RSpec mocking an :each block - ruby

I want to use RSpec mocks to provide canned input to a block.
Ruby:
class Parser
attr_accessor :extracted
def parse(fname)
File.open(fname).each do |line|
extracted = line if line =~ /^RCS file: (.*),v$/
end
end
end
RSpec:
describe Parser
before do
#parser = Parser.new
#lines = mock("lines")
#lines.stub!(:each)
File.stub!(:open).and_return(#lines)
end
it "should extract a filename into extracted" do
linetext = [ "RCS file: hello,v\n", "bla bla bla\n" ]
# HELP ME HERE ...
# the :each should be fed with 'linetext'
#lines.should_receive(:each)
#parser.should_receive('extracted=')
#parser.parse("somefile.txt")
end
end
It's a way to test that the internals of the block work correctly by passing fixtured data into it. But I can't figure out how to do the actual feeding with RSpec mocking mechanism.
update: looks like the problem was not with the linetext, but with the:
#parser.should_receive('extracted=')
it's not the way it's called, replacing it in the ruby code with self.extracted= helps a bit, but feels wrong somehow.

To flesh out the how 'and_yield' works: I don't think 'and_return' is really what you want here. That will set the return value of the File.open block, not the lines yielded to its block. To change the example slightly, say you have this:
Ruby
def parse(fname)
lines = []
File.open(fname){ |line| lines << line*2 }
end
Rspec
describe Parser do
it 'should yield each line' do
File.stub(:open).and_yield('first').and_yield('second')
parse('nofile.txt').should eq(['firstfirst','secondsecond'])
end
end
Will pass. If you replaced that line with an 'and_return' like
File.stub(:open).and_return(['first','second'])
It will fail because the block is being bypassed:
expected: ["firstfirst", "secondsecond"]
got: ["first", "second"]
So bottom line is use 'and_yield' to mock the input to 'each' type blocks. Use 'and_return' to mock the output of those blocks.

I don't have a computer with Ruby & RSpec available to check this, but I suspect you need to add a call to and_yields call [1] on the end of the should_receive(:each). However, you might find it simpler not to use mocks in this case e.g. you could return a StringIO instance containing linetext from the File.open stub.
[1] http://rspec.rubyforge.org/rspec/1.1.11/classes/Spec/Mocks/BaseExpectation.src/M000104.html

I would go with the idea of stubbing the File.open call
lines = "RCS file: hello,v\n", "bla bla bla\n"
File.stub!(:open).and_return(lines)
This should be good enough to test the code inside the loop.

This should do the trick:
describe Parser
before do
#parser = Parser.new
end
it "should extract a filename into extracted" do
linetext = [ "RCS file: hello,v\n", "bla bla bla\n" ]
File.should_receive(:open).with("somefile.txt").and_return(linetext)
#parser.parse("somefile.txt")
#parser.extracted.should == "hello"
end
end
There are some bugs in the Parser class (it won't pass the test), but that's how I'd write the test.

