apicollective / apibuilder-cli Goto Github PK

needs to run rspec test/spec/lib/*

Hi guys,
I'm trying to add a variable in the YAML file in order to reuse it in some points in the same file.

code:
  org:
    service-one:
      version: *version
      generators:
        dart_2_17_json:
          target: source

Running the command apibuilder update --path .apibuilder.yml it works running it on a macOS with ruby version 2.6.10 but it doesn't work running it on a macOS with ruby version 3.1.1.

/Users/federico/.rbenv/versions/3.1.1/lib/ruby/3.1.0/psych/visitors/to_ruby.rb:430:in `visit_Psych_Nodes_Alias': Unknown alias: version (Psych::BadAlias)

Searching I found this possible issue but I'm not sure.
https://stackoverflow.com/questions/71191685/visit-psych-nodes-alias-unknown-alias-default-psychbadalias

Do you have any idea how I can solve it keeping the latest ruby version?
Thanks

when you run apibuilder update --app <whatever> it removes any other generated files from .tracked_files. only the file that was just updated remains tracked

When running apibuilder update with a config like below (note the same api declared with different versions) only the files for the version declared latest are downloaded:

my-api:
  version: 1.0.3
  generators:
    - ...
my-api:
  version: 2.0.0
  generators:
    - ...

$ brew install apibuilder-cli
==> Downloading https://github.com/apicollective/apibuilder-cli/archive/0.1.25.tar.gz
Already downloaded: /Users/gary/Library/Caches/Homebrew/apibuilder-cli-0.1.25.tar.gz
==> ./install.sh /usr/local/Cellar/apibuilder-cli/0.1.25
🍺  /usr/local/Cellar/apibuilder-cli/0.1.25: 19 files, 305.6KB, built in 2 seconds

$ apibuilder
/usr/local/bin/apibuilder:68:in `load': cannot load such file -- /usr/local/bin/../src/apibuilder-cli.rb (LoadError)
	from /usr/local/bin/apibuilder:68:in `<main>'

There seem to be two loads of the above file:

load File.join('/usr/local/Cellar/apibuilder-cli/0.1.25/src/apibuilder-cli.rb')
#  - Upgrade the CLI to the latest version
#

load File.join(File.dirname(__FILE__), '../src/apibuilder-cli.rb')

• enable ability to ask users to update their CLI

How can I list more than one directory for the input?
I want to tell apidoc to look in 2 directories, for example:

apidoc --input dir1/ dir2/

e.g:

---
flow:
  api:
    ruby_client:
    - "/web/flowmisc/billing/lib/flow_commerce/flow_api_v0_client.rb"
  api-internal:
    ruby_client:
    - "/web/flowmisc/billing/lib/flow_commerce/flow_api_internal_v0_client.rb"

So if a different user has a different project location, this might not delete all files, or might delete extra files that are needed.

Currently it does not seem to be possible to download just a single file from a given generator. See e.g. https://app.apibuilder.io/apicollective/apibuilder-generator/0.12.58/http4s_0_15
With the http4s generator I will always only need one of the generated files, never all. Sepcifying the following in the .apibuilder file gives the error below:

      generators:
        http4s_0_15: src/main/scala/generated/ApicollectiveApibuilderGeneratorV0ModelsJson.scala

