feat 16099: Rebase strategy for backport PRs

jenkins/opensearch/backport-pr.jenkinsfile

@@ -0,0 +1,65 @@
+pipeline{
+    agent any
+    environment{
+        GITHUB_TOKEN = credentials('github-token')
+        REPO_URL = 'https://github.com/opensearch-project/opensearch-build.git'
+        REPO_DIR = 'opensearch-build'
+    }
+    stages{
+        stage('Determine PR type'){
+           steps{
+               echo 'Determining if the PR is Stalled or Backport..'
+               script{
+                   def prType = determinePRType()
+                   if(prType == 'stalled'){
+                       echo 'Processing stalled PR..'
+                       processStalledPR()
+                   }
+                   else if(prType == 'backport'){
+                       echo 'Processing Backport PR...'
+                       processBackportPR()
+                   }
+                   else{
+                   echo 'PR does not match criteria. Exiting Pipeline.'
+                   }
+               }
+           }
+        }
+    }
+   post{
+       success{
+           echo 'Pipeline completed successfully'
+       }
+       failure{
+           echo 'Pipeline failed. Check logs for details'
+       }
+   }
+}
+
+def determinePRType(token, repoURL, prId){
+    def result = sh(script: """
+        curl -s -H "Authorization: Bearer ${token}" -H "Accept: application/vnd.github.v3+json" \
+        ${repoURL}/pulls/${prId} | jq '.labels | map(.name)'
+    """, returnStdout: true).trim()
+    if(result.contains('stalled')){
+        return 'stalled'
+    }else if (result.contains('backport')){
+        return 'backport'
+    }else{
+        return null
+    }
+}
+
+def processStalledPR(){
+    echo 'Rebasing stalled PR branch onto target branch...'
+    sh """
+    scripts/pr-management/StalledPRs.py --repo $REPO_REPO_URL --token $GITHUB_TOKEN
+    """
+}
+
+def processBackportPR(){
+     echo 'Resolving backport PR conflicts...'
+     sh """
+     scripts/pr-management/BackportPRs.py --repo $REPO_URL --token $GITHUB_TOKEN
+     """
+}

scripts/pr-management/BackportPRs.py

@@ -0,0 +1,65 @@
+import os
+import requests
+import subprocess
+
+GITHUB_TOKEN = os.environ.get("GITHUB_TOKEN")
+HEADERS = {"Authorization": f"Bearer {GITHUB_TOKEN}", "Accept": "application/vnd.github.v3+json"}
+BASE_URL = "https://api.github.com"
+
+def fetch_backport_prs(owner, repo):
+    """Fetch backport PR's with the `backport` label"""
+    url = f"{BASE_URL}/search/issues"
+    query = f"repo:{owner}/{repo} label:backport is:pr is:open"
+    response = requests.get(url, headers=HEADERS, params={"q":query})
+    response.raise_for_status()
+    return response.json()["items"]
+
+def fetch_pr_details(owner, repo, pr_number):
+    """Fetch PR details to get source and target branches"""
+    url = f"{BASE_URL}/repos/{owner}/{repo}/pulls/{pr_number}"
+    response = requests.get(url, header=HEADERS)
+    response.raise_for_status()
+    return response.json()["items"]
+
+def resolve_changelog_conflict(repo_dir, pr_branch, target_branch):
+    """Resolve conflicts in CHANGELOG.md"""
+    subprocess.run(["git","checkout",pr_branch], cwd=repo_dir)
+    subprocess.run(["git","fetch","origin", target_branch], cwd=repo_dir)
+    subprocess.run(["git","rebase",f"origin/{target_branch}"], cwd=repo_dir)
+
+    conflicted_files = subprocess.check_output(
+        ["git","diff","--name-only","--diff-filter=U"], cwd=repo_dir
+    ).decode().strip().split("\n")
+    if "CHANGELOG.md" in conflicted_files:
+        print("Conflict detected in CHANGELOG.md. Resolving....")
+        changelog_file = f"{repo_dir}/CHANGELOG.md"
+        with open(changelog_file, "r") as file:
+            lines = file.readlines()
+        start_old = lines.index("<<<<<<< HEAD\n")
+        middle = lines.index("=======\n")
+        end_new = lines.index(">>>>>>> ", middle)
+        old_changes = lines[start_old + 1: middle]
+        new_changes = lines[middle + 1: end_new]
+        resolved_changes = (
+                ["# CHANGELOG\n\n"]
+                + ["## Existing Changes (from target branch):\n"] + old_changes
+                + ["\n## New Changes (from backport PR):\n"] + new_changes
+        )
+        with open(changelog_file, "w") as file:
+            file.writelines(resolved_changes)
+        subprocess.run(["git", "add", "CHANGELOG.md"], cwd=repo_dir)
+        subprocess.run(["git","commit","-m","Resolved CHANGELOG.md conflict"], cwd=repo_dir)
+    subprocess.run(["git","push","--force-with-lease"], cwd=repo_dir)
+
+def main_backport(owner, repo, repo_dir):
+    """Main function to handle backport PRs"""
+    backport_prs = fetch_backport_prs(owner,repo)
+    for pr in backport_prs:
+        pr_number = pr["number"]
+        pr_details = fetch_pr_details(owner, repo, pr_number)
+        pr_branch = pr_details["head"]["ref"]
+        target_branch = pr_details["base"]["ref"]
+        print(f"Handling Backport PR #{pr_number}: {pr_branch} -> {target_branch}")
+        resolve_changelog_conflict(repo_dir, pr_branch, target_branch)
+
+

