pull request review cache

docs/src/content/docs/reference/cli/commands.md

@@ -25,6 +25,7 @@ Options:
   -prc, --pull-request-comment [string]      create comment on a pull request.
   -prd, --pull-request-description [string]  upsert comment on a pull request description.
   -prr, --pull-request-reviews               create pull request reviews from annotations
+  --no-pull-request-reviews-cache            disable pull request reviews cache
   -j, --json                                 emit full JSON response to output
   -y, --yaml                                 emit full YAML response to output
   -p, --prompt                               dry run, don't execute LLM and return expanded prompt

packages/cli/src/cli.ts

@@ -112,6 +112,10 @@ export async function cli() {
             "-prr, --pull-request-reviews",
             "create pull request reviews from annotations"
         )
+        .option(
+            "--no-pull-request-reviews-cache",
+            "disable pull request reviews cache"
+        )
         .option("-j, --json", "emit full JSON response to output")
         .option("-y, --yaml", "emit full YAML response to output")
         .option(

packages/cli/src/run.ts

@@ -35,6 +35,10 @@ import {
     githubUpsetPullRequest,
     githubCreatePullRequestReviews,
     githubCreateIssueComment,
+    GITHUB_PULL_REQUEST_REVIEWS_CACHE,
+    JSONLineCache,
+    PullRequestReviewsCacheKey,
+    PullRequestReviewsCacheValue,
 } from "genaiscript-core"
 import { capitalize } from "inflection"
 import { basename, resolve, join, relative } from "node:path"
@@ -62,6 +66,7 @@ export async function runScript(
         pullRequestComment: string | boolean
         pullRequestDescription: string | boolean
         pullRequestReviews: boolean
+        pullRequestReviewsCache: boolean
         outData: string
         label: string
         temperature: string
@@ -101,6 +106,7 @@ export async function runScript(
     const maxTokens = normalizeInt(options.maxTokens)
     const maxToolCalls = normalizeInt(options.maxToolCalls)
     const cache = !!options.cache
+    const pullRequestReviewsCache = !!options.pullRequestReviewsCache
     const applyEdits = !!options.applyEdits
     const csvSeparator = options.csvSeparator || "\t"
     const removeOut = options.removeOut
@@ -368,8 +374,20 @@ ${Array.from(files)
 
     if (pullRequestReviews && res.annotations?.length) {
         const info = parseGHTokenFromEnv(process.env)
-        if (info.repository && info.issue)
-            await githubCreatePullRequestReviews(script, info, res.annotations)
+        if (info.repository && info.issue) {
+            const cache = pullRequestReviewsCache
+                ? JSONLineCache.byName<
+                      PullRequestReviewsCacheKey,
+                      PullRequestReviewsCacheValue
+                  >(GITHUB_PULL_REQUEST_REVIEWS_CACHE)
+                : undefined
+            await githubCreatePullRequestReviews(
+                script,
+                info,
+                res.annotations,
+                { cache }
+            )
+        }
     }
 
     if (pullRequestComment && res.text) {

packages/core/src/cache.ts

@@ -7,17 +7,17 @@ import { CORE_VERSION } from "./version"
 
 export type CacheEntry<K, V> = { sha: string; key: K; val: V }
 
-export class Cache<K, V> extends EventTarget {
+export class JSONLineCache<K, V> extends EventTarget {
     private _entries: Record<string, CacheEntry<K, V>>
     private constructor(public readonly name: string) {
         super()
     }
 
