format code and fix missing default EMBEDDING_DIM

templates/components/engines/python/agent/tools/__init__.py

@@ -13,7 +13,6 @@ class ToolType:
 
 
 class ToolFactory:
-
     TOOL_SOURCE_PACKAGE_MAP = {
         ToolType.LLAMAHUB: "llama_index.tools",
         ToolType.LOCAL: "app.engine.tools",

templates/components/engines/python/agent/tools/interpreter.py

@@ -3,7 +3,7 @@ import logging
 import base64
 import uuid
 from pydantic import BaseModel
-from typing import List, Tuple, Dict
+from typing import List, Tuple, Dict, Optional
 from llama_index.core.tools import FunctionTool
 from e2b_code_interpreter import CodeInterpreter
 from e2b_code_interpreter.models import Logs
@@ -14,18 +14,18 @@ logger = logging.getLogger(__name__)
 
 class InterpreterExtraResult(BaseModel):
     type: str
-    filename: str
-    url: str
+    content: Optional[str] = None
+    filename: Optional[str] = None
+    url: Optional[str] = None
 
 
 class E2BToolOutput(BaseModel):
     is_error: bool
     logs: Logs
-    results: List[InterpreterExtraResult] = []
+    results: List[InterpreterExtraResult | str] = []
 
 
 class E2BCodeInterpreter:
-
     output_dir = "tool-output"
 
     def __init__(self, api_key: str, filesever_url_prefix: str):
@@ -62,8 +62,9 @@ class E2BCodeInterpreter:
 
     def parse_result(self, result) -> List[InterpreterExtraResult]:
         """
-        The result could include multiple formats (e.g. png, svg, etc.) but encoded in base64
-        We save each result to disk and return saved file metadata (extension, filename, url)
+        The result format could be either a base64 string (png, svg, etc.) or a raw text (text, html, markdown,...)
+        If it's base64, we save each result to disk and return saved file metadata (extension, filename, url),
+        otherwise just return the raw text content
         """
         if not result:
             return []
@@ -72,31 +73,46 @@ class E2BCodeInterpreter:
 
         try:
             formats = result.formats()
-            base64_data_arr = [result[format] for format in formats]
+            data_list = [result[format] for format in formats]
 
-            for ext, base64_data in zip(formats, base64_data_arr):
-                if ext and base64_data:
-                    result = self.save_to_disk(base64_data, ext)
-                    filename = result["filename"]
-                    output.append(
-                        InterpreterExtraResult(
-                            type=ext, filename=filename, url=self.get_file_url(filename)
+            for ext, data in zip(formats, data_list):
+                match ext:
+                    case "png" | "jpeg" | "svg":
+                        result = self.save_to_disk(data, ext)
+                        filename = result["filename"]
+                        output.append(
+                            InterpreterExtraResult(
+                                type=ext,
+                                filename=filename,
+                                url=self.get_file_url(filename),
+                            )
                         )
-                    )
+                        break
+                    case "text" | "html" | "markdown":
+                        output.append(InterpreterExtraResult(type=ext, content=data))
         except Exception as error:
             logger.error("Error when saving data to disk", error)
 
         return output
 
-    def interpret(self, code: str) -> E2BToolOutput:
+    def interpret(self, code: str, file_path: Optional[str] = None) -> E2BToolOutput:
         with CodeInterpreter(api_key=self.api_key) as interpreter:
+            # Upload file to E2B sandbox
+            if file_path is not None:
+                with open(file_path, "rb") as f:
+                    remote_path = interpreter.upload_file(f)
+
+            # Execute the code to analyze the file
             logger.info(
                 f"\n{'='*50}\n> Running following AI-generated code:\n{code}\n{'='*50}"
             )
             exec = interpreter.notebook.exec_cell(code)
 
             if exec.error:
-                output = E2BToolOutput(is_error=True, logs=[exec.error])
+                logger.error(
+                    f"Error when executing code in E2B sandbox: {exec.error} {exec.logs}"
+                )
+                output = E2BToolOutput(is_error=True, logs=exec.logs, results=[])
             else:
                 if len(exec.results) == 0:
                     output = E2BToolOutput(is_error=False, logs=exec.logs, results=[])
@@ -108,9 +124,15 @@ class E2BCodeInterpreter:
             return output
 