Related

Stub a CSV file in Ruby Test::Unit

So I have a Ruby script which parses a CSV file as a list of rules and does some processing and returns an array of hashes as processed data.
CSV looks like this:
id_number,rule,notes
1,"dummy_rule","necessary"
2,"sample_rule","optional"
The parsing of CSV looks like this:
def parse_csv(file_name)
filtered_data = []
CSV.foreach(file_name, headers: true) do |row|
filtered_data << row # some processing
end
filtered_data
end
Now I am wondering if it is possible to stub/mock an actual CSV file for unit testing in such a way that I could pass a "filename" into this function. I could make a CSV object and use the generate function but then the actual filename would not be possible.
def test_parse_csv
expected_result = ["id_number"=>1, "rule"=>"dummy_rule", "notes"=>"optional"}]
# stub/mock csv with filename: say "rules.csv"
assert.equal(expected_result, parse_csv(file_name))
end
I use Test::Ruby
I also found a ruby gem library called mocha but I don't know how this works for CSVs
https://github.com/freerange/mocha
Any thoughts are welcome!
I would create a support directory inside your test directory then create a file..
# test/support/rules.csv
id_number,rule,notes
1,"dummy_rule","necessary"
2,"sample_rule","optional"
Your test should look like this, also adding an opening curly bracket which looks like you've missed on line #2.
def test_parse_csv
expected_result = [{"id_number"=>1, "rule"=>"dummy_rule", "notes"=>"optional"}]
assert.equal(expected_result, parse_csv(File.read(csv_file)).first)
end
def csv_file
Rails.root.join('test', 'support', 'rules.csv')
end
EDIT: Sorry, I should have noticed this wasn't Ruby on Rails... heres a ruby solution:
def test_parse_csv
expected_result = [{"id_number"=>1, "rule"=>"dummy_rule", "notes"=>"optional"}]
assert.equal(expected_result, parse_csv(csv_file).first)
end
def csv_file
File.read('test/support/rules.csv')
end
I did this in my test let(:raw_csv) { "row_number \n 1 \n 2" }, telling myself that a CSV file was surely read as a simple string with linebreaks and commas... and it works fine. Be careful with additional whitespaces and commas, it is easy to make a mistake.

Minitest: How to stub/mock the file result of Kernel.open on a URL

I have been trying to use Minitest to test my code (full repo) but am having trouble with one method which downloads a SHA1 hash from a .txt file on a website and returns the value.
Method:
def download_remote_sha1
#log.info('Downloading Elasticsearch SHA1.')
#remote_sha1 = ''
Kernel.open(#verify_url) do |file|
#remote_sha1 = file.read
end
#remote_sha1 = #remote_sha1.split(/\s\s/)[0]
#remote_sha1
end
You can see that I log what is occurring to the command line, create an object to hold my SHA1 value, open the url (e.g. https://download.elasticsearch.org/elasticsearch/elasticsearch/elasticsearch-1.4.2.deb.sha1.txt)
I then split the string so that I only have the SHA1 value.
The problem is that during a test, I want to stub the Kernel.open which uses OpenURI to open the URL. I would like to ensure that I'm not actually reaching out to download any file, but rather I'm just passing the block my own mock IO object testing just that it correctly splits stuff.
I attempted it like the block below but when #remote_sha1 = file.read occurs the file item is nil.
#mock_file = Minitest::Mock.new
#mock_file.expect(:read, 'd377e39343e5cc277104beee349e1578dc50f7f8 elasticsearch-1.4.2.deb')
Kernel.stub :open, #mock_file do
#downloader = ElasticsearchUpdate::Downloader.new(hash, true)
#downloader.download_remote_sha1.must_equal 'd377e39343e5cc277104beee349e1578dc50f7f8'
end
I was working on this question too, but matt figured it out first. To add to what matt posted:
When you write:
Kernel.stub(:open, #mock_file) do
#block code
end
...that means when Kernel.open() is called--in any code, anywhere before the stub() block ends--the return value of Kernel.open() will be #mock_file. However, you never use the return value of Kernel.open() in your code:
Kernel.open(#verify_url) do |f|
#remote_sha1 = f.read
end
If you wanted to use the return value of Kernel.open(), you would have to write:
return_val = Kernel.open(#verify_url) do |f|
#remote_sha1 = f.read
end
#do something with return_val
Therefore, the return value of Kernel.open() is irrelevant in your code--which means the second argument of stub() is irrelevant.
A careful examination of the source code for stub() reveals that stub() takes a third argument--an argument which will be passed to a block specified after the stubbed method call. You, in fact, have specified a block after your stubbed Kernel.open() method call:
stubbed method call -+ +- start of block
| | |
V V V
Kernel.open(#verify_url) do |f|
#remote_sha1 = f.read
end
^
|
end of block
So, in order to pass #mockfile to the block you need to specify it as the third argument to Kernel.stub():
Kernel.stub(:open, 'irrelevant', #mock_file) do
end
Here is a full example for future searchers:
require 'minitest/autorun'
class Dog
def initialize
#verify_url = 'http://www.google.com'
end
def download_remote_sha1
#remote_sha1 = ''
Kernel.open(#verify_url) do |f|
#remote_sha1 = f.read
end
#puts #remote_sha1[0..300]
#remote_sha1 = #remote_sha1.split(" ")[0] #Using a single space for the split() pattern will split on contiguous whitespace.
end
end
#Dog.new.download_remote_sha1
describe 'downloaded file' do
it 'should be an sha1 code' do
#mock_file = Minitest::Mock.new
#mock_file.expect(:read, 'd377e39343e5cc277104beee349e1578dc50f7f8 elasticsearch-1.4.2.deb')
Kernel.stub(:open, 'irrelevant', #mock_file) do
#downloader = Dog.new
#downloader.download_remote_sha1.must_equal 'd377e39343e5cc277104beee349e1578dc50f7f8'
end
end
end
xxx
The second argument to stub is what you want the return value to be for the duration of your test, but the way Kernel.open is used here requires the value it yields to the block to be changed instead.
You can achieve this by providing a third argument. Try changing the call to Kernel.stub to
Kernel.stub :open, true, #mock_file do
#...
Note the extra argument true, so that #mock_file is now the third argument and will be yielded to the block. The actual value of the second argument doesn’t really matter in this case, you might want to use #mock_file there too to more closely correspond to how open behaves.

