feat: add Language and parsingInstruction to LlamaParseReader (#779)

Co-authored-by: Alex Yang <[email protected]>

examples/readers/package.json

@@ -10,7 +10,8 @@
     "start:markdown": "node --import tsx ./src/markdown.ts",
     "start:pdf": "node --import tsx ./src/pdf.ts",
     "start:llamaparse": "node --import tsx ./src/llamaparse.ts",
-    "start:notion": "node --import tsx ./src/notion.ts"
+    "start:notion": "node --import tsx ./src/notion.ts",
+    "start:llamaparse2": "node --import tsx ./src/llamaparse_2.ts"
   },
   "dependencies": {
     "llamaindex": "*"

examples/readers/src/llamaparse_2.ts

@@ -0,0 +1,26 @@
+import fs from "fs/promises";
+import { LlamaParseReader } from "llamaindex";
+
+async function main() {
+  // Load PDF using LlamaParse. set apiKey here or in environment variable LLAMA_CLOUD_API_KEY
+  const reader = new LlamaParseReader({
+    resultType: "markdown",
+    language: "en",
+    parsingInstruction:
+      "The provided document is a manga comic book. Most pages do NOT have title. It does not contain tables. Try to reconstruct the dialogue happening in a cohesive way. Output any math equation in LATEX markdown (between $$)",
+  });
+  const documents = await reader.loadData("../data/manga.pdf"); // The manga.pdf in the data folder is just a copy of the TOS, due to copyright laws. You have to place your own. I used "The Manga Guide to Calculus" by Hiroyuki Kojima
+
+  // Assuming documents contain an array of pages or sections
+  const parsedManga = documents.map((page) => page.text).join("\n---\n");
+
+  // Output the parsed manga to .md file. Will be placed in ../example/readers/
+  try {
+    await fs.writeFile("./parsedManga.md", parsedManga);
+    console.log("Output successfully written to parsedManga.md");
+  } catch (err) {
+    console.error("Error writing to file:", err);
+  }
+}
+
+main().catch(console.error);

packages/core/src/readers/LlamaParseReader.ts

@@ -1,9 +1,7 @@
 import { defaultFS, getEnv, type GenericFileSystem } from "@llamaindex/env";
 import { filetypemime } from "magic-bytes.js";
 import { Document } from "../Node.js";
-import type { FileReader } from "./type.js";
-
-type ResultType = "text" | "markdown" | "json";
+import type { FileReader, Language, ResultType } from "./type.js";
 
 /**
  * Represents a reader for parsing files using the LlamaParse API.
@@ -20,7 +18,12 @@ export class LlamaParseReader implements FileReader {
   checkInterval = 1;
   // Whether to print the progress of the parsing.
   verbose = true;
+  // The result type for the parser.
   resultType: ResultType = "text";
+  // The language of the text to parse.
+  language: Language = "en";
+  // The parsing instruction for the parser.
+  parsingInstruction: string = "";
 
   constructor(params: Partial<LlamaParseReader> = {}) {
     Object.assign(this, params);
@@ -48,6 +51,8 @@ export class LlamaParseReader implements FileReader {
     const mimeType = await this.getMimeType(data);
     const body = new FormData();
     body.set("file", new Blob([data], { type: mimeType }), file);
+    body.append("language", this.language);
+    body.append("parsingInstruction", this.parsingInstruction);
 
     const headers = {
       Authorization: `Bearer ${this.apiKey}`,

packages/core/src/readers/type.ts

@@ -14,3 +14,91 @@ export interface BaseReader {
 export interface FileReader extends BaseReader {
   loadData(filePath: string, fs?: CompleteFileSystem): Promise<Document[]>;
 }
+
+// For LlamaParseReader.ts
+
+export type ResultType = "text" | "markdown" | "json";
+export type Language =
+  | "abq"
+  | "ady"
+  | "af"
+  | "ang"
+  | "ar"
+  | "as"
+  | "ava"
+  | "az"
+  | "be"
+  | "bg"
+  | "bh"
+  | "bho"
+  | "bn"
+  | "bs"
+  | "ch_sim"
+  | "ch_tra"
+  | "che"
+  | "cs"
+  | "cy"
+  | "da"
+  | "dar"
+  | "de"
+  | "en"
+  | "es"
+  | "et"
+  | "fa"
+  | "fr"
+  | "ga"
+  | "gom"
+  | "hi"
+  | "hr"
+  | "hu"
+  | "id"
+  | "inh"
+  | "is"
+  | "it"
+  | "ja"
+  | "kbd"
+  | "kn"
+  | "ko"
+  | "ku"
+  | "la"
+  | "lbe"
+  | "lez"
+  | "lt"
+  | "lv"
+  | "mah"
+  | "mai"
+  | "mi"
+  | "mn"
+  | "mr"
+  | "ms"
+  | "mt"
+  | "ne"
+  | "new"
+  | "nl"
+  | "no"
+  | "oc"
+  | "pi"
+  | "pl"
+  | "pt"
+  | "ro"
+  | "ru"
+  | "rs_cyrillic"
+  | "rs_latin"
+  | "sck"
+  | "sk"
+  | "sl"
+  | "sq"
+  | "sv"
+  | "sw"
+  | "ta"
+  | "tab"
+  | "te"
+  | "th"
+  | "tjk"
+  | "tl"
+  | "tr"
+  | "ug"
+  | "uk"
+  | "ur"
+  | "uz"
+  | "vi";