From 98695223cab9e4b0206228a37b07dbf874f846c0 Mon Sep 17 00:00:00 2001
From: wgroeneveld <wouter@brainbaking.com>
Date: Sat, 10 Apr 2021 10:17:38 +0200
Subject: [PATCH] rss link collecting impl

---
 .idea/inspectionProfiles/Project_Default.xml |  1 +
 app/rss/feed.go                              | 82 ++++++++++++++++++++
 app/rss/feed_test.go                         | 34 ++++++++
 app/webmention/send/rsslinkcollector.go      | 52 +++++++++++++
 app/webmention/send/rsslinkcollector_test.go | 62 +++++++++++++++
 app/webmention/send/send.go                  | 12 +++
 common/time.go                               | 10 +++
 common/time_test.go                          | 45 +++++++++++
 8 files changed, 298 insertions(+)
 create mode 100644 app/rss/feed.go
 create mode 100644 app/rss/feed_test.go
 create mode 100644 app/webmention/send/rsslinkcollector.go
 create mode 100644 app/webmention/send/rsslinkcollector_test.go
 create mode 100644 common/time_test.go

diff --git a/.idea/inspectionProfiles/Project_Default.xml b/.idea/inspectionProfiles/Project_Default.xml
index 09dbac8..73948b9 100644
--- a/.idea/inspectionProfiles/Project_Default.xml
+++ b/.idea/inspectionProfiles/Project_Default.xml
@@ -1,6 +1,7 @@
 <component name="InspectionProjectProfileManager">
   <profile version="1.0">
     <option name="myName" value="Project Default" />
+    <inspection_tool class="GoCommentStart" enabled="false" level="WEAK WARNING" enabled_by_default="false" />
     <inspection_tool class="GrazieInspection" enabled="false" level="TYPO" enabled_by_default="false" />
     <inspection_tool class="LanguageDetectionInspection" enabled="false" level="WARNING" enabled_by_default="false" />
     <inspection_tool class="SpellCheckingInspection" enabled="false" level="TYPO" enabled_by_default="false">
diff --git a/app/rss/feed.go b/app/rss/feed.go
new file mode 100644
index 0000000..5209730
--- /dev/null
+++ b/app/rss/feed.go
@@ -0,0 +1,82 @@
+package rss
+
+import (
+	"brainbaking.com/go-jamming/common"
+	"encoding/xml"
+	"errors"
+	"github.com/rs/zerolog/log"
+	"html/template"
+	"time"
+)
+
+// someone already did this for me, yay! https://siongui.github.io/2015/03/03/go-parse-web-feed-rss-atom/
+type Rss2 struct {
+	XMLName xml.Name `xml:"rss"`
+	Version string   `xml:"version,attr"`
+	// Required
+	Title       string `xml:"channel>title"`
+	Link        string `xml:"channel>link"`
+	Description string `xml:"channel>description"`
+	// Optional
+	PubDate  string `xml:"channel>pubDate"`
+	ItemList []Item `xml:"channel>item"`
+}
+
+type Item struct {
+	// Required
+	Title       string        `xml:"title"`
+	Link        string        `xml:"link"`
+	Description template.HTML `xml:"description"`
+	// Optional
+	Content  template.HTML `xml:"encoded"`
+	PubDate  string        `xml:"pubDate"`
+	Comments string        `xml:"comments"`
+}
+
+func (itm Item) PubDateAsTime() time.Time {
+	// format: Tue, 16 Mar 2021 17:07:14 +0000
+	t, err := time.Parse("Mon, 02 Jan 2006 15:04:05 +0000", itm.PubDate)
+	if err != nil {
+		log.Warn().Str("pubDate", itm.PubDate).Msg("Incorrectly formatted RSS date, reverting to now")
+		return common.Now()
+	}
+	return t
+}
+
+type Link struct {
+	Href string `xml:"href,attr"`
+}
+
+type Author struct {
+	Name  string `xml:"name"`
+	Email string `xml:"email"`
+}
+
+type Entry struct {
+	Title   string `xml:"title"`
+	Summary string `xml:"summary"`
+	Content string `xml:"content"`
+	Id      string `xml:"id"`
+	Updated string `xml:"updated"`
+	Link    Link   `xml:"link"`
+	Author  Author `xml:"author"`
+}
+
+func ParseFeed(content []byte) (Rss2, error) {
+	v := Rss2{}
+	err := xml.Unmarshal(content, &v)
+	if err != nil {
+		return v, err
+	}
+
+	if v.Version == "2.0" {
+		for i, _ := range v.ItemList {
+			if v.ItemList[i].Content != "" {
+				v.ItemList[i].Description = v.ItemList[i].Content
+			}
+		}
+		return v, nil
+	}
+
+	return v, errors.New("not RSS 2.0")
+}
diff --git a/app/rss/feed_test.go b/app/rss/feed_test.go
new file mode 100644
index 0000000..7001ace
--- /dev/null
+++ b/app/rss/feed_test.go
@@ -0,0 +1,34 @@
+package rss
+
+import (
+	"brainbaking.com/go-jamming/common"
+	"github.com/stretchr/testify/assert"
+	"testing"
+	"time"
+)
+
+func TestPubDateAsTimeIncorrectRevertsToNow(t *testing.T) {
+	common.Now = func() time.Time {
+		return time.Date(2020, time.January, 1, 12, 30, 0, 0, time.UTC)
+	}
+	itm := Item{
+		PubDate: "frutselbolletjes",
+	}
+
+	theTime := itm.PubDateAsTime()
+	assert.Equal(t, 2020, theTime.Year())
+	assert.Equal(t, time.January, theTime.Month())
+}
+
+func TestPubDateAsTime(t *testing.T) {
+	itm := Item{
+		PubDate: "Tue, 16 Mar 2021 17:07:14 +0000",
