Extract name if present
name_node = REXML::XPath.first(placemark, './/name')
data['name'] = name_node.text.strip if name_node
# Extract description if present
desc_node = REXML::XPath.first(placemark, './/description')
data['description'] = desc_node.text.strip if desc_node
# Extract all ExtendedData/Data elements
REXML::XPath.each(placemark, './/ExtendedData/Data') do |data_node|
name = data_node.attributes['name']
value_node = REXML::XPath.first(data_node, './value')
data[name] = value_node.text if name && value_node
end
data
rescue StandardError => e
Rails.logger.warn("Failed to extract extended data: #{e.message}")
{}
end
def bulk_insert_points(batch)
unique_batch = batch.uniq { |record| [record[:lonlat], record[:timestamp], record[:user_id]] }
# rubocop:disable Rails/SkipsModelValidations
Point.upsert_all(
unique_batch,
unique_by: %i[lonlat timestamp user_id],
returning: false,
on_duplicate: :skip
)
# rubocop:enable Rails/SkipsModelValidations
broadcast_import_progress(import, unique_batch.size)
rescue StandardError => e
create_notification("Failed to process KML file: #{e.message}")
end
def create_notification(message)
Notification.create!(
user_id: user_id,
title: 'KML Import Error',
content: message,
kind: :error
)
end
end