zed/script/github-triage-queue-board.py

#!/usr/bin/env python3
"""
Maintain a GitHub project as a search filter for second-line issue triage.

GitHub issue search can't express "created at least one business day ago", so
this script keeps the project's membership equal to the open issues created in
that window, adding issues that have become eligible and removing the ones that
have aged out.

Requires:
    requests (pip install requests)

Usage (called by the workflow, not directly):
    python github-triage-queue-board.py
"""

import os
import time
import urllib.parse
from datetime import datetime, timedelta
from zoneinfo import ZoneInfo

import requests

RETRYABLE_STATUS_CODES = {502, 503, 504}
MAX_RETRIES = 3
RETRY_DELAY_SECONDS = 5

GITHUB_API_URL = "https://api.github.com"
REPO_OWNER = "zed-industries"
REPO_NAME = "zed"
# The eligibility window's boundaries are calendar days in this timezone. GitHub
# search interprets bare dates as UTC, so the window is pinned to a fixed zone to
# stay aligned with a consistent business day regardless of where the job runs.
# ZoneInfo (not a fixed offset) so daylight saving is handled automatically.
BUSINESS_DAY_TZ = ZoneInfo("America/Los_Angeles")


def github_graphql(query, variables):
    for attempt in range(MAX_RETRIES + 1):
        response = requests.post(
            f"{GITHUB_API_URL}/graphql",
            headers=GITHUB_HEADERS,
            json={"query": query, "variables": variables},
            timeout=30,
        )
        if response.status_code in RETRYABLE_STATUS_CODES and attempt < MAX_RETRIES:
            print(
                f"GitHub API returned {response.status_code}, retrying in {RETRY_DELAY_SECONDS}s (attempt {attempt + 1}/{MAX_RETRIES})..."
            )
            time.sleep(RETRY_DELAY_SECONDS)
            continue
        response.raise_for_status()
        result = response.json()
        if "errors" in result:
            raise RuntimeError(f"GraphQL error: {result['errors']}")
        return result["data"]
    raise RuntimeError("github_graphql: retry loop exited without return")


def github_rest_get(path):
    for attempt in range(MAX_RETRIES + 1):
        response = requests.get(
            f"{GITHUB_API_URL}/{path}", headers=GITHUB_HEADERS, timeout=30
        )
        if response.status_code in RETRYABLE_STATUS_CODES and attempt < MAX_RETRIES:
            print(
                f"GitHub API returned {response.status_code}, retrying in {RETRY_DELAY_SECONDS}s (attempt {attempt + 1}/{MAX_RETRIES})..."
            )
            time.sleep(RETRY_DELAY_SECONDS)
            continue
        response.raise_for_status()
        return response.json()
    raise RuntimeError("github_rest_get: retry loop exited without return")


def business_day_end(day):
    # Latest instant of a calendar day in the business-day timezone.
    return datetime(day.year, day.month, day.day, 23, 59, 59, tzinfo=BUSINESS_DAY_TZ)


def eligibility_window(today):
    # Returns the (start, end) instants bounding the window. `start` is exclusive
    # and `end` inclusive (see eligible_issues); both are day-ends so the whole
    # from-day is excluded and the whole to-day included. Returning full
    # timestamps rather than bare dates keeps GitHub search, which reads bare
    # dates as UTC, from skewing the window by up to a day at the UTC boundary.
    weekday = today.isoweekday()  # Monday = 1
    # Nothing advances over the weekend, so compute as of the preceding Friday.
    if weekday >= 6:
        today -= timedelta(days=weekday - 5)
        weekday = 5
    one_biz_day_ago = today - timedelta(days=3 if weekday == 1 else 1)
    this_monday = today - timedelta(days=weekday - 1)
    eligible_from = this_monday - timedelta(days=4)
    return business_day_end(eligible_from), business_day_end(one_biz_day_ago)


def eligible_issues(window_start, window_end):
    # created:> excludes window_start; created:<= includes window_end.
    query = (
        f"repo:{REPO_OWNER}/{REPO_NAME} is:issue is:open "
        f"created:>{window_start.isoformat()} created:<={window_end.isoformat()}"
    )
    issues = {}
    page = 1
    while True:
        data = github_rest_get(
            f"search/issues?q={urllib.parse.quote(query)}&per_page=100&page={page}"
        )
        for item in data["items"]:
            issues[item["node_id"]] = item["number"]
        if len(data["items"]) < 100 or page >= 10:
            return issues
        page += 1


def fetch_project_id(project_number):
    data = github_graphql(
        """
        query($owner: String!, $number: Int!) {
          organization(login: $owner) { projectV2(number: $number) { id } }
        }
        """,
        {"owner": REPO_OWNER, "number": project_number},
    )
    project = data["organization"]["projectV2"]
    if not project:
        raise RuntimeError(f"project #{project_number} not found")
    return project["id"]


def project_items(project_id):
    # Yields (item_id, content_id, number) for each issue on the project.
    # content_id is the issue's global node id, which is the same value the REST
    # search API returns as node_id, so it can be compared directly against the
    # keys of eligible_issues().
    cursor = None
    while True:
        data = github_graphql(
            """
            query($project: ID!, $cursor: String) {
              node(id: $project) {
                ... on ProjectV2 {
                  items(first: 100, after: $cursor) {
                    pageInfo { hasNextPage endCursor }
                    nodes { id content { ... on Issue { id number } } }
                  }
                }
              }
            }
            """,
            {"project": project_id, "cursor": cursor},
        )
        items = data["node"]["items"]
        for item in items["nodes"]:
            content = item["content"] or {}
            if "id" in content:  # skip drafts / PRs (no Issue content)
                yield item["id"], content["id"], content["number"]
        if not items["pageInfo"]["hasNextPage"]:
            return
        cursor = items["pageInfo"]["endCursor"]


def add_to_project(project_id, content_id):
    github_graphql(
        "mutation($p: ID!, $c: ID!) { addProjectV2ItemById(input: {projectId: $p, contentId: $c}) { item { id } } }",
        {"p": project_id, "c": content_id},
    )


def remove_from_project(project_id, item_id):
    github_graphql(
        "mutation($p: ID!, $i: ID!) { deleteProjectV2Item(input: {projectId: $p, itemId: $i}) { deletedItemId } }",
        {"p": project_id, "i": item_id},
    )


def sync_project(project_id, dry_run=False):
    today = datetime.now(BUSINESS_DAY_TZ).date()
    window_start, window_end = eligibility_window(today)
    print(f"window: created >{window_start.isoformat()} <={window_end.isoformat()}")

    eligible = eligible_issues(window_start, window_end)
    current = list(project_items(project_id))
    on_board = {content_id for _, content_id, _ in current}

    for content_id, number in eligible.items():
        if content_id not in on_board:
            print(f"+ #{number}")
            if not dry_run:
                add_to_project(project_id, content_id)

    for item_id, content_id, number in current:
        if content_id not in eligible:
            print(f"- #{number}")
            if not dry_run:
                remove_from_project(project_id, item_id)


if __name__ == "__main__":
    GITHUB_HEADERS = {
        "Authorization": f"Bearer {os.environ['GITHUB_TOKEN']}",
        "Accept": "application/vnd.github+json",
        "X-GitHub-Api-Version": "2022-11-28",
    }
    project_number = int(os.environ["PROJECT_NUMBER"])
    project_id = fetch_project_id(project_number)
    sync_project(project_id, os.environ.get("DRY_RUN") == "true")