Task 2583: GitHub activity pipeline

config/settings.py

@@ -417,6 +417,15 @@
 # GitHub settings
 
 GITHUB_TOKEN = env("GITHUB_TOKEN", default=None)
+
+# Org scope for the profile GitHub activity card. The node ID is pinned so
+# boost_activity() never spends a REST call resolving it. Use the modern global
+# ID form. REST node_id and GraphQL organization.id both still return the
+# deprecated token form.
+BOOST_GITHUB_ORG = "boostorg"
+BOOST_GITHUB_ORG_NODE_ID = env("BOOST_GITHUB_ORG_NODE_ID", default="O_kgDOADBg4Q")
+# contributionsCollection is capped at one year by GitHub.
+BOOST_ACTIVITY_WINDOW_DAYS = 365
 JDOODLE_API_CLIENT_ID = env("JDOODLE_API_CLIENT_ID", "")
 JDOODLE_API_CLIENT_SECRET = env("JDOODLE_API_CLIENT_SECRET", "")
 

config/urls.py

@@ -65,6 +65,7 @@
 from users.views import (
     CurrentUserAPIView,
     CurrentUserProfileView,
+    GithubActivityFragmentView,
     CustomEmailVerificationSentView,
     CustomLoginView,
     CustomSignupView,
@@ -151,6 +152,11 @@
         ),
         path("accounts/", include("allauth.urls")),
         path("users/me/", CurrentUserProfileView.as_view(), name="profile-account"),
+        path(
+            "users/me/github-activity/",
+            GithubActivityFragmentView.as_view(),
+            name="profile-github-activity",
+        ),
         path("users/me/delete/", DeleteUserView.as_view(), name="profile-delete"),
         path(
             "users/me/cancel-delete/",

core/githubhelper.py

@@ -1,7 +1,7 @@
 import base64
 import re
 from collections import defaultdict
-from datetime import datetime
+from datetime import datetime, timezone as dt_timezone, timedelta
 from socket import gaierror
 import time
 from urllib.error import URLError
@@ -24,6 +24,34 @@
 
 logger = structlog.get_logger()
 
+_CONTRIBUTIONS_QUERY = """
+query BoostActivity($login: String!, $orgId: ID!, $from: DateTime!, $to: DateTime!) {
+  user(login: $login) {
+    contributionsCollection(organizationID: $orgId, from: $from, to: $to) {
+      totalCommitContributions
+      totalRepositoriesWithContributedCommits
+      totalPullRequestContributions
+      totalRepositoriesWithContributedPullRequests
+      totalPullRequestReviewContributions
+      totalRepositoriesWithContributedPullRequestReviews
+      repositoryContributions(first: 1) {
+        totalCount
+      }
+      pullRequestContributions(first: 50) {
+        nodes {
+          pullRequest {
+            title
+            url
+            repository { nameWithOwner }
+            comments { totalCount }
+          }
+        }
+      }
+    }
+  }
+}
+"""
+
 
 class GithubAPIClient:
     """A class to interact with the GitHub API."""
@@ -574,6 +602,23 @@ def get_artifact_content(self, url):
         with myzip.open(myzip.filelist[0]) as f:
             return f.read().decode()
 
+    def graphql(self, query: str, variables: dict) -> dict:
+        """Execute a GitHub GraphQL query using the app-level PAT."""
+        response = requests.post(
+            "https://api.github.com/graphql",
+            json={"query": query, "variables": variables},
+            headers={
+                "Authorization": f"Bearer {self.token}",
+                "Content-Type": "application/json",
+            },
+            timeout=30,
+        )
+        response.raise_for_status()
+        data = response.json()
+        if "errors" in data:
+            raise ValueError(f"GitHub GraphQL errors: {data['errors']}")
+        return data["data"]
+
 
 class GithubDataParser:
     def get_commits_per_month(self, commits: list[dict]):
@@ -712,3 +757,55 @@ def extract_name(self, val: str) -> str:
             val = val.replace(email.group(), "")
 
         return val.strip()
+
+
+def boost_activity(login: str) -> dict:
+    """Fetch boostorg contribution totals for a GitHub user (trailing 12 months).
+
+    Scoped to the ``boostorg`` org only. Raises on any API or GraphQL error, and
+    on an unknown login, so the caller can keep the last good snapshot rather
+    than overwrite it with zeros.
+
+    Returns a dict with keys:
+        total_commits, commit_repo_count, repos_created,
+        prs_opened, pr_repo_count, prs_reviewed, review_repo_count,
+        featured_pr (dict or None)
+    """
+    now = datetime.now(dt_timezone.utc)
+    data = GithubAPIClient().graphql(
+        _CONTRIBUTIONS_QUERY,
+        {
+            "login": login,
+            "orgId": settings.BOOST_GITHUB_ORG_NODE_ID,
+            "from": (
+                now - timedelta(days=settings.BOOST_ACTIVITY_WINDOW_DAYS)
+            ).isoformat(),
+            "to": now.isoformat(),
+        },
+    )
+
+    user_node = data.get("user")
+    if not user_node:
+        raise ValueError(f"GitHub user not found: {login}")
+
+    coll = user_node["contributionsCollection"]
+    prs = [
+        {
+            "title": node["pullRequest"]["title"],
+            "url": node["pullRequest"]["url"],
+            "repo": node["pullRequest"]["repository"]["nameWithOwner"],
+            "comment_count": node["pullRequest"]["comments"]["totalCount"],
+        }
+        for node in coll["pullRequestContributions"]["nodes"]
+    ]
+
+    return {
+        "total_commits": coll["totalCommitContributions"],
+        "commit_repo_count": coll["totalRepositoriesWithContributedCommits"],
+        "repos_created": coll["repositoryContributions"]["totalCount"],
+        "prs_opened": coll["totalPullRequestContributions"],
+        "pr_repo_count": coll["totalRepositoriesWithContributedPullRequests"],
+        "prs_reviewed": coll["totalPullRequestReviewContributions"],
+        "review_repo_count": coll["totalRepositoriesWithContributedPullRequestReviews"],
+        "featured_pr": max(prs, key=lambda p: p["comment_count"]) if prs else None,
+    }

