Keep Items in sync with their Amazon listing #62
Assignees
Labels
Comments
leesharma
added a commit
to leesharma/playtime
that referenced
this issue
Oct 3, 2017
In preparation for rubyforgood#62, this commit pulls the endpoint-specific information out of the HTTPClient class and into and endpoint-specific class. Now HTTPClient is responsible only for managing the environment information and directing the user to the relevant endpoint. Adding a new endpoint is as simple as adding a method (ex. `item_search`) and corresponding class (ex. `ItemSearchEndpoint). The interface is *not* in its final form yet. This is just a good breaking point for a commit.
leesharma
added a commit
to leesharma/playtime
that referenced
this issue
Oct 25, 2017
In preparation for rubyforgood#62, this commit pulls the endpoint-specific information out of the HTTPClient class and into and endpoint-specific class. Now HTTPClient is responsible only for managing the environment information and directing the user to the relevant endpoint. Adding a new endpoint is as simple as adding a method (ex. `item_search`) and corresponding class (ex. `ItemSearchEndpoint). The interface is *not* in its final form yet. This is just a good breaking point for a commit.
This was referenced Oct 26, 2017
leesharma
added a commit
that referenced
this issue
Nov 29, 2017
* Refactor AmazonProductAPI In preparation for #62, this commit pulls the endpoint-specific information out of the HTTPClient class and into and endpoint-specific class. Now HTTPClient is responsible only for managing the environment information and directing the user to the relevant endpoint. Adding a new endpoint is as simple as adding a method (ex. `item_search`) and corresponding class (ex. `ItemSearchEndpoint). The interface is *not* in its final form yet. This is just a good breaking point for a commit. * Add ItemSearch endpoint Building on the last commit, this adds a new ItemLookup endpoint. Now we can pull details on an individual item based on the ASIN. The endpoints and responses include a lot of duplication. Some refactoring will probably be needed. * Extract spec for ItemSearchEndpoint When the item search endpoint was extracted, the specs weren't extracted with it; this means that the HTTPClientSpec was testing everything relating to the item search endpoint. This commit extracts all specs relating to the endpoint into a new file. This can now be refactored and copied over for the `ItemLookupEndpointSpec`. * Add spec for ItemLookupEndpoint * Fix CodeClimate issue This is a bit of a hack solution, but I don't feel comfortable doing any major abstraction here yet; I don't think we have enough information. Hopefully this'll clear up the CodeClimate complaints about duplicated code! * Fix easy PR-related rubocop issues Two big points: 1. This only fixes easy rubocop issues related to this PR. This doesn't touch any of the new capistrano code; I want the diff for this PR to be fairly contained. 2. This includes two rubocop config changes: i. Exclude the vendor directory from linting ii. Allow multiline braces in tests (for exception checks, let, etc.) There are a few more risky/controversial rubocop changes I omitted from this commit. Those are coming next. * Fix all Lint/UriEscapeUnescape violations Fixes #56 This commit changes all the `URI.escape` calls to `CGI.escape`. All tests pass and the app still works. As a bonus, this seems to resolve issue #56 too-apostrophes are now properly escaped. This resolves all rubocop issues relating to this PR. * Give better name to AWS test credentials in specs Following the review suggestions. This renames `env` to `aws_credentials`. The latter is a better name because it represents a credentials object (built from env vars), not the ENV object itself. Hopefully this will be clearer! * Fix item lookup hash When I fixed the code style for Code Climate in #173, I used the wrong hash leading to no update attributes being found. This fixes the bug, adds a test, and renames some of the variables.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Item prices change frequently on Amazon, so fetching the price once and relying on that data forevermore is a risky strategy; there's a good chance it will get out of date pretty quickly. Same thing for removed/out of stock items.
I'd suggest adding a recurring background task to ping Amazon for changes and some sort of warning that we might be a little out of date on the app.
Any other ideas?
The text was updated successfully, but these errors were encountered: