fix: more durable @ references for commands (#2386)

This commit is contained in:
Aiden Cline
2025-09-02 21:24:56 -05:00
committed by GitHub
parent 751b81af34
commit f740663ded
3 changed files with 69 additions and 93 deletions

View File

@@ -1,48 +0,0 @@
import os from "os"
import path from "path"
/**
* Regular expression to match @ file references in text
* Matches @ followed by file paths, excluding commas, periods at end of sentences, and backticks
* Does not match when preceded by word characters or backticks (to avoid email addresses and quoted references)
*/
export const fileRegex = /(?<![\w`])@(\.?[^\s`,.]*(?:\.[^\s`,.]+)*)/g
/**
* File part type for chat input
*/
export type FilePart = {
type: "file"
url: string
filename: string
mime: string
}
/**
* Processes file references in a template string and returns file parts
* @param template - The template string containing @file references
* @param basePath - The base path to resolve relative file paths against
* @returns Array of file parts for the chat input
*/
export function processFileReferences(template: string, basePath: string): FilePart[] {
// intentionally doing match regex doing bash regex replacements
// this is because bash commands can output "@" references
const matches = template.matchAll(fileRegex)
const parts: FilePart[] = []
for (const match of matches) {
const filename = match[1]
const filepath = filename.startsWith("~/")
? path.join(os.homedir(), filename.slice(2))
: path.resolve(basePath, filename)
parts.push({
type: "file",
url: `file://${filepath}`,
filename,
mime: "text/plain",
})
}
return parts
}

View File

@@ -1,4 +1,6 @@
import os from "os"
import path from "path"
import fs from "fs/promises"
import { spawn } from "child_process"
import { Decimal } from "decimal.js"
import { z, ZodSchema } from "zod"
@@ -50,7 +52,6 @@ import { ulid } from "ulid"
import { defer } from "../util/defer"
import { Command } from "../command"
import { $ } from "bun"
import { processFileReferences } from "./file-reference"
export namespace Session {
const log = Log.create({ service: "session" })
@@ -1229,6 +1230,12 @@ export namespace Session {
})
export type CommandInput = z.infer<typeof CommandInput>
const bashRegex = /!`([^`]+)`/g
/**
* Regular expression to match @ file references in text
* Matches @ followed by file paths, excluding commas, periods at end of sentences, and backticks
* Does not match when preceded by word characters or backticks (to avoid email addresses and quoted references)
*/
export const fileRegex = /(?<![\w`])@(\.?[^\s`,.]*(?:\.[^\s`,.]+)*)/g
export async function command(input: CommandInput) {
log.info("command", input)
@@ -1259,8 +1266,36 @@ export namespace Session {
},
] as ChatInput["parts"]
const fileReferenceParts = processFileReferences(template, Instance.worktree)
parts.push(...fileReferenceParts)
const matches = Array.from(template.matchAll(fileRegex))
await Promise.all(
matches.map(async (match) => {
const name = match[1]
const filepath = name.startsWith("~/")
? path.join(os.homedir(), name.slice(2))
: path.resolve(Instance.worktree, name)
const stats = await fs.stat(filepath).catch(() => undefined)
if (!stats) {
const agent = await Agent.get(name)
if (agent) {
parts.push({
type: "agent",
name: agent.name,
})
}
return
}
if (stats.isDirectory()) return
parts.push({
type: "file",
url: `file://${filepath}`,
filename: name,
mime: "text/plain",
})
}),
)
return prompt({
sessionID: input.sessionID,