-    static byName<K, V>(name: string): Cache<K, V> {
+    static byName<K, V>(name: string): JSONLineCache<K, V> {
         name = name.replace(/[^a-z0-9_]/gi, "_")
         const key = "cacheKV." + name
         if (host.userState[key]) return host.userState[key]
-        const r = new Cache<K, V>(name)
+        const r = new JSONLineCache<K, V>(name)
         host.userState[key] = r
         return r
     }

packages/core/src/chat.ts

@@ -1,5 +1,5 @@
 import OpenAI from "openai"
-import { Cache } from "./cache"
+import { JSONLineCache } from "./cache"
 import { MarkdownTrace } from "./trace"
 import { PromptImage } from "./promptdom"
 import { AICIRequest } from "./aici"
@@ -16,6 +16,7 @@ import {
 import { validateFencesWithSchema, validateJSONWithSchema } from "./schema"
 import dedent from "ts-dedent"
 import {
+    CHAT_CACHE,
     DEFAULT_MODEL,
     DEFAULT_TEMPERATURE,
     MAX_DATA_REPAIRS,
@@ -87,18 +88,18 @@ export type ChatCompletationRequestCacheValue = {
     finishReason: ChatCompletionResponse["finishReason"]
 }
 
-export type ChatCompletationRequestCache = Cache<
+export type ChatCompletationRequestCache = JSONLineCache<
     ChatCompletionRequestCacheKey,
     ChatCompletationRequestCacheValue
 >
 
 export function getChatCompletionCache(
     name?: string
 ): ChatCompletationRequestCache {
-    return Cache.byName<
+    return JSONLineCache.byName<
         ChatCompletionRequestCacheKey,
         ChatCompletationRequestCacheValue
-    >(name || "chatv2")
+    >(name || CHAT_CACHE)
 }
 
 export interface ChatCompletionsProgressReport {

packages/core/src/constants.ts

@@ -170,4 +170,8 @@ export const DOCKER_CONTAINER_VOLUME = "/app"
 export const CLI_RUN_FILES_FOLDER = "files"
 
 export const GITHUB_API_VERSION = "2022-11-28"
-export const GITHUB_TOKEN = "GITHUB_TOKEN"
+export const GITHUB_TOKEN = "GITHUB_TOKEN"
+
+export const AI_REQUESTS_CACHE = "airequests"
+export const CHAT_CACHE = "chatv2"
+export const GITHUB_PULL_REQUEST_REVIEWS_CACHE = "prr"

packages/core/src/github.ts

@@ -1,9 +1,15 @@
 import { assert } from "node:console"
-import { GITHUB_API_VERSION, GITHUB_TOKEN } from "./constants"
+import {
+    GITHUB_API_VERSION,
+    GITHUB_PULL_REQUEST_REVIEWS_CACHE,
+    GITHUB_TOKEN,
+} from "./constants"
 import { createFetch } from "./fetch"
 import { host } from "./host"
 import { link, prettifyMarkdown } from "./markdown"
 import { logError, logVerbose, normalizeInt } from "./util"
+import { string } from "mathjs"
+import { JSONLineCache } from "./cache"
 
 export interface GithubConnectionInfo {
     token: string
@@ -211,6 +217,7 @@ async function githubCreatePullRequestReview(
 ) {
     assert(token)
     const { apiUrl, repository, issue, commitSha } = info
+
     const fetch = await createFetch({ retryOn: [] })
     const url = `${apiUrl}/repos/${repository}/pulls/${issue}/comments`
     const body = {
@@ -244,13 +251,36 @@ async function githubCreatePullRequestReview(
     return r
 }
 
+export interface PullRequestReviewsCacheKey {
+    repository: string
+    issue: number
+    scriptId: string
+    annotation: Diagnostic
+}
+
+export interface PullRequestReviewsCacheValue {
+    created: boolean
+    statusText: string
+    html_url: string
+}
+
+export type PullRequestReviewsCache = JSONLineCache<
+    PullRequestReviewsCacheKey,
+    PullRequestReviewsCacheValue
+>
+
 export async function githubCreatePullRequestReviews(
     script: PromptScript,
     info: GithubConnectionInfo,
-    annotations: Diagnostic[]
+    annotations: Diagnostic[],
+    options?: {
+        cache?: PullRequestReviewsCache
+    }
 ): Promise<boolean> {
+    const { repository, issue, sha } = info
+    const { cache } = options || {}
+
     if (!annotations?.length) return true
-    const { issue, sha } = info
     if (!issue) {
         logError("missing pull request number")
         return false
@@ -266,7 +296,25 @@ export async function githubCreatePullRequestReviews(
     }
 