 
-def code_interpret(code: str) -> Dict:
+def code_interpret(code: str, local_file_path: str) -> Dict:
     """
-    Execute python code in a Jupyter notebook cell and return any result, stdout, stderr, display_data, and error.
+    Use this tool to analyze the provided data in a sandbox environment.
+    The tool will:
+        1. Upload the provided file from local to the sandbox. The uploaded file path will be /home/user/{filename}
+        2. Execute python code in a Jupyter notebook cell to analyze the uploaded file in the sandbox.
+        3. Get the result from the code in stdout, stderr, display_data, and error.
+    You must to provide the code and the provided file path to run this tool.
+    Your code should read the file from the sandbox path /home/user/{filename}.
     """
     api_key = os.getenv("E2B_API_KEY")
     filesever_url_prefix = os.getenv("FILESERVER_URL_PREFIX")
@@ -126,7 +148,7 @@ def code_interpret(code: str) -> Dict:
     interpreter = E2BCodeInterpreter(
         api_key=api_key, filesever_url_prefix=filesever_url_prefix
     )
-    output = interpreter.interpret(code)
+    output = interpreter.interpret(code, local_file_path)
     return output.dict()
 
 

templates/components/engines/typescript/agent/tools/interpreter.ts

@@ -7,6 +7,7 @@ import path from "node:path";
 
 export type InterpreterParameter = {
   code: string;
+  localFilePath: string;
 };
 
 export type InterpreterToolParams = {
@@ -34,14 +35,21 @@ type InterpreterExtraType =
 
 export type InterpreterExtraResult = {
   type: InterpreterExtraType;
+  content?: string;
   filename: string;
   url: string;
 };
 
 const DEFAULT_META_DATA: ToolMetadata<JSONSchemaType<InterpreterParameter>> = {
   name: "interpreter",
-  description:
-    "Execute python code in a Jupyter notebook cell and return any result, stdout, stderr, display_data, and error.",
+  description: `Use this tool to analyze the provided data in a sandbox environment.
+    The tool will:
+        1. Upload the provided file from local to the sandbox. The uploaded file path will be /home/user/{filename}
+        2. Execute python code in a Jupyter notebook cell to analyze the uploaded file in the sandbox.
+        3. Get the result from the code in stdout, stderr, display_data, and error.
+    You must to provide the code and the provided file path to run this tool.
+    Your code should read the file from the sandbox path /home/user/{filename}.
+    `,
   parameters: {
     type: "object",
     properties: {
@@ -49,6 +57,10 @@ const DEFAULT_META_DATA: ToolMetadata<JSONSchemaType<InterpreterParameter>> = {
         type: "string",
         description: "The python code to execute in a single cell.",
       },
+      localFilePath: {
+        type: "string",
+        description: "The local file path to upload to the sandbox.",
+      },
     },
     required: ["code"],
   },
@@ -88,11 +100,22 @@ export class InterpreterTool implements BaseTool<InterpreterParameter> {
     return this.codeInterpreter;
   }
 
-  public async codeInterpret(code: string): Promise<InterpreterToolOutput> {
+  public async codeInterpret(
+    code: string,
+    localFilePath: string,
+  ): Promise<InterpreterToolOutput> {
+    const interpreter = await this.initInterpreter();
+    // Upload file to sandbox
+    console.log(`Uploading file ${localFilePath} to sandbox`);
+    const fileBuffer = fs.readFileSync(localFilePath);
+    const fileName = path.basename(localFilePath);
+    await interpreter.uploadFile(fileBuffer, fileName);
+    console.log(`Uploaded file ${fileName} to sandbox`);
+
+    // Execute code in sandbox
     console.log(
       `\n${"=".repeat(50)}\n> Running following AI-generated code:\n${code}\n${"=".repeat(50)}`,
     );
-    const interpreter = await this.initInterpreter();
     const exec = await interpreter.notebook.execCell(code);
     if (exec.error) console.error("[Code Interpreter error]", exec.error);
     const extraResult = await this.getExtraResult(exec.results[0]);
@@ -105,7 +128,7 @@ export class InterpreterTool implements BaseTool<InterpreterParameter> {
   }
 