+	}
+	theTime := itm.PubDateAsTime()
+	assert.Equal(t, 2021, theTime.Year())
+	assert.Equal(t, time.March, theTime.Month())
+	assert.Equal(t, 16, theTime.Day())
+	assert.Equal(t, 17, theTime.Hour())
+	assert.Equal(t, 7, theTime.Minute())
+	assert.Equal(t, 14, theTime.Second())
+}
diff --git a/app/webmention/send/rsslinkcollector.go b/app/webmention/send/rsslinkcollector.go
new file mode 100644
index 0000000..e69b16c
--- /dev/null
+++ b/app/webmention/send/rsslinkcollector.go
@@ -0,0 +1,52 @@
+package send
+
+import (
+	"brainbaking.com/go-jamming/app/rss"
+	"time"
+)
+
+type RSSItem struct {
+	link  string
+	hrefs []string
+}
+
+/**
+* a typical RSS item looks like this:
+-- if <time/> found in body, assume it's a lastmod update timestamp!
+ {
+    title: '@celia @kev I have read both you and Kev&#39;s post on...',
+    link: 'https://brainbaking.com/notes/2021/03/16h17m07s14/',
+    comments: 'https://brainbaking.com/notes/2021/03/16h17m07s14/#commento',
+    pubDate: 'Tue, 16 Mar 2021 17:07:14 +0000',
+    author: 'Wouter Groeneveld',
+    guid: {
+      '#text': 'https://brainbaking.com/notes/2021/03/16h17m07s14/',
+      '@_isPermaLink': 'true'
+    },
+    description: ' \n' +
+      '          \n' +
+      '\n' +
+      '          <p><span class="h-card"><a class="u-url mention" data-user="A5GVjIHI6MH82H6iLQ" href="https://fosstodon.org/@celia" rel="ugc">@<span>celia</span></a></span> <span class="h-card"><a class="u-url mention" data-user="A54b8g0RBaIgjzczMu" href="https://fosstodon.org/@kev" rel="ugc">@<span>kev</span></a></span> I have read both you and Kev&rsquo;s post on this and agree on some points indeed! But I&rsquo;m not yet ready to give up webmentions. As an academic, the idea of citing/mentioning each other is very alluring 🤓. Plus, I needed an excuse to fiddle some more with JS&hellip; <br><br>As much as I loved using Wordpress before, I can&rsquo;t imagine going back to writing stuff in there instead of in markdown. Gotta keep the workflow short, though. Hope it helps you focus on what matters - content!</p>\n' +
+      '\n' +
+      '\n' +
+      '          <p>\n' +
+      '            By <a href="/about">Wouter Groeneveld</a> on <time datetime='2021-03-20'>20 March 2021</time>.\n' +
+      '          </p>\n' +
+      '          '
+  }
+**/
+func Collect(xml string, since time.Time) ([]RSSItem, error) {
+	feed, err := rss.ParseFeed([]byte(xml))
+	if err != nil {
+		return nil, err
+	}
+	var items []RSSItem
+	for _, rssitem := range feed.ItemList {
+		if since.Before(rssitem.PubDateAsTime()) {
+			items = append(items, RSSItem{
+				link: rssitem.Link,
+			})
+		}
+	}
+	return items, nil
+}
diff --git a/app/webmention/send/rsslinkcollector_test.go b/app/webmention/send/rsslinkcollector_test.go
new file mode 100644
index 0000000..cc31ef9
--- /dev/null
+++ b/app/webmention/send/rsslinkcollector_test.go
@@ -0,0 +1,62 @@
+package send
+
+import (
+	"brainbaking.com/go-jamming/common"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/suite"
+	"io/ioutil"
+	"testing"
+)
+
+type CollectSuite struct {
+	suite.Suite
+	xml string
+}
+
+func (s *CollectSuite) SetupTest() {
+	file, _ := ioutil.ReadFile("../../../mocks/samplerss.xml")
+	s.xml = string(file)
+}
+
+func TestCollectSuite(t *testing.T) {
+	suite.Run(t, new(CollectSuite))
+}
+
+func (s *CollectSuite) TestCollectShouldNotContainHrefsFromBlockedDomains() {
+	items, err := Collect(s.xml, common.IsoToTime("2021-03-10T00:00:00.000Z"))
+	assert.NoError(s.T(), err)
+	last := items[len(items)-1]
+	assert.Equal(s.T(), "https://brainbaking.com/notes/2021/03/10h16m24s22/", last.link)
+	/*
+		assert.Equal(s.T(), []string{
+			"https://dog.estate/@eli_oat",
+			"https://twitter.com/olesovhcom/status/1369478732247932929",
+			"/aobut",
+		}, last.hrefs)
+
+	*/
+}
+
+func (s *CollectSuite) TestCollectShouldNotContainHrefsThatPointToImages() {
+
+}
+
+func (s *CollectSuite) TestCollectIfTimeTagFoundInContextThatActsAsAnUpdateStamp() {
+
+}
+
+func (s *CollectSuite) TestCollectsNotIfTimeTagFoundInContextButStillOlderThanSince() {
+
+}
+
+func (s *CollectSuite) TestCollectNothingIfDateInFutureAndSinceNothingNewInFeed() {
+
+}
+
+func (s *CollectSuite) TestCollectLatestXLinksWhenASinceParameterIsProvided() {
+
+}
+
+func (s *CollectSuite) TestCollectEveryExternalLinkWithoutAValidSinceDate() {
+
+}
diff --git a/app/webmention/send/send.go b/app/webmention/send/send.go
index 6254c31..d5d0d75 100644
--- a/app/webmention/send/send.go
+++ b/app/webmention/send/send.go
@@ -6,6 +6,7 @@ import (
 	"brainbaking.com/go-jamming/common"
 	"brainbaking.com/go-jamming/rest"
 	"github.com/rs/zerolog/log"
+	"time"
 )
 
