feat: Updated packages/create-llama/templates/type

.github/workflows/e2e.yml

@@ -8,6 +8,10 @@ on:
       - ".github/workflows/e2e.yml"
     branches: [main]
 
+env:
+      CI:
+        secrete: ''
+      PLAYWRIGHT_BROWSERS_PATH: './packages/create-llama/playwright-report/'
 jobs:
   e2e:
     name: create-llama
@@ -35,7 +39,7 @@ jobs:
         run: pnpm run build
         working-directory: ./packages/create-llama
       - name: Run Playwright tests
-        run: pnpm exec playwright test
+        run: pnpm exec playwright test 2>&1 | tee error_logs.txt
         env:
           OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
         working-directory: ./packages/create-llama
@@ -44,4 +48,5 @@ jobs:
         with:
           name: playwright-report
           path: ./packages/create-llama/playwright-report/
+          - ./error_logs.txt
           retention-days: 30

packages/core/src/llm/LLM.ts

@@ -10,7 +10,6 @@ import {
 
 import { ChatCompletionMessageParam } from "openai/resources";
 import { LLMOptions } from "portkey-ai";
-import { MessageContent } from "../ChatEngine";
 import { globalsHelper, Tokenizers } from "../GlobalsHelper";
 import {
   ANTHROPIC_AI_PROMPT,
@@ -48,8 +47,9 @@ export interface ChatResponse {
   delta?: string;
 }
 
-// NOTE in case we need CompletionResponse to diverge from ChatResponse in the future
-export type CompletionResponse = ChatResponse;
+export interface CompletionResponse {
+  text: string;
+}
 
 export interface LLMMetadata {
   model: string;
@@ -60,40 +60,57 @@ export interface LLMMetadata {
   tokenizer: Tokenizers | undefined;
 }
 
+export interface LLMChatParamsBase {
+  messages: ChatMessage[];
+  parentEvent?: Event;
+  extraParams?: Record<string, any>;
+}
+
+export interface LLMChatParamsStreaming extends LLMChatParamsBase {
+  stream: true;
+}
+
+export interface LLMChatParamsNonStreaming extends LLMChatParamsBase {
+  stream?: false | null;
+}
+
+export interface LLMCompletionParamsBase {
+  prompt: string;
+  parentEvent?: Event;
+}
+
+export interface LLMCompletionParamsStreaming extends LLMCompletionParamsBase {
+  stream: true;
+}
+
+export interface LLMCompletionParamsNonStreaming
+  extends LLMCompletionParamsBase {
+  stream?: false | null;
+}
+
 /**
  * Unified language model interface
  */
 export interface LLM {
   metadata: LLMMetadata;
-  // Whether a LLM has streaming support
-  hasStreaming: boolean;
   /**
    * Get a chat response from the LLM
-   * @param messages
    *
-   * The return type of chat() and complete() are set by the "streaming" parameter being set to True.
+   * @param params
    */
-  chat<
-    T extends boolean | undefined = undefined,
-    R = T extends true ? AsyncGenerator<string, void, unknown> : ChatResponse,
-  >(
-    messages: ChatMessage[],
-    parentEvent?: Event,
-    streaming?: T,
-  ): Promise<R>;
+  chat(params: LLMChatParamsStreaming): Promise<AsyncIterable<string>>;
+  chat(params: LLMChatParamsNonStreaming): Promise<ChatResponse>;
 
   /**
    * Get a prompt completion from the LLM
-   * @param prompt the prompt to complete
+   * @param params
    */
-  complete<
-    T extends boolean | undefined = undefined,
-    R = T extends true ? AsyncGenerator<string, void, unknown> : ChatResponse,
-  >(
-    prompt: MessageContent,
-    parentEvent?: Event,
-    streaming?: T,
-  ): Promise<R>;
+  complete(
+    params: LLMCompletionParamsStreaming,
+  ): Promise<AsyncIterable<string>>;
+  complete(
+    params: LLMCompletionParamsNonStreaming,
+  ): Promise<CompletionResponse>;
 
   /**
    * Calculates the number of tokens needed for the given chat messages
@@ -101,6 +118,39 @@ export interface LLM {
   tokens(messages: ChatMessage[]): number;
 }
 
