split docs/produtct dependencies

THIRD_PARTY_LICENSES.md

@@ -28192,12 +28192,11 @@ SOFTWARE.
 
 -----------
 
-The following npm packages may be included in this product:
+The following npm package may be included in this product:
 
  - @slidev/[email protected]
- - @slidev/[email protected]
 
-These packages each contain the following license and notice below:
+This package contains the following license and notice below:
 
 MIT License
 

docs/package.json

@@ -26,4 +26,4 @@
   "devDependencies": {
     "zx": "^8.1.4"
   }
-}
+}

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

@@ -84,10 +84,10 @@
   -rmo, --remove-out                  remove output folder if it exists
   --cli <string>                      override path to the cli
   -tp, --test-provider <string>       test provider
   -td, --test-delay <string>          delay between tests in seconds
   --no-cache                          disable LLM result cache
   -v, --verbose                       verbose output
-  -pv, --promptfoo-version [version]  promptfoo version, default is ^0.75.2
+  -pv, --promptfoo-version [version]  promptfoo version, default is 0.78.0
   -os, --out-summary <file>           append output summary in file
   -h, --help                          display help for command
 ```
@@ -291,11 +291,11 @@
 Options:
   -h, --help                   display help for command
 
 Commands:
-  fence <language>             Extracts a code fenced regions of the given type
+  fence <language> <file>      Extracts a code fenced regions of the given type
   pdf <file>                   Parse a PDF into text
   docx <file>                  Parse a DOCX into texts
-  html-to-text [file]          Parse an HTML file into text
+  html-to-text <file>          Parse an HTML file into text
   code <file> [query]          Parse code using tree sitter and executes a
                                query
   tokens [options] <files...>  Count tokens in a set of files
@@ -305,7 +305,7 @@
 ### `parse fence`
 
 ```
-Usage: genaiscript parse fence [options] <language>
+Usage: genaiscript parse fence [options] <language> <file>
 
 Extracts a code fenced regions of the given type
 
@@ -338,7 +338,7 @@
 ### `parse html-to-text`
 
 ```
-Usage: genaiscript parse html-to-text [options] [file]
+Usage: genaiscript parse html-to-text [options] <file>
 
 Parse an HTML file into text
 

docs/src/content/docs/reference/scripts/files.md

@@ -79,6 +79,14 @@
 const data = await workspace.readJSON("data.json")
 ```
 
+### `readXML`
+
+Reads the content of a file as XML.
+
+```ts
+const data = await workspace.readXML("data.xml")
+```
+
 ### `writeText`
 
 Writes text to a file, relative to the workspace root.

docs/src/content/docs/reference/scripts/parsers.md

@@ -100,10 +100,21 @@
 
 ## XML
 
 The `parsers.XML` function parses for the [XML format](https://en.wikipedia.org/wiki/XML).
 
 ```js
-const res = parsers.XML("<xml></xml>")
+const res = parsers.XML('<xml attr="1"><child /></xml>')
+```
+
+Attribute names are prepended with "@_".
+
+```json
+{
+    "xml": {
+        "@_attr": "1",
+        "child": {}
+    }
+}
 ```
 
 ## front matter

packages/cli/package.json

@@ -33,7 +33,7 @@
     "mathjs": "^13.0.3",
     "pdfjs-dist": "4.5.136",
     "tree-sitter-wasms": "^0.1.11",
-    "tsx": "^4.16.5",
+    "tsx": "^4.17.0",
     "typescript": "5.5.4",
     "vectra": "^0.7.6",
     "web-tree-sitter": "^0.22.2",
@@ -42,6 +42,9 @@
   "engines": {
     "node": ">=20.0.0"
   },
+  "peerDependencies": {
+    "promptfoo": "0.78.0"
+  },
   "devDependencies": {
     "@types/dockerode": "^3.3.31",
     "@types/fs-extra": "^11.0.4",
@@ -54,18 +57,16 @@
     "commander": "^12.1.0",
     "dotenv": "^16.4.5",
     "esbuild": "^0.23.0",
-    "execa": "^9.3.0",
+    "execa": "^9.3.1",
     "fs-extra": "^11.2.0",
-    "get-stdin": "^9.0.0",
     "glob": "^11.0.0",
     "memorystream": "^0.3.1",
     "node-sarif-builder": "^3.1.0",
-    "openai": "^4.55.0",
+    "openai": "^4.55.7",
     "ora": "^8.0.1",
     "pretty-bytes": "^6.1.1",
-    "promptfoo": "^0.75.2",
-    "prompts": "^2.4.2",
     "replace-ext": "^2.0.0",
+    "ws": "^8.18.0",
     "zx": "^8.1.4"
   },
   "scripts": {

packages/cli/src/cli.ts

@@ -249,7 +249,7 @@
         .alias("parsers")
         .description("Parse various outputs")
     parser
-        .command("fence <language>")
+        .command("fence <language> <file>")
         .description("Extracts a code fenced regions of the given type")
         .action(parseFence)
 
@@ -264,7 +264,7 @@
         .action(parseDOCX)
 
     parser
-        .command("html-to-text [file]")
+        .command("html-to-text <file>")
         .description("Parse an HTML file into text")
         .action(parseHTMLToText)
 

packages/cli/src/parse.ts

@@ -1,6 +1,5 @@
 import { createProgressSpinner } from "./spinner"
 import replaceExt from "replace-ext"
-import getStdin from "get-stdin"
 import { readFile } from "node:fs/promises"
 import { DOCXTryParse } from "../../core/src/docx"
 import { extractFenced } from "../../core/src/fence"
@@ -12,9 +11,9 @@
 import { YAMLStringify } from "../../core/src/yaml"
 import { resolveTokenEncoder } from "../../core/src/encoders"
 