How do I test reading a file?

I'm writing a test for one of my classes which has the following constructor:
def initialize(filepath)
#transactions = []
File.open(filepath).each do |line|
next if $. == 1
elements = line.split(/\t/).map { |e| e.strip }
transaction = Transaction.new(elements[0], Integer(1))
#transactions << transaction
end
end
I'd like to test this by using a fake file, not a fixture. So I wrote the following spec:
it "should read a file and create transactions" do
filepath = "path/to/file"
mock_file = double(File)
expect(File).to receive(:open).with(filepath).and_return(mock_file)
expect(mock_file).to receive(:each).with(no_args()).and_yield("phrase\tvalue\n").and_yield("yo\t2\n")
filereader = FileReader.new(filepath)
filereader.transactions.should_not be_nil
end
Unfortunately this fails because I'm relying on $. to equal 1 and increment on every line and for some reason that doesn't happen during the test. How can I ensure that it does?
Global variables make code hard to test. You could use each_with_index:
File.open(filepath) do |file|
file.each_with_index do |line, index|
next if index == 0 # zero based
# ...
end
end
But it looks like you're parsing a CSV file with a header line. Therefore I'd use Ruby's CSV library:
require 'csv'
CSV.foreach(filepath, col_sep: "\t", headers: true, converters: :numeric) do |row|
#transactions << Transaction.new(row['phrase'], row['value'])
end
You can (and should) use IO#each_line together with Enumerable#each_with_index which will look like:
File.open(filepath).each_line.each_with_index do |line, i|
next if i == 1
# …
end
Or you can drop the first line, and work with others:
File.open(filepath).each_line.drop(1).each do |line|
# …
end
If you don't want to mess around with mocking File for each test you can try FakeFS which implements an in memory file system based on StringIO that will clean up automatically after your tests.
This way your test's don't need to change if your implementation changes.
require 'fakefs/spec_helpers'
describe "FileReader" do
include FakeFS::SpecHelpers
def stub_file file, content
FileUtils.mkdir_p File.dirname(file)
File.open( file, 'w' ){|f| f.write( content ); }
end
it "should read a file and create transactions" do
file_path = "path/to/file"
stub_file file_path, "phrase\tvalue\nyo\t2\n"
filereader = FileReader.new(file_path)
expect( filereader.transactions ).to_not be_nil
end
end
Be warned: this is an implementation of most of the file access in Ruby, passing it back onto the original method where possible. If you are doing anything advanced with files you may start running into bugs in the FakeFS implementation. I got stuck with some binary file byte read/write operations which weren't implemented in FakeFS quite how Ruby implemented them.