View File

@@ -1,12 +1,8 @@
import { describe, expect, test, beforeAll, mock } from "bun:test"
import { describe, expect, test } from "bun:test"
import { Session } from "../../src/session/index"
describe("processFileReferences", () => {
let result: any
beforeAll(async () => {
mock.module("os", () => ({ default: { homedir: () => "/home/fake-user" } }))
const { processFileReferences } = await import("../../src/session/file-reference")
const template = `This is a @valid/path/to/a/file and it should also match at
describe("fileRegex", () => {
const template = `This is a @valid/path/to/a/file and it should also match at
the beginning of a line:
@another-valid/path/to/a/file
@@ -26,77 +22,70 @@ Also shouldn't forget @/absolute/paths.txt with and @/without/extensions,
as well as @~/home-files and @~/paths/under/home.txt.
If the reference is \`@quoted/in/backticks\` then it shouldn't match at all.`
result = processFileReferences(template, "/base")
})
const matches = Array.from(template.matchAll(Session.fileRegex))
test("should extract exactly 12 file references", () => {
expect(result.length).toBe(12)
})
test("all files should have correct type and mime", () => {
result.forEach((file: any) => {
expect(file.type).toBe("file")
expect(file.mime).toBe("text/plain")
})
expect(matches.length).toBe(12)
})
test("should extract valid/path/to/a/file", () => {
expect(result[0].filename).toBe("valid/path/to/a/file")
expect(result[0].url).toBe("file:///base/valid/path/to/a/file")
expect(matches[0][1]).toBe("valid/path/to/a/file")
})
test("should extract another-valid/path/to/a/file", () => {
expect(result[1].filename).toBe("another-valid/path/to/a/file")
expect(result[1].url).toBe("file:///base/another-valid/path/to/a/file")
expect(matches[1][1]).toBe("another-valid/path/to/a/file")
})
test("should extract paths ignoring comma after", () => {
expect(result[2].filename).toBe("commas")
expect(result[2].url).toBe("file:///base/commas")
expect(matches[2][1]).toBe("commas")
})
test("should extract a path with a file extension and comma after", () => {
expect(result[3].filename).toBe("file-extensions.md")
expect(result[3].url).toBe("file:///base/file-extensions.md")
expect(matches[3][1]).toBe("file-extensions.md")
})
test("should extract a path with multiple dots and comma after", () => {
expect(result[4].filename).toBe("multiple.extensions.bak")
expect(result[4].url).toBe("file:///base/multiple.extensions.bak")
expect(matches[4][1]).toBe("multiple.extensions.bak")
})
test("should extract hidden directory", () => {
expect(result[5].filename).toBe(".config/")
expect(result[5].url).toBe("file:///base/.config")
expect(matches[5][1]).toBe(".config/")
})
test("should extract hidden file", () => {
expect(result[6].filename).toBe(".bashrc")
expect(result[6].url).toBe("file:///base/.bashrc")
expect(matches[6][1]).toBe(".bashrc")
})
test("should extract a file ignoring period at end of sentence", () => {
expect(result[7].filename).toBe("foo.md")
expect(result[7].url).toBe("file:///base/foo.md")
expect(matches[7][1]).toBe("foo.md")
})
test("should extract an absolute path with an extension", () => {
expect(result[8].filename).toBe("/absolute/paths.txt")
expect(result[8].url).toBe("file:///absolute/paths.txt")
expect(matches[8][1]).toBe("/absolute/paths.txt")
})
test("should extract an absolute path without an extension", () => {
expect(result[9].filename).toBe("/without/extensions")
expect(result[9].url).toBe("file:///without/extensions")
expect(matches[9][1]).toBe("/without/extensions")
})
test("should extract an absolute path in home directory", () => {
expect(result[10].filename).toBe("~/home-files")
expect(result[10].url).toBe("file:///home/fake-user/home-files")
expect(matches[10][1]).toBe("~/home-files")
})
test("should extract an absolute path under home directory", () => {
expect(result[11].filename).toBe("~/paths/under/home.txt")
expect(result[11].url).toBe("file:///home/fake-user/paths/under/home.txt")
expect(matches[11][1]).toBe("~/paths/under/home.txt")
})
test("should not match when preceded by backtick", () => {
const backtickTest = "This `@should/not/match` should be ignored"
const backtickMatches = Array.from(backtickTest.matchAll(Session.fileRegex))
expect(backtickMatches.length).toBe(0)
})
test("should not match email addresses", () => {
const emailTest = "Contact user@example.com for help"
const emailMatches = Array.from(emailTest.matchAll(Session.fileRegex))
expect(emailMatches.length).toBe(0)
})
})