-export async function parseFence(language: string) {
-    const stdin = await getStdin()
-    const fences = extractFenced(stdin || "").filter(
+export async function parseFence(language: string, file: string) {
+    const res = await parsePdf(file)
+    const fences = extractFenced(res.content || "").filter(
         (f) => f.language === language
     )
     console.log(fences.map((f) => f.content).join("\n\n"))
@@ -32,9 +31,7 @@
 }
 
 export async function parseHTMLToText(file: string) {
-    const html = file
-        ? await readFile(file, { encoding: "utf-8" })
-        : await getStdin()
+    const html = await readFile(file, { encoding: "utf-8" })
     const text = HTMLToText(html)
     console.log(text)
 }

packages/cli/src/version.ts

@@ -1,6 +1,6 @@
 import packageJson from "../package.json"
 
 export const NODE_MIN_VERSION = packageJson.engines.node
-export const PROMPTFOO_VERSION = packageJson.devDependencies.promptfoo
+export const PROMPTFOO_VERSION = packageJson.peerDependencies.promptfoo
 export const TYPESCRIPT_VERSION = packageJson.dependencies.typescript
 export const DOCKERODE_VERSION = packageJson.dependencies.dockerode

packages/core/package.json

@@ -38,7 +38,7 @@
     "gpt-tokenizer": "^2.2.1",
     "html-escaper": "^3.0.3",
     "html-to-text": "^9.0.5",
-    "ignore": "^5.3.1",
+    "ignore": "^5.3.2",
     "inflection": "^3.0.0",
     "ini": "^4.1.3",
     "json5": "^2.2.3",
@@ -50,7 +50,7 @@
     "mime-types": "^2.1.35",
     "minimatch": "^10.0.1",
     "minisearch": "^7.1.0",
-    "openai": "^4.55.0",
+    "openai": "^4.55.7",
     "parse-diff": "^0.11.1",
     "prettier": "^3.3.3",
     "pretty-bytes": "^6.1.1",
@@ -59,7 +59,7 @@
     "toml": "^3.0.0",
     "tree-sitter-wasms": "^0.1.11",
     "ts-dedent": "^2.2.0",
-    "tsx": "^4.16.5",
+    "tsx": "^4.17.0",
     "typescript": "5.5.4",
     "vectra": "^0.7.6",
     "xlsx": "https://cdn.sheetjs.com/xlsx-0.20.2/xlsx-0.20.2.tgz",

packages/core/src/filesystem.ts

@@ -5,6 +5,7 @@ import { readText, writeText } from "./fs"
 import { host } from "./host"
 import { JSON5parse } from "./json5"
 import { logVerbose } from "./util"
+import { XMLParse, XMLTryParse } from "./xml"
 
 export function createFileSystem(): Omit<WorkspaceFileSystem, "grep"> {
     const fs: Omit<WorkspaceFileSystem, "grep"> = {
@@ -61,6 +62,11 @@ export function createFileSystem(): Omit<WorkspaceFileSystem, "grep"> {
             const res = JSON5parse(file.content)
             return res
         },
+        readXML: async (f: string | WorkspaceFile) => {
+            const file = await fs.readText(f)
+            const res = XMLParse(file.content)
+            return res
+        },
     }
     ;(fs as any).readFile = readText
     return Object.freeze(fs)

packages/core/src/promptcontext.ts

@@ -90,6 +90,7 @@ export async function createPromptContext(
     const workspace: WorkspaceFileSystem = {
         readText: (f) => runtimeHost.workspace.readText(f),
         readJSON: (f) => runtimeHost.workspace.readJSON(f),
+        readXML: (f) => runtimeHost.workspace.readXML(f),
         writeText: (f, c) => runtimeHost.workspace.writeText(f, c),
         findFiles: async (pattern, options) => {
             const res = await runtimeHost.workspace.findFiles(pattern, options)

packages/core/src/types/prompt_template.d.ts

@@ -506,6 +506,11 @@ interface WorkspaceFileSystem {
      */
     readJSON(path: string | WorkspaceFile): Promise<any>
 
+    /**
+     * Reads the content of a file and parses to XML, using the XML parser.
+     */
+    readXML(path: string | WorkspaceFile): Promise<any>
+
     /**
      * Writes a file as text to the file system
      * @param path

packages/core/src/xml.test.ts

@@ -1,10 +1,15 @@
-import { XMLParse } from './xml'
-import { describe, test } from 'node:test'
-import assert from 'node:assert/strict'
+import { XMLParse } from "./xml"
+import { describe, test } from "node:test"
+import assert from "node:assert/strict"
 
-describe('xml', () => {
-    test('parse', () => {
-        const x = XMLParse('<root><a>1</a><b>2</b></root>')
+describe("xml", () => {
+    test("parse elements", () => {
+        const x = XMLParse("<root><a>1</a><b>2</b></root>")
         assert.deepStrictEqual(x, { root: { a: 1, b: 2 } })
     })
-})
+
+    test("parse attribute", () => {
+        const x = XMLParse('<root a="1"><b>2</b></root>')
+        assert.deepStrictEqual(x, { root: { b: 2, "@_a": 1 } })
+    })
+})

packages/core/src/xml.ts

@@ -15,6 +15,13 @@
 }
 
 export function XMLParse(text: string, options?: XMLParseOptions) {
-    const parser = new XMLParser(options)
+    const parser = new XMLParser({
+        ignoreAttributes: false,
+        attributeNamePrefix: "@_",
+        allowBooleanAttributes: true,
+        ignoreDeclaration: true,
+        parseAttributeValue: true,
+        ...(options || {}),
+    })
     return parser.parse(text)
 }