Mock file input as file path on Rspec

I have a question on how to use rspec to mock a file input. I have a following code for the class, but not exactly know a why to mock a file input. filepath is /path/to/the/file
I did my search on Google and usually turns out to be loading the actual file instead of mocking, but I'm actually looking the opposite where only mock, but not using the actual file.
module Service
class Signing
def initialize(filepath)
#config = YAML.load_file(filepath)
raise "Missing config file." if #config.nil?
end
def sign() …
end
private
def which() …
end
end
end
Is it possible to use EOF delimiter for this file input mocking?
file = <<EOF
A_NAME: ABC
A_ALIAS: my_alias
EOF
You could stub out YAML.load_file and return parsed YAML from your text, like this:
yaml_text = <<-EOF
A_NAME: ABC
A_ALIAS: my_alias
EOF
yaml = YAML.load(yaml_text)
filepath = "bogus_filename.yml"
YAML.stub(:load_file).with(filepath).and_return(yaml)
This doesn't quite stub out the file load itself, but to do that you'd have to make assumptions about what YAML.load_file does under the covers, and that's not a good idea. Since it's safe to assume that the YAML implementation is already tested, you can use the code above to replace the entire call with your parsed-from-text fixture.
If you want to test that the correct filename is passed to load_file, replace the stub with an expectation:
YAML.should_receive(:load_file).with(filepath).and_return(yaml)
If the idea is to put an expectation on something, I don't see much benefit on this approach of calling YAML.load to fake the return. YAML.load_file actually returns a hash, so instead of doing all that my suggestion would be to simply return a hash:
parsed_yaml = {
"somekey" => {
"someotherkey" => "abc"
}
}
YAML.should_receive(:load_file).with(filepath).and_return(parsed_yaml)
As this is supposed to be a unit test and not an integration test, I think this would make more sense.

RSpec: how to test file operations and file content

In my app, I have the following code:
File.open "filename", "w" do |file|
file.write("text")
end
I want to test this code via RSpec. What are the best practices for doing this?
I would suggest using StringIO for this and making sure your SUT accepts a stream to write to instead of a filename. That way, different files or outputs can be used (more reusable), including the string IO (good for testing)
So in your test code (assuming your SUT instance is sutObject and the serializer is named writeStuffTo:
testIO = StringIO.new
sutObject.writeStuffTo testIO
testIO.string.should == "Hello, world!"
String IO behaves like an open file. So if the code already can work with a File object, it will work with StringIO.
For very simple i/o, you can just mock File. So, given:
def foo
File.open "filename", "w" do |file|
file.write("text")
end
end
then:
describe "foo" do
it "should create 'filename' and put 'text' in it" do
file = mock('file')
File.should_receive(:open).with("filename", "w").and_yield(file)
file.should_receive(:write).with("text")
foo
end
end
However, this approach falls flat in the presence of multiple reads/writes: simple refactorings which do not change the final state of the file can cause the test to break. In that case (and possibly in any case) you should prefer #Danny Staple's answer.
This is how to mock File (with rspec 3.4), so you could write to a buffer and check its content later:
it 'How to mock File.open for write with rspec 3.4' do
#buffer = StringIO.new()
#filename = "somefile.txt"
#content = "the content fo the file"
allow(File).to receive(:open).with(#filename,'w').and_yield( #buffer )
# call the function that writes to the file
File.open(#filename, 'w') {|f| f.write(#content)}
# reading the buffer and checking its content.
expect(#buffer.string).to eq(#content)
end
You can use fakefs.
It stubs filesystem and creates files in memory
You check with
File.exists? "filename"
if file was created.
You can also just read it with
File.open
and run expectation on its contents.
For someone like me who need to modify multiple files in multiple directories (e.g. generator for Rails), I use temp folder.
Dir.mktmpdir do |dir|
Dir.chdir(dir) do
# Generate a clean Rails folder
Rails::Generators::AppGenerator.start ['foo', '--skip-bundle']
File.open(File.join(dir, 'foo.txt'), 'w') {|f| f.write("write your stuff here") }
expect(File.exist?(File.join(dir, 'foo.txt'))).to eq(true)
end
end

Resources