core/tests/test_githubhelper.py

@@ -1,11 +1,13 @@
 import datetime
+import json
 from unittest.mock import MagicMock, Mock
 
 import pytest
+import requests
 import responses
 from ghapi.all import GhApi
 
-from core.githubhelper import GithubAPIClient, GithubDataParser
+from core.githubhelper import GithubAPIClient, GithubDataParser, boost_activity
 
 """GithubAPIClient Tests"""
 
@@ -258,3 +260,110 @@ def test_extract_contributor_data():
     assert expected["valid_email"] is False
     assert expected["display_name"] == result["display_name"]
     assert "email" in result
+
+
+"""boost_activity Tests"""
+
+GRAPHQL_URL = "https://api.github.com/graphql"
+
+
+def _contributions_payload(prs=None, **overrides):
+    """Build a contributionsCollection GraphQL response."""
+    collection = {
+        "totalCommitContributions": 5,
+        "totalRepositoriesWithContributedCommits": 3,
+        "totalPullRequestContributions": 5,
+        "totalRepositoriesWithContributedPullRequests": 5,
+        "totalPullRequestReviewContributions": 2,
+        "totalRepositoriesWithContributedPullRequestReviews": 1,
+        "repositoryContributions": {"totalCount": 0},
+        "pullRequestContributions": {"nodes": prs or []},
+    }
+    collection.update(overrides)
+    return {"data": {"user": {"contributionsCollection": collection}}}
+
+
+def _pr_node(repo, comments, title="A PR", url="https://github.com/x/y/pull/1"):
+    return {
+        "pullRequest": {
+            "title": title,
+            "url": url,
+            "repository": {"nameWithOwner": repo},
+            "comments": {"totalCount": comments},
+        }
+    }
+
+
+@responses.activate
+def test_boost_activity_queries_boostorg_only(settings):
+    """Exactly one GraphQL call, scoped to the boostorg node ID."""
+    settings.BOOST_GITHUB_ORG_NODE_ID = "O_kgDOADBg4Q"
+    responses.add(responses.POST, GRAPHQL_URL, json=_contributions_payload())
+
+    result = boost_activity("testuser")
+
+    assert len(responses.calls) == 1
+    body = json.loads(responses.calls[0].request.body)
+    assert body["variables"]["orgId"] == "O_kgDOADBg4Q"
+    assert body["variables"]["login"] == "testuser"
+    assert result["total_commits"] == 5
+    assert result["commit_repo_count"] == 3
+
+
+@responses.activate
+def test_boost_activity_raises_on_graphql_error():
+    """A GraphQL errors payload raises instead of returning partial data."""
+    responses.add(
+        responses.POST,
+        GRAPHQL_URL,
+        json={"errors": [{"message": "Could not resolve to a node"}]},
+    )
+
+    with pytest.raises(ValueError):
+        boost_activity("testuser")
+
+
+@responses.activate
+def test_boost_activity_raises_on_http_error():
+    """A transient 502 must raise so callers keep the last good snapshot."""
+    responses.add(responses.POST, GRAPHQL_URL, status=502, json={})
+
+    with pytest.raises(requests.exceptions.HTTPError):
+        boost_activity("testuser")
+
+
+@responses.activate
+def test_boost_activity_raises_on_unknown_login():
+    """A null user is a success payload, but must not be stored as zeros."""
+    responses.add(responses.POST, GRAPHQL_URL, json={"data": {"user": None}})
+
+    with pytest.raises(ValueError):
+        boost_activity("nosuchuser")
+
+
+@responses.activate
+def test_boost_activity_featured_pr_is_highest_comment_count():
+    """The featured PR is the one with the most conversation comments."""
+    responses.add(
+        responses.POST,
+        GRAPHQL_URL,
+        json=_contributions_payload(
+            prs=[
+                _pr_node("boostorg/url", 5),
+                _pr_node("boostorg/beast", 9),
+                _pr_node("boostorg/json", 2),
+            ]
+        ),
+    )
+
+    featured = boost_activity("testuser")["featured_pr"]
+
+    assert featured["repo"] == "boostorg/beast"
+    assert featured["comment_count"] == 9
+
+
+@responses.activate
+def test_boost_activity_featured_pr_none_when_no_prs():
+    responses.add(responses.POST, GRAPHQL_URL, json=_contributions_payload(prs=[]))
+
+    assert boost_activity("testuser")["featured_pr"] is None

docs/env_vars.md

@@ -9,6 +9,13 @@ This project uses environment variables to configure certain aspects of the appl
 - In **deployed environments**, this should be set to a valid access token associated with the GitHub organization. Edit `kube/boost/values.yaml` (or the environment-specific yaml file) to change this value.
 
