avorion-docgen/class.go

package main

import (
	_ "embed"
	"fmt"
	"log"
	"os"
	"path/filepath"
	"regexp"
	"strings"
	"text/template"

	"github.com/PuerkitoBio/goquery"
	"github.com/davecgh/go-spew/spew"
)

//go:embed class.tmpl
var templatestr string
var classTemplate *template.Template

var fns = template.FuncMap{
	"plus1": func(x int) int {
		return x + 1
	},
}

func init() {
	var err error
	classTemplate, err = template.New("class").Funcs(fns).Parse(templatestr)
	if err != nil {
		Error.Printf("Error parsing template: %v", err)
		return
	}
}

type (
	Class struct {
		ClassName    string
		Description  string
		Fields       []Field
		Methods      []Method
		Constructors []Constructor
	}
	Field struct {
		Name    string
		Type    string
		Comment string
	}
	Method struct {
		Name    string
		Params  []Param
		Returns []Return
		Comment string
	}
	Param struct {
		Name    string
		Type    string
		Comment string
	}
	Return struct {
		Type    string
		Comment string
	}
	Constructor struct {
		Params  []Param
		Returns string
		Comment string
	}
)

func (c *Class) GetOutFile(root string) (*os.File, error) {
	if c.ClassName == "" {
		return nil, fmt.Errorf("ClassName is empty")
	}
	filename := fmt.Sprintf("%s.lua", c.ClassName)
	filename = strings.ReplaceAll(filename, " ", "")
	filename = strings.ReplaceAll(filename, "-", "")
	filename = strings.ReplaceAll(filename, ",", "")
	filename = strings.ReplaceAll(filename, ":", "")
	filePath := filepath.Join(root, filename)
	filePath = filepath.Clean(filePath)
	f, err := os.Create(filePath)
	if err != nil {
		return nil, err
	}
	return f, nil
}

func (c *Class) Write(root string, tmpl *template.Template) error {
	outfile, err := c.GetOutFile(root)
	if err != nil {
		return fmt.Errorf("error creating output file %v: %v", c.ClassName, err)
	}
	// spew.Dump(c)
	err = tmpl.Execute(outfile, c)
	if err != nil {
		return fmt.Errorf("error writing output file %v: %v", c.ClassName, err)
	}
	return nil
}

func ParseClass(file string) (*Class, error) {
	log.Printf("Parsing file: '%s'", file)
	res := Class{
		Fields:       []Field{},
		Methods:      []Method{},
		Constructors: []Constructor{},
	}
	filehandle, err := os.Open(file)
	if err != nil {
		return nil, fmt.Errorf("error opening file: %w", err)
	}
	defer filehandle.Close()

	doc, err := goquery.NewDocumentFromReader(filehandle)
	if err != nil {
		return nil, fmt.Errorf("error parsing file: %w", err)
	}

	dataContainer := doc.Find("div.floatright")
	if dataContainer.Length() == 0 {
		return nil, fmt.Errorf("no data container found")
	}

	className, err := getClassName(dataContainer)
	if err != nil {
		return nil, fmt.Errorf("error getting class name: %w", err)
	}
	res.ClassName = className

	classDescription, err := getClassDescription(dataContainer)
	if err != nil {
		return nil, fmt.Errorf("error getting class description: %w", err)
	}
	res.Description = classDescription

	constructor, err := getConstructor(dataContainer)
	if err != nil {
		return nil, fmt.Errorf("error getting constructor: %w", err)
	}
	res.Constructors = append(res.Constructors, constructor)

	// Doing div+div specifically skips only the constructor which is usually the first div
	// So the constructor would then be located at h1 + p + div OR h1 + div if there is no description (no p)
	dataElements := dataContainer.ChildrenFiltered("div + div")
	if dataElements.Length() == 0 {
		return nil, fmt.Errorf("no data elements found")
	}

	dataElements.Each(func(i int, s *goquery.Selection) {
		id, ok := s.Attr("id")
		if !ok {
			return
		}
		if id == "Properties" {
			s.Children().Each(func(i int, s *goquery.Selection) {
				field, err := parseField(s)
				if err != nil {
					Error.Printf("Error parsing field: %v", err)
					return
				}
				res.Fields = append(res.Fields, field)
			})
		} else {
			method, err := parseMethod(s)
			if err != nil {
				Error.Printf("Error parsing method: %v", err)
				return
			}
			res.Methods = append(res.Methods, method)
		}
	})

	return &res, nil
}

func getClassName(dataContainer *goquery.Selection) (string, error) {
	class := dataContainer.ChildrenFiltered("h1")
	if class.Length() == 0 {
		return "", fmt.Errorf("no class found")
	}
	res := CleanUp(class.Text())
	return res, nil
}

var manySpaceRe = regexp.MustCompile(`\s{2,}`)

func getClassDescription(dataContainer *goquery.Selection) (string, error) {
	class := dataContainer.ChildrenFiltered("h1 + p")
	if class.Length() == 0 {
		return "", fmt.Errorf("no class description found")
	}
	res := CleanUp(class.Text())
	return res, nil
}

func getConstructor(dataContainer *goquery.Selection) (Constructor, error) {
	resConstructor := Constructor{}

