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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 | /**
* Streaming PR review — real-time Claude review via SSE.
*
* Additive to the existing batch ai-review.ts. This module streams tokens
* from Claude as they arrive, allowing the browser to display the review
* in real time rather than waiting 10–30 seconds for a batch result.
*
* Usage:
* for await (const token of streamPrReview(...)) {
* // send token as SSE event
* }
*/
import { eq, and, like } from "drizzle-orm";
import { db } from "../db";
import { pullRequests, prComments } from "../db/schema";
import { getRepoPath } from "../git/repository";
import { getAnthropic, isAiAvailable, MODEL_SONNET } from "./ai-client";
import { getBotUserIdOrFallback } from "./bot-user";
// ---------------------------------------------------------------------------
// Public types
// ---------------------------------------------------------------------------
export interface StreamingReviewToken {
type: "token" | "section_start" | "section_end" | "done" | "error";
content?: string; // for type="token"
section?: string; // for section_start/end: "summary" | "finding" | "verdict"
error?: string; // for type="error"
}
// ---------------------------------------------------------------------------
// In-memory streaming state
// ---------------------------------------------------------------------------
/** PR IDs currently being streamed. Prevents duplicate concurrent streams. */
const _streamingPrs = new Set<string>();
/**
* Returns true when a streaming review is already in progress for the given
* PR id.
*/
export function isReviewStreaming(prId: string): boolean {
return _streamingPrs.has(prId);
}
// ---------------------------------------------------------------------------
// Marker embedded in the saved comment body so we can detect duplicates
// ---------------------------------------------------------------------------
export const STREAM_REVIEW_MARKER = "<!-- gluecron:stream-review:v1 -->";
/** Max bytes of diff sent to Claude. Matches the batch reviewer's cap. */
const DIFF_BYTE_CAP = 100_000;
// ---------------------------------------------------------------------------
// Section detection helpers
// ---------------------------------------------------------------------------
/**
* Detect section transitions by examining accumulated lines.
*
* Sections (in order):
* 1. "summary" — opening paragraph
* 2. "finding" — numbered list items (lines starting with a digit + ".")
* 3. "verdict" — line starting with "Verdict:"
*
* Returns the section name when a transition is detected, or null otherwise.
*/
function detectSectionTransition(
accumulated: string,
prevSection: string | null
): string | null {
// Split into lines so we can inspect the latest complete line.
const lines = accumulated.split("\n");
// Look at lines from the end (most recent first)
for (let i = lines.length - 1; i >= 0; i--) {
const line = lines[i].trim();
if (!line) continue;
// Verdict line — always wins
if (line.toLowerCase().startsWith("verdict:") && prevSection !== "verdict") {
return "verdict";
}
// Finding: numbered list item (e.g. "1. ", "2. ", "12. ")
if (/^\d+\.\s/.test(line) && prevSection === "summary") {
return "finding";
}
break; // Only inspect the last non-empty line
}
return null;
}
// ---------------------------------------------------------------------------
// Core streaming generator
// ---------------------------------------------------------------------------
/**
* Start a streaming review for a PR.
*
* Yields `StreamingReviewToken` objects as Claude produces them.
* On completion, saves the full review text as a PR comment
* (idempotent — skipped when the marker already exists in comments).