 
+## `BOOST_GITHUB_ORG_NODE_ID`
+
+- The GraphQL node ID of the GitHub org whose contributions the profile activity card counts.
+- Optional. Defaults to the `boostorg` org (`O_kgDOADBg4Q`), so neither local nor deployed environments need to set it.
+- Set it only to scope the card at a different org, such as a fork or a test org. To find the value for another org, query GitHub's GraphQL API for that org's `id` and use the `next_global_id` from the deprecation warning — both the REST `node_id` field and GraphQL's `organization.id` still return the older, deprecated ID format.
+
+
 ## `ENVIRONMENT_NAME`
 
 - Used to indicate the name of the environment where the application is running.

env.template

@@ -16,6 +16,10 @@ SECRET_KEY="top-secret"
 
 GITHUB_TOKEN=
 
+# Optional. Which GitHub org the profile activity card counts contributions in.
+# Defaults to boostorg, so only set this to point at a fork or a test org.
+# BOOST_GITHUB_ORG_NODE_ID=O_kgDOADBg4Q
+
 # AWS_ACCESS_KEY_ID="changeme"
 # AWS_SECRET_ACCESS_KEY="changeme"
 # BUCKET_NAME="boost.revsys.dev"

static/css/v3/user-profile-page.css

@@ -291,6 +291,20 @@
     animation: create-post-spin 0.8s linear infinite;
 }
 
+.user-profile__github-status {
+    display: flex;
+    align-items: center;
+    justify-content: flex-end;
+    gap: var(--space-s);
+    min-height: 16px;
+    margin-top: var(--space-default);
+}
+
+.user-profile__github-status-text {
+    color: var(--color-text-tertiary);
+    font-size: var(--font-size-xs);
+}
+
 @keyframes create-post-spin {
     to {
         transform: rotate(360deg);

templates/admin/user_change_form.html

@@ -0,0 +1,30 @@
+{% extends "admin/change_form.html" %}
+{% load i18n admin_urls %}
+
+{% block submit_buttons_bottom %}
+  {{ block.super }}
+  {% if original and original.github_username %}
+    <div class="submit-row">
+      {# POST, not a link: this queues work, so it must not be reachable by GET. #}
+      <form method="post"
+            action="{% url 'admin:user_refresh_github_activity' original.pk %}"
+            style="display: inline-block;">
+        {% csrf_token %}
+        <button type="submit"
+                class="button"
+                style="background: #417690; color: white; padding: 10px 15px; border: none; border-radius: 4px; cursor: pointer;">
+          {% trans "Refresh GitHub Activity" %}
+        </button>
+      </form>
+      <span style="margin-left: 10px; color: #666; font-size: 13px;">
+        {% trans "Queues a background task to fetch this user's Boost org activity from GitHub." %}
+      </span>
+    </div>
+  {% elif original and not original.github_username %}
+    <div class="submit-row">
+      <span style="color: #888; font-size: 13px;">
+        {% trans "No GitHub username set - cannot fetch activity." %}
+      </span>
+    </div>
+  {% endif %}
+{% endblock %}

templates/v3/includes/_github_activity_card.html

@@ -0,0 +1,25 @@
+{% comment %}
+  GitHub activity card plus its sync status line.
+
+  While a background refresh is running this element re-requests itself and
+  swaps itself out. Each response decides whether polling continues: the
+  hx-* attributes are only rendered while still refreshing and under the
+  attempt cap, so the chain stops on its own once the data lands (or once we
+  give up and ask the user to reload).
+
+  Inputs:
+    data: dict, required - from users.profile_cards.github_activity_card
+    attempt: int, required - how many polls have happened so far
+    poll_exhausted: bool, required - whether the attempt cap has been reached
+    poll_url: str, required - url of the fragment endpoint
+    poll_interval: int, required - seconds between polls
+{% endcomment %}
+<div id="github-activity-card"
+     {% if data.refreshing and not poll_exhausted %}
+     hx-get="{{ poll_url }}?attempt={{ next_attempt }}"
+     hx-trigger="load delay:{{ poll_interval }}s"
+     hx-swap="outerHTML"
+     {% endif %}>
+  {% include 'v3/includes/_markdown_card.html' with title=data.title markdown=data.markdown_text button_url=data.button_url button_label=data.button_label %}
+  {% include 'v3/includes/_github_activity_status.html' with refreshing=data.refreshing last_synced=data.last_synced poll_exhausted=poll_exhausted only %}
+</div>