commit d4a380c2346e616fcbde234e5a2572802eeb66fb
Author: Araozu <fernando@araozu.dev>
Date:   Tue Sep 10 10:48:19 2024 -0500

    osm parsing

diff --git a/go.mod b/go.mod
new file mode 100644
index 0000000..1f8b2e8
--- /dev/null
+++ b/go.mod
@@ -0,0 +1,3 @@
+module osm_parser
+
+go 1.23.0
diff --git a/main.go b/main.go
new file mode 100644
index 0000000..39c0bed
--- /dev/null
+++ b/main.go
@@ -0,0 +1,177 @@
+package main
+
+import (
+	"encoding/xml"
+	"errors"
+	"fmt"
+	"log"
+	"os"
+	"regexp"
+)
+
+type OsmBounds struct {
+	MinLat float64 `xml:"minlat,attr"`
+	MinLon float64 `xml:"minlon,attr"`
+	MaxLat float64 `xml:"maxlat,attr"`
+	MaxLon float64 `xml:"maxlon,attr"`
+}
+
+type OsmNode struct {
+	Id        int64   `xml:"id,attr"`
+	Lat       float64 `xml:"lat,attr"`
+	Lon       float64 `xml:"lon,attr"`
+	Version   int64   `xml:"version,attr"`
+	Timestamp string  `xml:"timestamp,attr"`
+	Changeset int64   `xml:"changeset,attr"`
+	Uid       int64   `xml:"uid,attr"`
+	User      string  `xml:"user,attr"`
+}
+
+type OsmMember struct {
+	Type string `xml:"type,attr"`
+	Ref  int64  `xml:"ref,attr"`
+	Role string `xml:"role,attr"`
+}
+
+type OsmTag struct {
+	K string `xml:"k,attr"`
+	V string `xml:"v,attr"`
+}
+
+type OsmRelation struct {
+	Id        int64       `xml:"id,attr"`
+	Version   int32       `xml:"version,attr"`
+	Timestamp string      `xml:"timestamp,attr"`
+	Changeset int64       `xml:"changeset,attr"`
+	Uid       int64       `xml:"uid,attr"`
+	User      string      `xml:"user,attr"`
+	Members   []OsmMember `xml:"member"`
+	Tags      []OsmTag    `xml:"tag"`
+}
+
+type OsmDocument struct {
+	Note      string        `xml:"note"`
+	Bounds    OsmBounds     `xml:"bounds"`
+	Nodes     []OsmNode     `xml:"node"`
+	Relations []OsmRelation `xml:"relation"`
+}
+
+type CombiData struct {
+	/// Name of the operator of the route.
+	/// E.g.: "C7 - AqpMasivo"
+	Company string
+	/// Name of the route
+	/// E.g.: "Combi B: Villa Santa Rosa -> Terminal Terrestre"
+	Name string
+	/// Contains a single, identifiable name for the route.
+	/// This is the Name field, but without any extra info
+	/// E.g.: "B"
+	ParsedName string
+}
+
+func main() {
+	log.Println("Begin processing")
+
+	xmlData, err := os.ReadFile("aqp_map.xml")
+	if err != nil {
+		panic(err)
+	}
+
+	var osmDocument OsmDocument
+	err = xml.Unmarshal(xmlData, &osmDocument)
+	if err != nil {
+		panic(err)
+	}
+	log.Println("XML unmarshal finished")
+
+	// Get the relation with id 17642638,
+	// that relation hosts the SIT data
+	var sitId int64 = 17642638
+	var sitRelation *OsmRelation
+	for _, relation := range osmDocument.Relations {
+		if relation.Id == sitId {
+			sitRelation = &relation
+		}
+	}
+	if sitRelation == nil {
+		panic("SIT relation with id 17642638 not found!")
+	}
+
+	// (naively) get all the submembers
+	sitMembers := make([]OsmRelation, 0)
+	for _, member := range sitRelation.Members {
+		// search the member
+		// TODO: determine if this is a performance bottleneck,
+		// and requires optimization (insertion sort & binary search)
+		for _, relation := range osmDocument.Relations {
+			if relation.Id == member.Ref {
+				sitMembers = append(sitMembers, relation)
+			}
+		}
+	}
+
+	// transform sitMembers into CombiData
+	combis := make([]CombiData, 0)
+	for _, member := range sitMembers {
+		var operatorTag *OsmTag
+		var nameTag *OsmTag
+
+		for _, tag := range member.Tags {
+			if tag.K == "operator" {
+				operatorTag = &tag
+				continue
+			}
+			if tag.K == "name" {
+				nameTag = &tag
+				continue
+			}
+		}
+
+		if operatorTag == nil {
+			log.Fatalf("Found a SIT member without an operator tag, with id %d\n", member.Id)
+		}
+		if nameTag == nil {
+			log.Fatalf("Found a SIT member without a name tag, with id %d\n", member.Id)
+		}
+
+		parsedName, err := parseRoute(nameTag.V)
+		if err != nil {
+			log.Printf("SIT member id: %s\n", member.Id)
+			panic(err)
+		}
+
+		combis = append(combis, CombiData{
+			Company:    operatorTag.V,
+			Name:       nameTag.V,
+			ParsedName: parsedName,
+		})
+	}
+
+	for _, combi := range combis {
+		log.Printf("%s: `%s`\n", combi.Company, combi.ParsedName)
+	}
+}
+
+func parseRoute(routeName string) (string, error) {
+	// There are 3 types of route name formats:
+	// Combi xx : ____
+	// Curster xx : ____
+	// xx : ___
+
+	// attempt first format
+	match := regexp.MustCompile("Combi (.+) :.+").FindStringSubmatch(routeName)
+	if match != nil {
+		return match[1], nil
+	}
+	match = regexp.MustCompile("Custer (.+) :.+").FindStringSubmatch(routeName)
+	if match != nil {
+		return match[1], nil
+	}
+	match = regexp.MustCompile("(.+) :.+").FindStringSubmatch(routeName)
+	if match != nil {
+		return match[1], nil
+	}
+
+	// Return an error
+	return "", errors.New(fmt.Sprintf("Route %s did not match any pattern", routeName))
+}