scripts/pr-management/README.md

@@ -0,0 +1,16 @@
+# PR Management Scripts
+
+This folder contains scripts for automating tasks related to pull requests.
+
+## BackportPRs.py
+
+This script handles backport PRs by:
+- Detecting and resolving conflicts in `CHANGELOG.md`.
+- Combining old and new changes during conflict resolution.
+- Committing the resolved file back to the PR branch.
+
+## StalledPRs.py
+
+This script handles Stalled PRs by:
+- Fetching all the stalled PRs using the Stalled label
+- Rebase the PRs onto the latest target branch and push updates

scripts/pr-management/StalledPRs.py

@@ -0,0 +1,36 @@
+import os
+import requests
+import subprocess
+
+
+GITHUB_TOKEN = os.environ.get("GITHUB_TOKEN")
+HEADERS = {"Authorization": f"Bearer {GITHUB_TOKEN}", "Accept": "application/vnd.github.v3+json"}
+BASE_URL = "https://api.github.com"
+
+def fetch_stalled_prs(owner, repo):
+    """Fetch stalled PRs with the `stalled` label"""
+    url = f"{BASE_URL}"/search/issues
+    query = f"repo:{owner}/{repo} label:stalled is:pr is:open"
+    response = requests.get(url, headers=HEADERS, params={"q": query})
+    response.raise_for_status()
+    return response.json()["items"]
+
+def rebase_pr(repo_dir, pr_branch, target_branch):
+    """Rebase a stalled PR onto the target branch."""
+    subprocess.run(["git","checkout",pr_branch], cwd=repo_dir)
+    subprocess.run(["git","fetch","origin", target_branch], cwd=repo_dir)
+    subprocess.run(["git","rebase",f"origin"/{target_branch}], cwd=repo_dir)
+    subprocess.run(["git","push","--force-with-lease"], cwd=repo_dir)
+
+def main_stalled(owner, repo, repo_dir):
+    """Main function to handle stalled PRs"""
+    stalled_prs = fetch_stalled_prs(owner,repo)
+    for pr in stalled_prs:
+        pr_number = pr["number"]
+        pr_details = fetch_stalled_prs(owner, repo, pr_number)
+        pr_branch = pr_details["head"]["ref"]
+        target_branch = pr_details["base"]["ref"]
+        print(f"Handling Stalled PR #{pr_number}: {pr_branch} -> {target_branch}")
+        rebase_pr(repo_dir, pr_branch, target_branch)
+
+