1.0.0

- used mod instead of dep (removed vendor dir)
- refactoring (most of the logic now in the GHDoc)
- handle "API rate limit exceeded" error
- added bash/zsh autocomplete section into README

README.md

@@ -33,6 +33,7 @@ Table of Contents
     * [Depth](#depth)
     * [No Escape](#no-escape)
     * [Github token](#github-token)
+    * [Bash/ZSH auto\-complete](#bashzsh-auto-complete)
   * [LICENSE](#license)
 
 Created by [gh-md-toc](https://github.com/ekalinin/github-markdown-toc.go)
@@ -334,6 +335,21 @@ Table of Contents
 * [LICENSE](#license)
 ```
 
+Bash/ZSH auto-complete
+----------------------
+
+
+
+Just add a simple command into your `~/.bashrc` or `~/.zshrc`:
+
+```bash
+# for zsh
+eval "$(gh-md-toc --completion-script-zsh)"
+
+# for bash
+eval "$(gh-md-toc --completion-script-bash)"
+```
+
 LICENSE
 =======
 

ghdoc.go

@@ -0,0 +1,173 @@
+package main
+
+import (
+	"fmt"
+	"io/ioutil"
+	"log"
+	"net/url"
+	"os"
+	"regexp"
+	"strconv"
+	"strings"
+)
+
+// GHToc GitHub TOC
+type GHToc []string
+
+// Print print TOC to the console
+func (toc *GHToc) Print() {
+	for _, tocItem := range *toc {
+		fmt.Println(tocItem)
+	}
+	fmt.Println()
+}
+
+// GHDoc GitHub document
+type GHDoc struct {
+	Path     string
+	AbsPaths bool
+	Depth    int
+	Escape   bool
+	GhToken  string
+	Indent   int
+	Debug    bool
+	html     string
+}
+
+// NewGHDoc create GHDoc
+func NewGHDoc(Path string, AbsPaths bool, Depth int, Escape bool, Token string, Indent int, Debug bool) *GHDoc {
+	return &GHDoc{Path, AbsPaths, Depth, Escape, Token, Indent, Debug, ""}
+}
+
+func (doc *GHDoc) d(msg string) {
+	if doc.Debug {
+		log.Println(msg)
+	}
+}
+
+// IsRemoteFile checks if path is for remote file or not
+func (doc *GHDoc) IsRemoteFile() bool {
+	u, err := url.Parse(doc.Path)
+	if err != nil || u.Scheme == "" {
+		doc.d("IsRemoteFile: false")
+		return false
+	}
+	doc.d("IsRemoteFile: true")
+	return true
+}
+
+// Convert2HTML downloads remote file
+func (doc *GHDoc) Convert2HTML() error {
+	doc.d("Convert2HTML: start.")
+	defer doc.d("Convert2HTML: done.")
+
+	if doc.IsRemoteFile() {
+		htmlBody, ContentType, err := httpGet(doc.Path)
+		doc.d("Convert2HTML: remote file. content-type: " + ContentType)
+		if err != nil {
+			return err
+		}
+
+		// if not a plain text - return the result (should be html)
+		if strings.Split(ContentType, ";")[0] != "text/plain" {
+			doc.html = string(htmlBody)
+			return nil
+		}
+
+		// if remote file's content is a plain text
+		// we need to convert it to html
+		tmpfile, err := ioutil.TempFile("", "ghtoc-remote-txt")
+		if err != nil {
+			return err
+		}
+		defer tmpfile.Close()
+		doc.Path = tmpfile.Name()
+		if err = ioutil.WriteFile(tmpfile.Name(), htmlBody, 0644); err != nil {
+			return err
+		}
+	}
+	doc.d("Convert2HTML: local file: " + doc.Path)
+	if _, err := os.Stat(doc.Path); os.IsNotExist(err) {
+		return err
+	}
+	htmlBody, err := ConvertMd2Html(doc.Path, doc.GhToken)
+	doc.d("Convert2HTML: converted to html, size: " + strconv.Itoa(len(htmlBody)))
+	if err != nil {
+		return err
+	}
+	// doc.d("Convert2HTML: " + htmlBody)
+	doc.html = htmlBody
+	return nil
+}
+
+// GrabToc gets TOC from html
+func (doc *GHDoc) GrabToc() *GHToc {
+	doc.d("GrabToc: start, html size: " + strconv.Itoa(len(doc.html)))
+	defer doc.d("GrabToc: done.")
+
+	re := `(?si)<h(?P<num>[1-6])>\s*` +
+		`<a\s*id="user-content-[^"]*"\s*class="anchor"\s*` +
+		`href="(?P<href>[^"]*)"[^>]*>\s*` +
+		`.*?</a>(?P<name>.*?)</h`
+	r := regexp.MustCompile(re)
+	listIndentation := generateListIndentation(doc.Indent)
+
+	toc := GHToc{}
+	minHeaderNum := 6
+	var groups []map[string]string
+	doc.d("GrabToc: matching ...")
+	for idx, match := range r.FindAllStringSubmatch(doc.html, -1) {
+		doc.d("GrabToc: match #" + strconv.Itoa(idx) + " ...")
+		group := make(map[string]string)
+		// fill map for groups
+		for i, name := range r.SubexpNames() {
+			if i == 0 || name == "" {
+				continue
+			}
+			doc.d("GrabToc: process group: " + name + " ...")
+			group[name] = removeStuf(match[i])
+		}
+		// update minimum header number
+		n, _ := strconv.Atoi(group["num"])
+		if n < minHeaderNum {
+			minHeaderNum = n
+		}
+		groups = append(groups, group)
+	}
+
+	var tmpSection string
+	doc.d("GrabToc: processing groups ...")
