Use GitHub API token (#37)

conbench/entities/summary.py

@@ -1,4 +1,5 @@
 import decimal
+import os
 
 import flask as f
 import marshmallow
@@ -91,7 +92,13 @@ def create(data):
         if not commit:
             name = repository.split("github.com/")[1]
             url = f"https://api.github.com/repos/{name}/commits/{sha}"
-            response = requests.get(url)
+
+            token, session = os.getenv("GITHUB_API_TOKEN"), None
+            if token:
+                session = requests.Session()
+                session.headers = {"Authorization": f"Bearer {token}"}
+
+            response = session.get(url) if session else requests.get(url)
             github = parse_commit(response.json())
             commit = Commit.create(
                 {

migrations/versions/662175f2e6c6_backfill_parent.py

@@ -5,6 +5,8 @@
 Create Date: 2021-04-21 09:23:28.589522
 
 """
+import os
+
 from alembic import op
 import requests
 
@@ -23,12 +25,18 @@ def upgrade():
     commit_table = Commit.__table__
     connection = op.get_bind()
 
+    token, session = os.getenv("GITHUB_API_TOKEN"), None
+    if token:
+        session = requests.Session()
+        session.headers = {"Authorization": f"Bearer {token}"}
+
     commits = connection.execute(
         commit_table.select().where(commit_table.c.parent == None)  # noqa
     )
+
     for commit in commits:
         url = f"https://api.github.com/repos/apache/arrow/commits/{commit.sha}"
-        response = requests.get(url)
+        response = session.get(url) if session else requests.get(url)
         parent = response.json()["parents"][0]["sha"]
         connection.execute(
             commit_table.update()