CodeIssuesDiscussionsWikiPull RequestsProjectsCommitsActionsReleasesContributorsPulse● GatesSecuritySettingsDeploymentsPipelineInsightsAgents✨ Explain✨ Ask AI✨ Workspace✨ Spec✨ Tests▓ Debt Map✨ NL Search🏛 Archaeology
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 | /**
* PR Split Suggestions — AI-powered guidance for decomposing large PRs.
*
* When a PR has >400 lines changed, Claude Sonnet is asked to suggest how
* to split it into 2-4 smaller, independently-mergeable PRs grouped by
* logical concern (schema / API / UI / tests etc.).
*
* Results are cached in memory for 1 hour per PR — the AI call is expensive
* and the diff doesn't change between page loads.
*
* Returns `null` when:
* - PR has <=400 changed lines
* - AI is unavailable (no ANTHROPIC_API_KEY)
* - Claude returns fewer than 2 suggestions
* - Any error occurs (always degrades gracefully)
*/
import { getAnthropic, MODEL_SONNET, extractText, parseJsonResponse, isAiAvailable } from "./ai-client";
import { join } from "path";
import { config } from "./config";
// ---------------------------------------------------------------------------
// Public types
// ---------------------------------------------------------------------------
export interface SplitPr {
title: string;
rationale: string;
files: string[];
estimatedLines: number;
suggestedBranch: string;
}
export interface SplitSuggestion {
originalPrTitle: string;
totalFiles: number;
totalLines: number;
suggestedPrs: SplitPr[];
mergeOrder: string[];
}
// ---------------------------------------------------------------------------
// In-memory cache (1h TTL per prId)
// ---------------------------------------------------------------------------
interface CacheEntry {
suggestion: SplitSuggestion | null;
cachedAt: number;
}
const _cache = new Map<string, CacheEntry>();
const CACHE_TTL_MS = 60 * 60 * 1000; // 1 hour
function getCached(prId: string): SplitSuggestion | null | undefined {
const entry = _cache.get(prId);
if (!entry) return undefined; // cache miss
if (Date.now() - entry.cachedAt > CACHE_TTL_MS) {
_cache.delete(prId);
return undefined;
}
return entry.suggestion;
}
function setCached(prId: string, suggestion: SplitSuggestion | null): void {
_cache.set(prId, { suggestion, cachedAt: Date.now() });
}
// ---------------------------------------------------------------------------
// Diff stat parsing
// ---------------------------------------------------------------------------
interface FileStatLine {
path: string;
added: number;
deleted: number;
total: number;
}
function parseNumstat(raw: string): FileStatLine[] {
return raw
.trim()
.split("\n")
.filter(Boolean)
.map((line) => {
const parts = line.split("\t");
if (parts.length < 3) return null;
const added = parts[0] === "-" ? 0 : parseInt(parts[0], 10) || 0;
const deleted = parts[1] === "-" ? 0 : parseInt(parts[1], 10) || 0;
return { path: parts[2], added, deleted, total: added + deleted };
})
.filter((x): x is FileStatLine => x !== null);
}
function getRepoDir(owner: string, repo: string): string {
return join(config.gitReposPath, `${owner}/${repo}.git`);
}
async function getDiffStat(
owner: string,
repo: string,
baseBranch: string,
headBranch: string
): Promise<FileStatLine[]> {
const repoDir = getRepoDir(owner, repo);
const proc = Bun.spawn(
["git", "--git-dir", repoDir, "diff", "--numstat", `${baseBranch}...${headBranch}`],
{ stdout: "pipe", stderr: "pipe" }
);
const raw = await new Response(proc.stdout).text();
await proc.exited;
return parseNumstat(raw);
}
// ---------------------------------------------------------------------------
// Public API
// ---------------------------------------------------------------------------
/**
* Suggest how to split a large PR.
* Returns null when the PR is small, AI is unavailable, or any error occurs.
*/
export async function suggestPrSplit(
prId: string,
prTitle: string,
ownerName: string,
repoName: string,
baseBranch: string,
headBranch: string
): Promise<SplitSuggestion | null> {
// Check cache first
const cached = getCached(prId);
if (cached !== undefined) return cached;
try {
const fileStats = await getDiffStat(ownerName, repoName, baseBranch, headBranch);
const totalLines = fileStats.reduce((s, f) => s + f.total, 0);
const totalFiles = fileStats.length;
if (totalLines < 400) {
setCached(prId, null);
return null;
}
if (!isAiAvailable()) {
setCached(prId, null);
return null;
}
const fileList = fileStats
.map((f) => `${f.path} +${f.added} -${f.deleted}`)
.join("\n");
const prompt = `This PR is too large to review effectively (${totalLines} lines across ${totalFiles} files).
Suggest how to split it into 2-4 smaller PRs that can be reviewed and merged independently.
PR title: ${prTitle}
Files changed:
${fileList}
Return JSON with this exact shape (no extra keys, no prose outside the JSON block):
{
"suggestedPrs": [
{
"title": "...",
"rationale": "...",
"files": ["..."],
"estimatedLines": N,
"suggestedBranch": "..."
}
],
"mergeOrder": ["PR title 1", "PR title 2"]
}
Rules:
- Group by logical concern (schema changes together, API layer together, UI together)
- Each suggested PR should be independently mergeable
- Suggest merge order to minimise conflicts
- suggestedBranch should be kebab-case derived from the PR title (e.g. feat/auth-schema-only)
- Return between 2 and 4 suggested PRs`;
const anthropic = getAnthropic();
const message = await anthropic.messages.create({
model: MODEL_SONNET,
max_tokens: 1024,
messages: [{ role: "user", content: prompt }],
});
const text = extractText(message);
const parsed = parseJsonResponse<{
suggestedPrs: SplitPr[];
mergeOrder: string[];
}>(text);
if (!parsed || !Array.isArray(parsed.suggestedPrs) || parsed.suggestedPrs.length < 2) {
setCached(prId, null);
return null;
}
const suggestion: SplitSuggestion = {
originalPrTitle: prTitle,
totalFiles,
totalLines,
suggestedPrs: parsed.suggestedPrs,
mergeOrder: Array.isArray(parsed.mergeOrder) ? parsed.mergeOrder : [],
};
setCached(prId, suggestion);
return suggestion;
} catch {
// Always degrade gracefully
setCached(prId, null);
return null;
}
}
|