*/
export async function* streamPrReview(
prId: string,
ownerName: string,
repoName: string,
baseBranch: string,
headBranch: string
): AsyncGenerator<StreamingReviewToken> {
// Guard: AI must be configured
if (!isAiAvailable()) {
yield { type: "error", error: "AI not available — ANTHROPIC_API_KEY is not set" };
return;
}
// Guard: prevent duplicate concurrent streams
if (_streamingPrs.has(prId)) {
yield { type: "error", error: "A streaming review is already in progress for this PR" };
return;
}
// Idempotency: skip if a stream-review comment already exists
try {
const [existing] = await db
.select({ id: prComments.id })
.from(prComments)
.where(
and(
eq(prComments.pullRequestId, prId),
eq(prComments.isAiReview, true),
like(prComments.body, `%${STREAM_REVIEW_MARKER}%`)
)
)
.limit(1);
if (existing) {
yield { type: "error", error: "Streaming review already completed for this PR" };
return;
}
} catch {
// DB check failed — proceed optimistically
}
// Mark PR as streaming
_streamingPrs.add(prId);
// Accumulate full text for DB persistence
let accumulated = "";
let currentSection: string | null = null;
try {
// Compute diff via git
let diffText = "";
try {
const cwd = getRepoPath(ownerName, repoName);
const proc = Bun.spawn(
["git", "diff", `${baseBranch}...${headBranch}`, "--"],
{ cwd, stdout: "pipe", stderr: "pipe" }
);
const raw = await new Response(proc.stdout).text();
await proc.exited;
diffText = raw;
} catch {
diffText = "";
}
if (!diffText.trim()) {
yield { type: "error", error: "No diff found between branches — nothing to review" };
return;
}
// Cap diff size
if (diffText.length > DIFF_BYTE_CAP) {
diffText = diffText.slice(0, DIFF_BYTE_CAP);
}
// Build messages
const systemPrompt =
"You are a senior code reviewer. Review this PR diff section by section. " +
"Start with a one-paragraph summary of the overall change, then list specific findings " +
"(each as: file:line — issue description), then give a verdict " +
"(Approve / Request Changes / Comment). Be direct and specific. " +
"Format findings as a numbered list. Start the verdict line with 'Verdict:'.";
const userMessage = `PR diff:\n${diffText}`;
// Open the streaming request
const anthropic = getAnthropic();
// Emit the summary section start before streaming begins
yield { type: "section_start", section: "summary" };
currentSection = "summary";
const stream = anthropic.messages.stream({
model: MODEL_SONNET,
max_tokens: 2048,
system: systemPrompt,
messages: [{ role: "user", content: userMessage }],
});
// Stream tokens
for await (const chunk of stream) {
if (
chunk.type === "content_block_delta" &&
chunk.delta.type === "text_delta"
) {
const text = chunk.delta.text;
if (!text) continue;
accumulated += text;
// Check for section transitions
const newSection = detectSectionTransition(accumulated, currentSection);
if (newSection && newSection !== currentSection) {
// Close the old section
yield { type: "section_end", section: currentSection ?? "summary" };
// Open the new section
yield { type: "section_start", section: newSection };
currentSection = newSection;
}
// Yield the token
yield { type: "token", content: text };
}
}
// Close the final section
if (currentSection) {
yield { type: "section_end", section: currentSection };
}
// Persist the full review as a PR comment (idempotent)
if (accumulated.trim()) {
try {
// Re-check idempotency before insert
const [existingCheck] = await db
.select({ id: prComments.id })
.from(prComments)
.where(
and(
eq(prComments.pullRequestId, prId),
eq(prComments.isAiReview, true),
like(prComments.body, `%${STREAM_REVIEW_MARKER}%`)
)
)
.limit(1);
if (!existingCheck) {
// Resolve author for the comment
const [pr] = await db
.select({ authorId: pullRequests.authorId })
.from(pullRequests)
.where(eq(pullRequests.id, prId))
.limit(1);
if (pr) {
const commentAuthorId = await getBotUserIdOrFallback(pr.authorId);
const commentBody = `## AI Stream Review\n\n${accumulated}\n\n${STREAM_REVIEW_MARKER}`;
await db.insert(prComments).values({
pullRequestId: prId,
authorId: commentAuthorId,
isAiReview: true,
body: commentBody,
});
}
}
} catch (err) {
// Non-fatal — review was streamed, just not persisted
if (process.env.DEBUG_AI_REVIEW === "1") {
console.error("[streaming-review] failed to persist review comment:", err);
}
}
}
yield { type: "done" };
} catch (err) {
const message = err instanceof Error ? err.message : String(err);
yield { type: "error", error: message };
} finally {
_streamingPrs.delete(prId);
}
}
|