Fix unnecessary re-fetching of pull request data (singer-io#49)

GitHub’s /pulls endpoint doesn’t have a ?since filter, so instead we
check the updated_at values of each record imported. Since we have
sorted descending by this value, we can break out of the loop once we
have found the first updated_at value that is less than the bookmark
time.

Fixes singer-io#43

tap_github.py

@@ -187,15 +187,30 @@ def get_all_pull_requests(schemas, repo_path, state, mdata):
     '''
     https://developer.github.com/v3/pulls/#list-pull-requests
     '''
+
+    bookmark_value = get_bookmark(state, repo_path, "pull_requests", "since");
+    if bookmark_value:
+        bookmark_time = singer.utils.strptime_to_utc(bookmark_value)
+    else:
+        bookmark_time = 0
+
     with metrics.record_counter('pull_requests') as counter:
         with metrics.record_counter('reviews') as reviews_counter:
             for response in authed_get_all_pages(
                     'pull_requests',
-                    'https://api.github.com/repos/{}/pulls?state=all'.format(repo_path)
+                    'https://api.github.com/repos/{}/pulls?state=all&sort=updated&direction=desc'.format(repo_path)
             ):
                 pull_requests = response.json()
                 extraction_time = singer.utils.now()
                 for pr in pull_requests:
+
+                    # skip records that haven't been updated since the last run
+                    # the GitHub API doesn't currently allow a ?since param for pulls
+                    # once we find the first piece of old data we can return, thanks to
+                    # the sorting
+                    if bookmark_time and singer.utils.strptime_to_utc(pr.get('updated_at')) < bookmark_time:
+                        return state
+
                     pr_num = pr.get('number')
                     pr['_sdc_repository'] = repo_path