   async call(input: InterpreterParameter): Promise<InterpreterToolOutput> {
-    const result = await this.codeInterpret(input.code);
+    const result = await this.codeInterpret(input.code, input.localFilePath);
     await this.codeInterpreter?.close();
     return result;
   }
@@ -119,18 +142,26 @@ export class InterpreterTool implements BaseTool<InterpreterParameter> {
     try {
       const formats = res.formats(); // formats available for the result. Eg: ['png', ...]
       const base64DataArr = formats.map((f) => res[f as keyof Result]); // get base64 data for each format
+      console.log("data", base64DataArr);
 
       // save base64 data to file and return the url
       for (let i = 0; i < formats.length; i++) {
         const ext = formats[i];
         const base64Data = base64DataArr[i];
-        if (ext && base64Data) {
+        if (ext === "png" && base64Data) {
           const { filename } = this.saveToDisk(base64Data, ext);
           output.push({
             type: ext as InterpreterExtraType,
             filename,
             url: this.getFileUrl(filename),
           });
+        } else {
+          output.push({
+            type: ext as InterpreterExtraType,
+            content: base64Data,
+            filename: `output.${ext}`,
+            url: "",
+          });
         }
       }
     } catch (error) {

templates/components/vectordbs/python/milvus/vectordb.py

@@ -1,4 +1,5 @@
 import os
+
 from llama_index.vector_stores.milvus import MilvusVectorStore
 
 
@@ -15,6 +16,6 @@ def get_vector_store():
         user=os.getenv("MILVUS_USERNAME"),
         password=os.getenv("MILVUS_PASSWORD"),
         collection_name=collection,
-        dim=int(os.getenv("EMBEDDING_DIM")),
+        dim=int(os.getenv("EMBEDDING_DIM", 768)),
     )
     return store

templates/types/streaming/fastapi/app/api/routers/models.py

@@ -1,5 +1,6 @@
 import os
 import logging
+import tempfile
 from pydantic import BaseModel, Field, validator
 from pydantic.alias_generators import to_camel
 from typing import List, Any, Optional, Dict
@@ -21,6 +22,29 @@ class CsvFile(BaseModel):
     filesize: int
     id: str
     type: str
+    local_file_path: Optional[str] = None
+
+    def __init__(self, **data):
+        super().__init__(**data)
+
+        # Write the content to a temporary file
+        saved_path = self.write_to_temp_file(self.content)
+        self.local_file_path = saved_path
+
+    @staticmethod
+    def write_to_temp_file(file_content: str) -> str:
+        """
+        Write the content to a temporary file and return the file path
+        """
+        csv_file = tempfile.NamedTemporaryFile(mode="w", delete=False, suffix=".csv")
+        csv_file.write(file_content)
+        file_path = csv_file.name
+        return file_path
+
+    def __del__(self):
+        # Remove the temporary file once the object is deleted
+        if self.local_file_path:
+            os.remove(self.local_file_path)
 
 
 class DataParserOptions(BaseModel):
@@ -47,9 +71,16 @@ class DataParserOptions(BaseModel):
 
     def to_raw_content(self) -> str:
         if self.csv_files is not None and len(self.csv_files) > 0:
-            return "Use data from following CSV raw contents" + "\n".join(
-                [f"```csv\n{csv_file.content}\n```" for csv_file in self.csv_files]
-            )
+            saved_path = self.csv_files[0].local_file_path
+            saved_file_name = os.path.basename(saved_path)
+            content = self.csv_files[0].content
+            csv_meta = {
+                "local_file_path": saved_path,
+                "example_data": content[: min(200, len(content))],
+                "sandbox_file_path": f"/home/user/{saved_file_name}",
+            }
+
+            return f"Provided CSV file metadata:\n{csv_meta}"
 