+abstract class BaseLLM implements LLM {
+  abstract metadata: LLMMetadata;
+
+  complete(
+    params: LLMCompletionParamsStreaming,
+  ): Promise<AsyncIterable<string>>;
+  complete(
+    params: LLMCompletionParamsNonStreaming,
+  ): Promise<CompletionResponse>;
+  async complete(
+    params: LLMCompletionParamsStreaming | LLMCompletionParamsNonStreaming,
+  ): Promise<CompletionResponse | AsyncIterable<string>> {
+    const { prompt, parentEvent, stream } = params;
+    if (stream) {
+      return this.chat({
+        messages: [{ content: prompt, role: "user" }],
+        parentEvent,
+        stream: true,
+      });
+    }
+    const chatResponse = await this.chat({
+      messages: [{ content: prompt, role: "user" }],
+      parentEvent,
+    });
+    return { text: chatResponse.message.content as string };
+  }
+
+  abstract chat(params: LLMChatParamsStreaming): Promise<AsyncIterable<string>>;
+  abstract chat(params: LLMChatParamsNonStreaming): Promise<ChatResponse>;
+
+  abstract tokens(messages: ChatMessage[]): number;
+}
+
 export const GPT4_MODELS = {
   "gpt-4": { contextWindow: 8192 },
   "gpt-4-32k": { contextWindow: 32768 },
@@ -125,9 +175,7 @@ export const ALL_AVAILABLE_OPENAI_MODELS = {
 /**
  * OpenAI LLM implementation
  */
-export class OpenAI implements LLM {
-  hasStreaming: boolean = true;
-
+export class OpenAI extends BaseLLM {
   // Per completion OpenAI params
   model: keyof typeof ALL_AVAILABLE_OPENAI_MODELS;
   temperature: number;
@@ -135,7 +183,7 @@ export class OpenAI implements LLM {
   maxTokens?: number;
   additionalChatOptions?: Omit<
     Partial<OpenAILLM.Chat.ChatCompletionCreateParams>,
-    "max_tokens" | "messages" | "model" | "temperature" | "top_p" | "streaming"
+    "max_tokens" | "messages" | "model" | "temperature" | "top_p" | "stream"
   >;
 
   // OpenAI session params
@@ -155,6 +203,7 @@ export class OpenAI implements LLM {
       azure?: AzureOpenAIConfig;
     },
   ) {
+    super();
     this.model = init?.model ?? "gpt-3.5-turbo";
     this.temperature = init?.temperature ?? 0.1;
     this.topP = init?.topP ?? 1;
@@ -247,10 +296,12 @@ export class OpenAI implements LLM {
     }
   }
 
-  async chat<
-    T extends boolean | undefined = undefined,
-    R = T extends true ? AsyncGenerator<string, void, unknown> : ChatResponse,
-  >(messages: ChatMessage[], parentEvent?: Event, streaming?: T): Promise<R> {
+  chat(params: LLMChatParamsStreaming): Promise<AsyncIterable<string>>;
+  chat(params: LLMChatParamsNonStreaming): Promise<ChatResponse>;
+  async chat(
+    params: LLMChatParamsNonStreaming | LLMChatParamsStreaming,
+  ): Promise<ChatResponse | AsyncIterable<string>> {
+    const { messages, parentEvent, stream } = params;
     const baseRequestParams: OpenAILLM.Chat.ChatCompletionCreateParams = {
       model: this.model,
       temperature: this.temperature,
@@ -266,11 +317,8 @@ export class OpenAI implements LLM {
       ...this.additionalChatOptions,
     };
     // Streaming
-    if (streaming) {
-      if (!this.hasStreaming) {
-        throw Error("No streaming support for this LLM.");
-      }
-      return this.streamChat(messages, parentEvent) as R;
+    if (stream) {
+      return this.streamChat(params);
     }
     // Non-streaming
     const response = await this.session.openai.chat.completions.create({
@@ -281,27 +329,13 @@ export class OpenAI implements LLM {
     const content = response.choices[0].message?.content ?? "";
     return {
       message: { content, role: response.choices[0].message.role },
-    } as R;
+    };
   }
 
-  async complete<
-    T extends boolean | undefined = undefined,
-    R = T extends true ? AsyncGenerator<string, void, unknown> : ChatResponse,
-  >(prompt: string, parentEvent?: Event, streaming?: T): Promise<R> {
-    return this.chat(
-      [{ content: prompt, role: "user" }],
-      parentEvent,
-      streaming,
-    );
-  }
-
-  //We can wrap a stream in a generator to add some additional logging behavior
-  //For future edits: syntax for generator type is <typeof Yield, typeof Return, typeof Accept>
-  //"typeof Accept" refers to what types you'll accept when you manually call generator.next(<AcceptType>)
-  protected async *streamChat(
-    messages: ChatMessage[],
-    parentEvent?: Event,
-  ): AsyncGenerator<string, void, unknown> {
+  protected async *streamChat({
+    messages,
+    parentEvent,
+  }: LLMChatParamsStreaming): AsyncIterable<string> {
     const baseRequestParams: OpenAILLM.Chat.ChatCompletionCreateParams = {
       model: this.model,
       temperature: this.temperature,
@@ -360,14 +394,6 @@ export class OpenAI implements LLM {
     }
     return;
   }
-
-  //streamComplete doesn't need to be async because it's child function is already async
-  protected streamComplete(
-    query: string,
-    parentEvent?: Event,
-  ): AsyncGenerator<string, void, unknown> {
-    return this.streamChat([{ content: query, role: "user" }], parentEvent);
-  }
 }
 