 type Sender struct {
@@ -15,6 +16,17 @@ type Sender struct {
 
 func (snder *Sender) Send(domain string, since string) {
 	log.Info().Str("domain", domain).Str("since", since).Msg(` OK: someone wants to send mentions`)
+	feed, err := snder.RestClient.GetBody("https://" + domain + "/index.xml")
+	if err != nil {
+		log.Err(err).Str("domain", domain).Msg("Unable to retrieve RSS feed, aborting send")
+		return
+	}
+
+	snder.parseRssFeed(feed, common.IsoToTime(since))
+}
+
+func (snder *Sender) parseRssFeed(feed string, since time.Time) {
+
 }
 
 func mention() {
diff --git a/common/time.go b/common/time.go
index 7f40d01..d8aee82 100644
--- a/common/time.go
+++ b/common/time.go
@@ -5,3 +5,13 @@ import "time"
 // https://labs.yulrizka.com/en/stubbing-time-dot-now-in-golang/
 // None of the above are very appealing. For now, just use the lazy way.
 var Now = time.Now
+
+// since should be in ISO String format, as produced by clients using day.js - e.g. 2021-04-09T15:51:43.732Z
+func IsoToTime(since string) time.Time {
+	layout := "2006-01-02T15:04:05.000Z"
+	t, err := time.Parse(layout, since)
+	if err != nil {
+		return Now()
+	}
+	return t
+}
diff --git a/common/time_test.go b/common/time_test.go
new file mode 100644
index 0000000..d2cceea
--- /dev/null
+++ b/common/time_test.go
@@ -0,0 +1,45 @@
+package common
+
+import (
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/suite"
+	"testing"
+	"time"
+)
+
+type TimeSuite struct {
+	suite.Suite
+	nowtime time.Time
+}
+
+func (s *TimeSuite) SetupTest() {
+	s.nowtime = time.Date(2020, time.January, 1, 12, 30, 0, 0, time.UTC)
+	Now = func() time.Time {
+		return s.nowtime
+	}
+}
+
+func TestSendSuite(t *testing.T) {
+	suite.Run(t, new(TimeSuite))
+}
+
+func (s *TimeSuite) TestIsoToTimeInISOString() {
+	expectedtime := time.Date(2021, time.March, 9, 15, 51, 43, 732, time.UTC)
+	since := IsoToTime("2021-03-09T15:51:43.732Z")
+	assert.Equal(s.T(), expectedtime.Year(), since.Year())
+	assert.Equal(s.T(), expectedtime.Month(), since.Month())
+	assert.Equal(s.T(), expectedtime.Day(), since.Day())
+	assert.Equal(s.T(), expectedtime.Hour(), since.Hour())
+	assert.Equal(s.T(), expectedtime.Minute(), since.Minute())
+	assert.Equal(s.T(), expectedtime.Second(), since.Second())
+}
+
+func (s *TimeSuite) TestIsoToTimeInvalidStringReturnsNow() {
+	since := IsoToTime("woef ik ben een hondje")
+	assert.Equal(s.T(), s.nowtime, since)
+}
+
+func (s *TimeSuite) TestIsoToTimeEmptyReturnsNow() {
+	since := IsoToTime("")
+	assert.Equal(s.T(), s.nowtime, since)
+}