     def to_response_data(self) -> list[dict] | None:
         output = []

templates/types/streaming/fastapi/app/engine/generate.py

@@ -21,7 +21,6 @@ STORAGE_DIR = os.getenv("STORAGE_DIR", "storage")
 
 
 def get_doc_store():
-
     # If the storage directory is there, load the document store from it.
     # If not, set up an in-memory document store since we can't load from a directory that doesn't exist.
     if os.path.exists(STORAGE_DIR):

templates/types/streaming/nextjs/app/api/chat/llamaindex-stream.ts

@@ -14,11 +14,13 @@ import {
 } from "llamaindex";
 
 import { AgentStreamChatResponse } from "llamaindex/agent/base";
+import path from "path";
 import {
   CsvFile,
   appendCsvData,
   appendImageData,
   appendSourceData,
+  writeTempCsvFiles,
 } from "./stream-helper";
 
 type LlamaIndexResponse =
@@ -51,14 +53,22 @@ export const convertMessageContent = (
   }
 
   if (additionalData?.csvFiles?.length) {
-    const rawContents = additionalData.csvFiles.map((csv) => {
-      return "```csv\n" + csv.content + "\n```";
-    });
+    const tmpFile = writeTempCsvFiles(additionalData.csvFiles);
+    // Get a few lines of the CSV file as sample content
+    const sampleContent = additionalData.csvFiles
+      .map((csv) => csv.content.split("\n").slice(1, 4).join("\n"))
+      .join("\n\n");
+    const metadata = {
+      localFilePath: tmpFile.name,
+      sampleContent: sampleContent,
+      sandboxFilePath: `/home/user/${path.basename(tmpFile.name)}`,
+    };
     const csvContent =
-      "Use data from following CSV raw contents:\n" + rawContents.join("\n\n");
+      "Provided CSV file metadata:\n" + JSON.stringify(metadata, null, 2);
+    console.log(csvContent);
     content.push({
       type: "text",
-      text: `${csvContent}\n\n${textMessage}`,
+      text: `${textMessage}\n\n${csvContent}`,
     });
   }
 

templates/types/streaming/nextjs/app/api/chat/stream-helper.ts

@@ -1,4 +1,5 @@
 import { StreamData } from "ai";
+import fs from "fs";
 import {
   CallbackManager,
   Metadata,
@@ -6,6 +7,7 @@ import {
   ToolCall,
   ToolOutput,
 } from "llamaindex";
+import tmp from "tmp";
 
 export function appendImageData(data: StreamData, imageUrl?: string) {
   if (!imageUrl) return;
@@ -127,6 +129,7 @@ export type CsvFile = {
   filename: string;
   filesize: number;
   id: string;
+  localFilePath: string;
 };
 
 export function appendCsvData(data: StreamData, csvFiles?: CsvFile[]) {
@@ -138,3 +141,10 @@ export function appendCsvData(data: StreamData, csvFiles?: CsvFile[]) {
     },
   });
 }
+
+export function writeTempCsvFiles(csvFiles: CsvFile[]) {
+  const csvFile = csvFiles[0];
+  const tmpFile = tmp.fileSync({ postfix: ".csv" });
+  fs.writeFileSync(tmpFile.name, csvFile.content);
+  return tmpFile;
+}

templates/types/streaming/nextjs/package.json

@@ -36,7 +36,8 @@
     "vaul": "^0.9.1",
     "@llamaindex/pdf-viewer": "^1.1.1",
     "@e2b/code-interpreter": "^0.0.5",
-    "uuid": "^9.0.1"
+    "uuid": "^9.0.1",
+    "tmp": "^0.2.3"
   },
   "devDependencies": {
     "@types/node": "^20.10.3",