     // code annotations
-    for (const annotation of annotations)
-        await githubCreatePullRequestReview(script, info, token, annotation)
+    for (const annotation of annotations) {
+        const cacheKey = {
+            repository,
+            issue,
+            scriptId: script.id,
+            annotation,
+        }
+        const cached = await cache?.get(cacheKey)
+        if (cached)
+            logVerbose("ignore cached pull request review, " + cached.html_url)
+        else {
+            const res = await githubCreatePullRequestReview(
+                script,
+                info,
+                token,
+                annotation
+            )
+            if (res.created) await cache?.set(cacheKey, res)
+        }
+    }
     return true
 }

packages/sample/genaisrc/summarize-link.genai.js

@@ -7,7 +7,7 @@ script({
     temperature: 0,
     tests: {
         files: [
-            "https://raw.githubusercontent.com/microsoft/genaiscript/main/packages/sample/src/rag/markdown.md",
+            `https://raw.githubusercontent.com/microsoft/genaiscript/main/packages/sample/src/rag/markdown.md`,
         ],
         keywords: "markdown",
     },

packages/vscode/src/airequesttree.ts

@@ -5,7 +5,7 @@ import {
     ExtensionState,
 } from "./state"
 import { CACHE_AIREQUEST_PREFIX, CHANGE, CacheEntry } from "genaiscript-core"
-import { Cache } from "genaiscript-core"
+import { JSONLineCache } from "genaiscript-core"
 import { infoUri } from "./markdowndocumentprovider"
 import { toMarkdownString } from "./markdown"
 
@@ -14,7 +14,7 @@ type AIRequestTreeNode = CacheEntry<AIRequestSnapshotKey, AIRequestSnapshot>
 class AIRequestTreeDataProvider
     implements vscode.TreeDataProvider<AIRequestTreeNode>
 {
-    cache: Cache<AIRequestSnapshotKey, AIRequestSnapshot>
+    cache: JSONLineCache<AIRequestSnapshotKey, AIRequestSnapshot>
     constructor(readonly state: ExtensionState) {
         this.cache = state.aiRequestCache()
         this.cache.addEventListener(CHANGE, () => this.refresh())

packages/vscode/src/state.ts

@@ -13,7 +13,7 @@ import {
     isCancelError,
     delay,
     CHANGE,
-    Cache,
+    JSONLineCache,
     logInfo,
     logMeasure,
     parseAnnotations,
@@ -31,6 +31,7 @@ import {
     fixPromptDefinitions,
     DEFAULT_MODEL,
     resolveModelConnectionInfo,
+    AI_REQUESTS_CACHE,
 } from "genaiscript-core"
 import { ExtensionContext } from "vscode"
 import { VSCodeHost } from "./vshost"
@@ -124,7 +125,7 @@ export function snapshotAIRequest(r: AIRequest): AIRequestSnapshot {
 }
 
 function getAIRequestCache() {
-    return Cache.byName<AIRequestSnapshotKey, AIRequestSnapshot>("airequests")
+    return JSONLineCache.byName<AIRequestSnapshotKey, AIRequestSnapshot>(AI_REQUESTS_CACHE)
 }
 
 export class ExtensionState extends EventTarget {
@@ -133,7 +134,7 @@ export class ExtensionState extends EventTarget {
     private _project: Project = undefined
     private _aiRequest: AIRequest = undefined
     private _diagColl: vscode.DiagnosticCollection
-    private _aiRequestCache: Cache<AIRequestSnapshotKey, AIRequestSnapshot> =
+    private _aiRequestCache: JSONLineCache<AIRequestSnapshotKey, AIRequestSnapshot> =
         undefined
     readonly output: vscode.LogOutputChannel