/team: Fetch contributors data from all major repos.

Also wait 2 seconds before trying again.

tools/update-authors-json

@@ -9,10 +9,12 @@ from lib import sanity_check
 sanity_check.check_venv(__file__)
 
 from typing import Any, Dict, List, Optional, Union, Text
+from mypy_extensions import TypedDict
 
 import os
 import sys
 import argparse
+from time import sleep
 from datetime import date
 import subprocess
 
@@ -25,7 +27,6 @@ from django.conf import settings
 from zerver.lib.utils import split_by
 
 FIXTURE_FILE = os.path.join(os.path.dirname(__file__), '../zerver/fixtures/authors.json')
-GITHUB_LINK = 'https://api.github.com/repos/zulip/zulip/stats/contributors'
 
 parser = argparse.ArgumentParser()
 parser.add_argument('--max-retries', type=int, default=3,
@@ -38,22 +39,18 @@ parser.add_argument('--not-required', action='store_true', default=False,
                     help='Consider failures to reach GitHub nonfatal')
 args = parser.parse_args()
 
-def fetch_data(retries, link):
-    # type: (int, str) -> Optional[List[Dict[str, Any]]]
-    for _ in range(retries):
-        try:
-            r = requests.get(link)  # type: requests.Response
-            if r.status_code == 200:
-                return r.json()
-            else:
-                print('Github API return non 200 status.')
-        except requests.exceptions.RequestException as e:
-            print(e)
+ContributorsJSON = TypedDict('ContributorsJSON', {
+    'date': str,
+    'contrib': Dict[str, Dict[str, Union[str, int]]],
+})
 
-    return None
 
-def write_to_disk(json_data, out_file):
-    # type: (Dict[str, Any], str) -> None
+def fetch_contributors(repo_link: str) -> Optional[List[Dict[str, Dict[str, Any]]]]:
+    r = requests.get(repo_link)  # type: requests.Response
+    return r.json() if r.status_code == 200 else None
+
+
+def write_to_disk(json_data: ContributorsJSON, out_file: str) -> None:
     with open(out_file, 'w') as f:
         try:
             f.write("{}\n".format(json.dumps(json_data)))
@@ -61,53 +58,66 @@ def write_to_disk(json_data, out_file):
             print(e)
             sys.exit(1)
 
-def run_production():
-    # type: () -> None
+
+def run_production() -> None:
     """
-    Fetch data from Github and stored it in
-    `static/generated/github-contributors.json`
+    Get contributors data from Github and insert them into a temporaril
+    dictionary. Retry fetching each repository if responded with non HTTP
+    200 status.
     """
-    json_data = fetch_data(args.max_retries, GITHUB_LINK)  # type: Optional[List[Dict[str, Any]]]
-    if json_data:
-        # Successfully fetch data from Github
-        contribs = []
-        for user in json_data:
-            author = user.get('author')
-            if author is None:
-                print("Unable to access fields for %s" % (user,))
-                continue
-            result_user = dict(
-                avatar=author.get('avatar_url'),
-                name=author.get('login'),
-                commits=user.get('total')
-            )
-            contribs.append(result_user)
+    repositories = {
+        'server': 'https://api.github.com/repos/zulip/zulip/stats/contributors',
+        'desktop': 'https://api.github.com/repos/zulip/zulip-electron/stats/contributors',
+        'mobile': 'https://api.github.com/repos/zulip/zulip-mobile/stats/contributors',
+        'python-zulip-api': 'https://api.github.com/repos/zulip/python-zulip-api/stats/contributors',
+        'zulipbot': 'https://api.github.com/repos/zulip/zulipbot/stats/contributors',
+    }
 
-        out_contrib_data = sorted(
-            contribs,
-            key=lambda k: k.get('commits'),
-            reverse=True
-        )  # type: List[Dict[str, Union[Text, int]]]
+    data = dict(date=str(date.today()), contrib={})  # type: ContributorsJSON
+    contribs_data = {}  # type: Dict[str, Dict[str, Union[str, int]]]
 
-        out_data = dict(
-            data=out_contrib_data,
-            date=str(date.today())
-        )  # type: Dict[str, Any]
+    for t in range(args.max_retries):
+        repo_done = []
+        for k, v in repositories.items():
+            contribs = fetch_contributors(v)
+            if contribs:
+                repo_done.append(k)
+                for contrib in contribs:
+                    username = contrib.get('author').get('login')
+                    contrib_data = {
+                        'avatar': contrib.get('author').get('avatar_url'),
+                        k: contrib.get('total'),
+                    }
+                    if username in contribs_data:
+                        contribs_data[username].update(contrib_data)
+                    else:
+                        contribs_data[username] = contrib_data
+        for k in repo_done:
+            del repositories[k]
 
-        write_to_disk(out_data, settings.CONTRIBUTORS_DATA)
+        if not repositories:
+            break
 
-    elif not args.not_required:
-        print('Fail to fetch data from Github.')
-        sys.exit(1)
+        # Wait before retrying failed requests for Github to aggregate data.
+        sleep(2)
+    else:
+        print("ERROR: Failed fetching contributors data from Github.")
+        if not args.not_required:
+            sys.exit(1)
 
-def copy_fixture():
-    # type: () -> None
+    data['contrib'] = contribs_data
+
+    write_to_disk(data, settings.CONTRIBUTORS_DATA)
+
+
+def copy_fixture() -> None:
     """
     Copy test fixture file from zerver/fixtures. This is used to avoid
     constantly fetching data from Github during testing.
     """
     subprocess.check_call(['cp', FIXTURE_FILE, settings.CONTRIBUTORS_DATA])
 
+
 if args.use_fixture:
     copy_fixture()
 else: