From 3607a16fab201125f2799e6d93ab433b4d872880 Mon Sep 17 00:00:00 2001
From: Eric Boucher <eric.p.boucher@gmail.com>
Date: Tue, 17 May 2022 13:38:40 -0400
Subject: [PATCH 1/2] Paginate backwards for stargazers

---
 tap_github/client.py             |  6 ++--
 tap_github/repository_streams.py | 56 ++++++++++++++++++++++++++++++++
 2 files changed, 59 insertions(+), 3 deletions(-)

diff --git a/tap_github/client.py b/tap_github/client.py
index 222f4ba3..9f110743 100644
--- a/tap_github/client.py
+++ b/tap_github/client.py
@@ -77,11 +77,11 @@ def get_next_page_token(
         else:
             results = resp_json.get("items")
 
-        # Exit early if the response has no items. ? Maybe duplicative the "next" link check.
+        # Exit early if the response has no items. ? Maybe duplicative of the "next" link check.
         if not results:
             return None
 
-        # Unfortunately endpoints such as /starred, /stargazers, /events and /pulls do not support
+        # Unfortunately endpoints such as /starred, /events and /pulls do not support
         # the "since" parameter out of the box. So we use a workaround here to exit early.
         # For such streams, we sort by descending dates (most recent first), and paginate
         # "back in time" until we reach records before our "since" parameter.
@@ -132,7 +132,7 @@ def get_url_params(
             params["sort"] = "updated"
             params["direction"] = "desc" if self.missing_since_parameter else "asc"
 
-        # Unfortunately the /starred, /stargazers (starred_at) and /events (created_at) endpoints do not support
+        # Unfortunately the /starred (starred_at) and /events (created_at) endpoints do not support
         # the "since" parameter out of the box. But we use a workaround in 'get_next_page_token'.
         elif self.replication_key in ["starred_at", "created_at"]:
             params["sort"] = "created"
diff --git a/tap_github/repository_streams.py b/tap_github/repository_streams.py
index c05acafa..6871b577 100644
--- a/tap_github/repository_streams.py
+++ b/tap_github/repository_streams.py
@@ -6,6 +6,9 @@
 from singer_sdk import typing as th  # JSON Schema typing helpers
 from singer_sdk.helpers.jsonpath import extract_jsonpath
 
+from dateutil.parser import parse
+from urllib.parse import parse_qs, urlparse
+
 from tap_github.client import GitHubGraphqlStream, GitHubRestStream
 from tap_github.schema_objects import (
     user_object,
@@ -1496,6 +1499,59 @@ def http_headers(self) -> dict:
         headers["Accept"] = "application/vnd.github.v3.star+json"
         return headers
 
+    def get_next_page_token(
+        self, response: requests.Response, previous_token: Optional[Any]
+    ) -> Optional[Any]:
+        """Return a token for identifying next page or None if no more pages.
+
+        /stargazers does not have "since" or "direction" parameters. If a since parameter is passed,
+        we manually paginate backwards and try to exit early.
+
+        WARNING - This does not work for repositories with too many stars like facebook/react.
+        See https://github.com/MeltanoLabs/tap-github/issues/120
+        """
+        # Leverage header links returned by the GitHub API.
+        if not ("next" in response.links.keys() or "prev" in response.links.keys()):
+            return None
+
+        results = response.json()
+        request_parameters = parse_qs(str(urlparse(response.request.url).query))
+
+        # parse_qs interprets "+" as a space, revert this to keep an aware datetime
+        try:
+            since = (
+                request_parameters["since"][0].replace(" ", "+")
+                if "since" in request_parameters
+                else ""
+            )
+        except IndexError:
+            since = ""
+
+        # if "since" is present, paginate backwards.
+        next_page_key = "prev" if since else "next"
+
+        # start with the last page
+        if (previous_token or 1) == 1 and "last" in response.links.keys():
+            next_page_key = "last"
+        # if possible, exit early
+        elif not results or len(results) < self.MAX_PER_PAGE:
+            return None
+        elif since and (parse(results[-1][self.replication_key]) < parse(since)):
+            return None
+
+        # Use header links returned by the GitHub API.
+        parsed_url = urlparse(response.links[next_page_key]["url"])
+        captured_page_value_list = parse_qs(parsed_url.query).get("page")
+        next_page_string = (
+            captured_page_value_list[0] if captured_page_value_list else None
+        )
+        if next_page_string and next_page_string.isdigit():
+            print(next_page_key)
+            print(int(next_page_string))
+            return int(next_page_string)
+
+        return (previous_token or 1) + 1
+
     def post_process(self, row: dict, context: Optional[Dict] = None) -> dict:
         """
         Add a user_id top-level field to be used as state replication key.

From 67fbe9476120bbbe33c2a95fe6e44f8968da5b87 Mon Sep 17 00:00:00 2001
From: Eric Boucher <eric.p.boucher@gmail.com>
Date: Tue, 17 May 2022 23:01:56 -0400
Subject: [PATCH 2/2] remove :footprints:

---
 tap_github/repository_streams.py | 2 --
 1 file changed, 2 deletions(-)

diff --git a/tap_github/repository_streams.py b/tap_github/repository_streams.py
index 6871b577..349b6298 100644
--- a/tap_github/repository_streams.py
+++ b/tap_github/repository_streams.py
@@ -1546,8 +1546,6 @@ def get_next_page_token(
             captured_page_value_list[0] if captured_page_value_list else None
         )
         if next_page_string and next_page_string.isdigit():
-            print(next_page_key)
-            print(int(next_page_string))
             return int(next_page_string)
 
         return (previous_token or 1) + 1