Fetching code from https://api.apibuilder.io
  apicollective/apibuilder-generator/latest/http4s_0_15...
/Users/gheine/.rvm/rubies/ruby-2.3.1/lib/ruby/2.3.0/fileutils.rb:253:in `mkdir': File exists @ dir_s_mkdir - /Code/src/main/scala/generated/ApicollectiveApibuilderGeneratorV0ModelsJson.scala (Errno::EEXIST)
	from /Users/gheine/.rvm/rubies/ruby-2.3.1/lib/ruby/2.3.0/fileutils.rb:253:in `fu_mkdir'
	from /Users/gheine/.rvm/rubies/ruby-2.3.1/lib/ruby/2.3.0/fileutils.rb:227:in `block (2 levels) in mkdir_p'
	from /Users/gheine/.rvm/rubies/ruby-2.3.1/lib/ruby/2.3.0/fileutils.rb:225:in `reverse_each'
	from /Users/gheine/.rvm/rubies/ruby-2.3.1/lib/ruby/2.3.0/fileutils.rb:225:in `block in mkdir_p'
	from /Users/gheine/.rvm/rubies/ruby-2.3.1/lib/ruby/2.3.0/fileutils.rb:211:in `each'
	from /Users/gheine/.rvm/rubies/ruby-2.3.1/lib/ruby/2.3.0/fileutils.rb:211:in `mkdir_p'
	from /web/apidoc-cli/bin/apibuilder:371:in `block (4 levels) in <main>'
	from /web/apidoc-cli/bin/apibuilder:365:in `each'
	from /web/apidoc-cli/bin/apibuilder:365:in `block (3 levels) in <main>'
	from /web/apidoc-cli/bin/apibuilder:360:in `each'
	from /web/apidoc-cli/bin/apibuilder:360:in `block (2 levels) in <main>'
	from /web/apidoc-cli/bin/apibuilder:359:in `each'
	from /web/apidoc-cli/bin/apibuilder:359:in `block in <main>'
	from /web/apidoc-cli/bin/apibuilder:358:in `map'
	from /web/apidoc-cli/bin/apibuilder:358:in `<main>'

It's unclear to me which syntax the CLI wants me to use to filter files. It looks like it wants to support globs, but the implementation to turn a glob into a regular expression feels incomplete. At least, I could not find a way to exclude a single file by name. Since the patterns that can be used are poorly documented and feel limited by the implementation, I propose the file filter should accept a regular expression without substitution. In my case, this would allow me to exclude a single file using a negative lookahead ^(?!filename).*\.js$.

Basically, I'm proposing this line is changed from:

file_regex = Regexp::new("^" + filename.gsub('.', '\.').gsub('*', '[^.]*').gsub('?', '[^.]?') + "$")

to:

file_regex = Regexp::new(filename)

Need a way to download a zip file, for example from here: http://apidoc.me/gilt/checkout-api/latest/android_client (generator takes ~2 minutes to warm up on lambda, so if you get an error on that page - try again after ~2 minutes)

The output of running apidoc in terminal is as follows:

$ apidoc                                                                                                               [14:16:50]
** ERROR: Missing command is required. Available commands:
 - cli
 - code
 - download
 - example
 - list
 - update
 - upload

However, only the following are documented: list, code, and update. Running apidoc upload yields a cryptic error to anyone unfamiliar with the tool:

$ apidoc upload                                                                                                        [14:17:40]
org, application, path are required

It would be great if apidoc --help returned some more information w/ specific examples.

after brew install apibuilder-cli:
When running /opt/homebrew/bin/create-config I get
/opt/homebrew/bin/create-config:9:in load': cannot load such file -- /opt/homebrew/bin/../src/apibuilder-cli.rb (LoadError) from /opt/homebrew/bin/create-config:9:in

Also it seems like apibuilder-cli wasn't added to my path after brew install

For example:

http://apidoc.me/gilt/api-user-registration/0.0.2/play_2_x_routes used to return contents of the routes file, and with it this command...

$ /web/apidoc-cli/bin/apidoc code gilt api-user-registration 0.0.2 play_2_x_routes

... used to pull contents of routes.

But now that generators could contain multiple files, routes live at: http://apidoc.me/gilt/api-user-registration/0.0.2/play_2_x_routes/routes

and, seemingly, with this change cli command returns a listing of files

$ /web/apidoc-cli/bin/apidoc code gilt api-user-registration 0.0.2 play_2_x_routes
play_2_x_routes Files:
 - routes

Or does it always have to connect to the website?

According to the apibuilder documentation, you should be able to use the env var APIBUILDER_API_URI to override the API that is hit

when executing commands - e.g. list applications gilt - if there is another page, maybe add a prompt or a note that there are more with instructions for how to get more

When I enable directory creation in the global settings, the CLI exits with no implicit conversion of nil into String for generators that do not respond with a recommended directory path. A recommended directory is an optional field of the file model, thereby the CLI should gracefully fallback to no directory creation.

settings:
  code.create.directories: true