 export const ALL_AVAILABLE_LLAMADEUCE_MODELS = {
@@ -425,16 +451,16 @@ export enum DeuceChatStrategy {
 /**
  * Llama2 LLM implementation
  */
-export class LlamaDeuce implements LLM {
+export class LlamaDeuce extends BaseLLM {
   model: keyof typeof ALL_AVAILABLE_LLAMADEUCE_MODELS;
   chatStrategy: DeuceChatStrategy;
   temperature: number;
   topP: number;
   maxTokens?: number;
   replicateSession: ReplicateSession;
-  hasStreaming: boolean;
 
   constructor(init?: Partial<LlamaDeuce>) {
+    super();
     this.model = init?.model ?? "Llama-2-70b-chat-4bit";
     this.chatStrategy =
       init?.chatStrategy ??
@@ -447,7 +473,6 @@ export class LlamaDeuce implements LLM {
       init?.maxTokens ??
       ALL_AVAILABLE_LLAMADEUCE_MODELS[this.model].contextWindow; // For Replicate, the default is 500 tokens which is too low.
     this.replicateSession = init?.replicateSession ?? new ReplicateSession();
-    this.hasStreaming = init?.hasStreaming ?? false;
   }
 
   tokens(messages: ChatMessage[]): number {
@@ -592,10 +617,12 @@ If a question does not make any sense, or is not factually coherent, explain why
     };
   }
 
-  async chat<
-    T extends boolean | undefined = undefined,
-    R = T extends true ? AsyncGenerator<string, void, unknown> : ChatResponse,
-  >(messages: ChatMessage[], _parentEvent?: Event, streaming?: T): Promise<R> {
+  chat(params: LLMChatParamsStreaming): Promise<AsyncIterable<string>>;
+  chat(params: LLMChatParamsNonStreaming): Promise<ChatResponse>;
+  async chat(
+    params: LLMChatParamsNonStreaming | LLMChatParamsStreaming,
+  ): Promise<ChatResponse | AsyncIterable<string>> {
+    const { messages, parentEvent, stream } = params;
     const api = ALL_AVAILABLE_LLAMADEUCE_MODELS[this.model]
       .replicateApi as `${string}/${string}:${string}`;
 
@@ -629,14 +656,7 @@ If a question does not make any sense, or is not factually coherent, explain why
         //^ We need to do this because Replicate returns a list of strings (for streaming functionality which is not exposed by the run function)
         role: "assistant",
       },
-    } as R;
-  }
-
-  async complete<
-    T extends boolean | undefined = undefined,
-    R = T extends true ? AsyncGenerator<string, void, unknown> : ChatResponse,
-  >(prompt: string, parentEvent?: Event, streaming?: T): Promise<R> {
-    return this.chat([{ content: prompt, role: "user" }], parentEvent);
+    };
   }
 }
 
@@ -650,9 +670,7 @@ export const ALL_AVAILABLE_ANTHROPIC_MODELS = {
  * Anthropic LLM implementation
  */
 
-export class Anthropic implements LLM {
-  hasStreaming: boolean = true;
-
+export class Anthropic extends BaseLLM {
   // Per completion Anthropic params
   model: keyof typeof ALL_AVAILABLE_ANTHROPIC_MODELS;
   temperature: number;
@@ -668,6 +686,7 @@ export class Anthropic implements LLM {
   callbackManager?: CallbackManager;
 
   constructor(init?: Partial<Anthropic>) {
+    super();
     this.model = init?.model ?? "claude-2";
     this.temperature = init?.temperature ?? 0.1;
     this.topP = init?.topP ?? 0.999; // Per Ben Mann
@@ -719,20 +738,15 @@ export class Anthropic implements LLM {
     );
   }
 
-  async chat<
-    T extends boolean | undefined = undefined,
-    R = T extends true ? AsyncGenerator<string, void, unknown> : ChatResponse,
-  >(
-    messages: ChatMessage[],
-    parentEvent?: Event | undefined,
-    streaming?: T,
-  ): Promise<R> {
+  chat(params: LLMChatParamsStreaming): Promise<AsyncIterable<string>>;
+  chat(params: LLMChatParamsNonStreaming): Promise<ChatResponse>;
+  async chat(
+    params: LLMChatParamsNonStreaming | LLMChatParamsStreaming,
+  ): Promise<ChatResponse | AsyncIterable<string>> {
+    const { messages, parentEvent, stream } = params;
     //Streaming
-    if (streaming) {
-      if (!this.hasStreaming) {
-        throw Error("No streaming support for this LLM.");
-      }
-      return this.streamChat(messages, parentEvent) as R;
+    if (stream) {
+      return this.streamChat(messages, parentEvent);
     }
 
