15 minute fix: Shorten long titles from feed items if needed (#13961)

* Shorten long titles from feed items if needed

Fixes #13959

When the title is too long to create a valid article, put a shortened
title in the generated article markdown.

Added unit tests to cover the default case and the shortening case

* Fix Struct initialization

Rubocop suggested struct over openstruct because of performance, but
Struct.new creates a "class" with a positional initializer, while
Openstruct's new takes a hash like keyword set and builds an object
from those directly.

I mistakenly reused the openstruct syntax after the conversion. This
changes the feed and item to use Struct's initializer.

* Use String#truncate in place of custom trimming

`separator` enables a natural break (preventing space between last
character and ellipsis), and omission text "..." is the default behavior.

* Use verifying doubles in test, use method access to members in code

The instance_double for Feedjira::Parser::RSSEntry
necessitated changing hash key access to reader method access for the
data members (title, categories, summary, description).
This commit is contained in:
Daniel Uber 2021-06-11 12:51:51 -05:00 committed by GitHub
parent 26592a3a9c
commit b4ce4ec975
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 58 additions and 4 deletions

View file

@ -6,8 +6,8 @@ module Feeds
def initialize(item, user, feed, feed_source_url)
@item = item
@title = item[:title].strip
@categories = item[:categories] || []
@title = item.title.strip
@categories = item.categories || []
@user = user
@feed = feed
@feed_source_url = feed_source_url
@ -16,7 +16,7 @@ module Feeds
def call
body = <<~HEREDOC
---
title: #{@title}
title: #{processed_title}
published: false
date: #{@item.published}
tags: #{get_tags}
@ -31,6 +31,10 @@ module Feeds
private
def processed_title
@title.truncate(128, omission: "...", separator: " ")
end
def get_tags
@categories.first(4).map do |tag|
tag.delete(" ").gsub(/[^[:alnum:]]/i, "")[0..19]
@ -54,7 +58,7 @@ module Feeds
end
def get_content
@item[:content] || @item[:summary] || @item[:description]
@item.content || @item.summary || @item.description
end
def thorough_parsing(content, feed_url)

View file

@ -0,0 +1,50 @@
require "rails_helper"
RSpec.describe Feeds::AssembleArticleMarkdown, type: :service do
let(:user) { create(:user, feed_mark_canonical: true) }
let(:feed_source_url) { "https://feed.source/url" }
let(:feed) { instance_double("Feedjira::Parser::RSS", url: "https://feed.source/") }
let(:title) { "A title" }
let(:content) { "Some content that came in with the item, should be the body" }
let(:item) do
instance_double(
"Feedjira::Parser::RSSEntry",
title: title,
categories: %w[tag1 tag2 tag3 tag4 tag5],
published: "2020-12-20",
content: content,
)
end
let(:feeds_assemble_article_markdown) { described_class.new(item, user, feed, feed_source_url) }
it "creates markdown head matter" do
body = feeds_assemble_article_markdown.call
expect(body).to include("title: A title")
expect(body).to include("date: 2020-12-20")
end
it "includes the content from the feed" do
body = feeds_assemble_article_markdown.call
expect(body).to include(content)
end
context "when item has a long title" do
let(:title) { "Words " * 25 }
it "limits the title to 128 characters by truncation" do
body = feeds_assemble_article_markdown.call
title_line = body.lines.detect { |line| line.starts_with?("title:") }
shortened_title = title_line.split(":").second.strip
# we expect 21 "Words" plus an ellipsis, total of 128 characters
expected_title = "Words Words Words Words Words Words Words Words Words Words Words Words Words Words Words Words Words Words Words Words Words..." # rubocop:disable Layout/LineLength
expect(shortened_title.size).to eq(128)
expect(shortened_title).to eq(expected_title)
end
end
end