	constructorBlock := dataContainer.Find("h1 + p + div")
	if constructorBlock.Length() == 0 {
		constructorBlock = dataContainer.Find("h1 + div")
		if constructorBlock.Length() == 0 {
			return resConstructor, fmt.Errorf("no constructor found")
		}
	}

	types := constructorBlock.Find("div.function span.type")
	if types.Length() == 0 {
		return resConstructor, fmt.Errorf("no types found")
	}
	params := constructorBlock.Find("div.function span.parameter")
	if params.Length() == 0 {
		return resConstructor, fmt.Errorf("no params found")
	}

	types.Each(func(i int, s *goquery.Selection) {
		resConstructor.Params = append(resConstructor.Params, Param{
			Name:    CleanUp(params.Eq(i).Text()),
			Type:    MapType(CleanUp(types.Eq(i).Text())),
			Comment: "",
		})
	})

	descriptor := constructorBlock.Find("div:not(.function)")
	if descriptor.Length() == 0 {
		return resConstructor, fmt.Errorf("no descriptor found")
	}

	// 0 nothing
	// 1 parameters
	// 2 returns
	state := 0
	children := descriptor.Eq(0).Children()
	childrenLength := children.Length()
	children.Each(func(i int, s *goquery.Selection) {
		if i == childrenLength-1 {
			// For some stupid reason the last child is always a mispalced </p> tag that does not close any open <p>
			// I assume this is a bug with their documentation generator
			// So we just ignore it
			return
		}
		text := CleanUp(s.Text())
		if text == "" {
			return
		}
		if s.Is("p") {
			switch text {
			case "Parameters":
				state = 1
				return
			case "Returns":
				state = 2
				return
			}
		} else {
			switch state {
			case 0:
				return
			case 1:
				param := s.ChildrenFiltered("span.parameter").Text()
				paramTrimmed := CleanUp(param)
				for i := range resConstructor.Params {
					cparam := &resConstructor.Params[i]
					if paramTrimmed == cparam.Name {
						cleanText := strings.TrimPrefix(text, param)
						cparam.Comment = CleanUp(cleanText)
					}
				}
			case 2:
				cleaned := CleanUp(s.Text())
				if resConstructor.Returns != "" {
					resConstructor.Returns += "\n"
				}
				resConstructor.Returns = cleaned
			}
		}
	})

	// spew.Dump(resConstructor)
	return resConstructor, nil
}

func parseField(s *goquery.Selection) (Field, error) {
	res := Field{}

	id, ok := s.Attr("id")
	if !ok {
		return res, fmt.Errorf("no id found")
	}
	res.Name = id

	typeElement := s.Find("span.type")
	if typeElement.Length() == 0 {
		return res, fmt.Errorf("no type found")
	}
	res.Type = MapType(CleanUp(typeElement.Text()))

	readonlyElement := s.Find("td[align='right']")
	if readonlyElement.Length() != 0 {
		// return res, fmt.Errorf("no readonly found")
		res.Comment = CleanUp(readonlyElement.Text())
	}

	comments := s.ChildrenFiltered("div")
	if comments.Length() == 0 {
		return res, fmt.Errorf("no comments found")
	}
	comments.Each(func(i int, s *goquery.Selection) {
		text := CleanUp(s.Text())
		if text == "" {
			return
		}
		if res.Comment != "" {
			res.Comment += ". "
		}
		res.Comment += text
	})
	res.Comment = strings.ReplaceAll(res.Comment, "\n--", ". ")

	return res, nil
}

func parseMethod(s *goquery.Selection) (Method, error) {
	res := Method{}

	id, ok := s.Attr("id")
	if !ok {
		return res, fmt.Errorf("no id found")
	}
	res.Name = id

	returns := s.Find("span.keyword")
	if returns.Length() != 0 {
		returnsText := CleanUp(returns.Text())
		returnsText = strings.ReplaceAll(returnsText, "function", "")
		returnsText = CleanUp(returnsText)

		res.Returns = append(res.Returns, Return{
			Type:    MapType(returnsText),
			Comment: "",
		})
	}

	types := s.Find("span.type")
	parameters := s.Find("span.parameter")
	types.Each(func(i int, s *goquery.Selection) {
		res.Params = append(res.Params, Param{
			Name:    MapName(CleanUp(parameters.Eq(i).Text())),
			Type:    MapType(CleanUp(types.Eq(i).Text())),
			Comment: "",
		})
	})

	// <div id="add" class="codecontainer">
	// 	<div id="add" class="function">
	// 		<p>
	// 			<span class="keyword">function var</span> add(<span class="type">CargoBay</span> <span
	// 				class="parameter">other</span>) <br />
	// 		</p>
	// 	</div>
	// 	<div id="add" class="">
	// 		<p><span class="docheader">Returns</span></p>
	// 		<div class="indented">
	// 			<p>
	// 				nothing
	// 			</p>
	// 		</div>
	// 		</p>
	// 	</div>
	// </div>

	spew.Dump(res)
	return res, nil
}

func CleanUp(s string) string {
	s = strings.TrimSpace(s)
	s = strings.ReplaceAll(s, "\t", " ")
	s = strings.ReplaceAll(s, "\n", "")
	s = manySpaceRe.ReplaceAllString(s, " ")
	s = strings.ReplaceAll(s, ". ", "\n--")
	return s
}