+	for _, group := range groups {
+		// format result
+		n, _ := strconv.Atoi(group["num"])
+		if doc.Depth > 0 && n > doc.Depth {
+			continue
+		}
+
+		link := group["href"]
+		if doc.AbsPaths {
+			link = doc.Path + link
+		}
+
+		tmpSection = removeStuf(group["name"])
+		if doc.Escape {
+			tmpSection = EscapeSpecChars(tmpSection)
+		}
+		tocItem := strings.Repeat(listIndentation(), n-minHeaderNum) + "* " +
+			"[" + tmpSection + "]" +
+			"(" + link + ")"
+		//fmt.Println(tocItem)
+		toc = append(toc, tocItem)
+	}
+
+	return &toc
+}
+
+// GetToc return GHToc for a document
+func (doc *GHDoc) GetToc() *GHToc {
+	if err := doc.Convert2HTML(); err != nil {
+		log.Fatal(err)
+		return nil
+	}
+	return doc.GrabToc()
+}

go.mod

@@ -0,0 +1,7 @@
+module github.com/ekalinin/github-markdown-toc.go
+
+require (
+	github.com/alecthomas/template v0.0.0-20160405071501-a0175ee3bccc
+	github.com/alecthomas/units v0.0.0-20151022065526-2efee857e7cf
+	gopkg.in/alecthomas/kingpin.v2 v2.2.4
+)

go.sum

@@ -0,0 +1,6 @@
+github.com/alecthomas/template v0.0.0-20160405071501-a0175ee3bccc h1:cAKDfWh5VpdgMhJosfJnn5/FoN2SRZ4p7fJNX58YPaU=
+github.com/alecthomas/template v0.0.0-20160405071501-a0175ee3bccc/go.mod h1:LOuyumcjzFXgccqObfd/Ljyb9UuFJ6TxHnclSeseNhc=
+github.com/alecthomas/units v0.0.0-20151022065526-2efee857e7cf h1:qet1QNfXsQxTZqLG4oE62mJzwPIB8+Tee4RNCL9ulrY=
+github.com/alecthomas/units v0.0.0-20151022065526-2efee857e7cf/go.mod h1:ybxpYRFXyAe+OPACYpWeL0wqObRcbAqCMya13uyzqw0=
+gopkg.in/alecthomas/kingpin.v2 v2.2.4 h1:CC8tJ/xljioKrK6ii3IeWVXU4Tw7VB+LbjZBJaBxN50=
+gopkg.in/alecthomas/kingpin.v2 v2.2.4/go.mod h1:FMv+mEhP44yOT+4EoQTLFTRgOQ1FBLkstjWtayDeSgw=

internals.go

@@ -0,0 +1,111 @@
+package main
+
+import (
+	"bytes"
+	"errors"
+	"io"
+	"io/ioutil"
+	"net/http"
+	"os"
+	"strings"
+)
+
+// check checks if there whas an error and do panic if it was
+func check(e error) {
+	if e != nil {
+		panic(e)
+	}
+}
+
+// doHTTPReq executes a particullar http request
+func doHTTPReq(req *http.Request) ([]byte, string, error) {
+	req.Header.Set("User-Agent", userAgent)
+	client := &http.Client{}
+	resp, err := client.Do(req)
+	if err != nil {
+		return []byte{}, "", err
+	}
+
+	defer resp.Body.Close()
+	body, err := ioutil.ReadAll(resp.Body)
+	if err != nil {
+		return []byte{}, "", err
+	}
+
+	if resp.StatusCode == http.StatusForbidden {
+		return []byte{}, resp.Header.Get("Content-type"), errors.New(string(body))
+	}
+
+	return body, resp.Header.Get("Content-type"), nil
+}
+
+// Executes HTTP GET request
+func httpGet(urlPath string) ([]byte, string, error) {
+	req, err := http.NewRequest("GET", urlPath, nil)
+	if err != nil {
+		return []byte{}, "", err
+	}
+	return doHTTPReq(req)
+}
+
+// httpPost executes HTTP POST with file content
+func httpPost(urlPath string, filePath string) (string, error) {
+	file, err := os.Open(filePath)
+	if err != nil {
+		return "", err
+	}
+	defer file.Close()
+
+	body := &bytes.Buffer{}
+	io.Copy(body, file)
+
+	req, err := http.NewRequest("POST", urlPath, body)
+	if err != nil {
+		return "", err
+	}
+
+	req.Header.Set("Content-Type", "text/plain")
+
+	resp, _, err := doHTTPReq(req)
+	return string(resp), err
+}
+
+// removeStuf trims spaces, removes new lines and code tag from a string
+func removeStuf(s string) string {
+	res := strings.Replace(s, "\n", "", -1)
+	res = strings.Replace(res, "<code>", "", -1)
+	res = strings.Replace(res, "</code>", "", -1)
+	res = strings.TrimSpace(res)
+
+	return res
+}
+
+// generate func of custom spaces indentation
+func generateListIndentation(spaces int) func() string {
+	return func() string {
+		return strings.Repeat(" ", spaces)
+	}
+}
+
+// Public
+
+// EscapeSpecChars Escapes special characters
+func EscapeSpecChars(s string) string {
+	specChar := []string{"\\", "`", "*", "_", "{", "}", "#", "+", "-", ".", "!"}
+	res := s
+
+	for _, c := range specChar {
+		res = strings.Replace(res, c, "\\"+c, -1)
+	}
+	return res
+}
+
+// ConvertMd2Html Sends Markdown to the github converter
+// and returns html
+func ConvertMd2Html(localpath string, token string) (string, error) {
+	url := "https://api.github.com/markdown/raw"
+	if token != "" {
+		url += "?access_token=" + token
+	}
+	return httpPost(url, localpath)
+}