Stream `extra_metrics` fails on repos with large number of issues/PRs

[laurentS]

When running the tap on https://github.com/microsoft/TypeScript with the extra_metrics stream, it crashes because the number of open issues is shown as 5k+ on the project page.

When navigating to https://github.com/microsoft/TypeScript/issues the actual number is 5988 (as of writing this), so really closer to 6k.

Stack trace:

File "tap_github/repository_streams.py", line 2071, in parse_response
     yield from scrape_metrics(response, self.logger)
                ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "tap_github/scraping.py", line 126, in scrape_metrics
     issues = parse_counter(soup.find("span", id="issues-repo-tab-count"), logger)
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "tap_github/scraping.py", line 109, in parse_counter
     return int(title_string.strip().replace(",", ""))
            ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
ValueError: invalid literal for int() with base 10: '5000+'

It would make sense to source the number of open issues and PRs from the graphql api endpoint instead.
I will open a PR to fix this.

[laurentS]

Dependents and contributors counts are included in the same stream, but are not available through the API, as a temp fix, I'll make sure the 5k+ token does not crash the tap and returns something. Not ideal, but should allow the tap to keep running.

For the record, the following graphql query:

{
  repository(name: "TypeScript", owner: "microsoft") {
    issues(states: OPEN) {
      totalCount
    }
    pullRequests(states: OPEN) {
      totalCount
    }
  }
}

provides the number of open issues and PRs as shown on the web page.