     //Non-streaming
@@ -748,13 +762,13 @@ export class Anthropic implements LLM {
       message: { content: response.completion.trimStart(), role: "assistant" },
       //^ We're trimming the start because Anthropic often starts with a space in the response
       // That space will be re-added when we generate the next prompt.
-    } as R;
+    };
   }
 
   protected async *streamChat(
     messages: ChatMessage[],
     parentEvent?: Event | undefined,
-  ): AsyncGenerator<string, void, unknown> {
+  ): AsyncIterable<string> {
     // AsyncIterable<AnthropicStreamToken>
     const stream: AsyncIterable<AnthropicStreamToken> =
       await this.session.anthropic.completions.create({
@@ -775,36 +789,9 @@ export class Anthropic implements LLM {
     }
     return;
   }
-
-  async complete<
-    T extends boolean | undefined = undefined,
-    R = T extends true ? AsyncGenerator<string, void, unknown> : ChatResponse,
-  >(
-    prompt: string,
-    parentEvent?: Event | undefined,
-    streaming?: T,
-  ): Promise<R> {
-    if (streaming) {
-      return this.streamComplete(prompt, parentEvent) as R;
-    }
-    return this.chat(
-      [{ content: prompt, role: "user" }],
-      parentEvent,
-      streaming,
-    ) as R;
-  }
-
-  protected streamComplete(
-    prompt: string,
-    parentEvent?: Event | undefined,
-  ): AsyncGenerator<string, void, unknown> {
-    return this.streamChat([{ content: prompt, role: "user" }], parentEvent);
-  }
 }
 
-export class Portkey implements LLM {
-  hasStreaming: boolean = true;
-
+export class Portkey extends BaseLLM {
   apiKey?: string = undefined;
   baseURL?: string = undefined;
   mode?: string = undefined;
@@ -813,6 +800,7 @@ export class Portkey implements LLM {
   callbackManager?: CallbackManager;
 
   constructor(init?: Partial<Portkey>) {
+    super();
     this.apiKey = init?.apiKey;
     this.baseURL = init?.baseURL;
     this.mode = init?.mode;
@@ -834,45 +822,27 @@ export class Portkey implements LLM {
     throw new Error("metadata not implemented for Portkey");
   }
 
-  async chat<
-    T extends boolean | undefined = undefined,
-    R = T extends true ? AsyncGenerator<string, void, unknown> : ChatResponse,
-  >(
-    messages: ChatMessage[],
-    parentEvent?: Event | undefined,
-    streaming?: T,
-    params?: Record<string, any>,
-  ): Promise<R> {
-    if (streaming) {
-      return this.streamChat(messages, parentEvent, params) as R;
+  chat(params: LLMChatParamsStreaming): Promise<AsyncIterable<string>>;
+  chat(params: LLMChatParamsNonStreaming): Promise<ChatResponse>;
+  async chat(
+    params: LLMChatParamsNonStreaming | LLMChatParamsStreaming,
+  ): Promise<ChatResponse | AsyncIterable<string>> {
+    const { messages, parentEvent, stream, extraParams } = params;
+    if (stream) {
+      return this.streamChat(messages, parentEvent, extraParams);
     } else {
-      const resolvedParams = params || {};
+      const bodyParams = extraParams || {};
       const response = await this.session.portkey.chatCompletions.create({
         messages,
-        ...resolvedParams,
+        ...bodyParams,
       });
 
       const content = response.choices[0].message?.content ?? "";
       const role = response.choices[0].message?.role || "assistant";
-      return { message: { content, role: role as MessageType } } as R;
+      return { message: { content, role: role as MessageType } };
     }
   }
 
-  async complete<
-    T extends boolean | undefined = undefined,
-    R = T extends true ? AsyncGenerator<string, void, unknown> : ChatResponse,
-  >(
-    prompt: string,
-    parentEvent?: Event | undefined,
-    streaming?: T,
-  ): Promise<R> {
-    return this.chat(
-      [{ content: prompt, role: "user" }],
-      parentEvent,
-      streaming,
-    );
-  }
-
   async *streamChat(
     messages: ChatMessage[],
     parentEvent?: Event,
@@ -919,11 +889,4 @@ export class Portkey implements LLM {
     }
     return;
   }
-
-  streamComplete(
-    query: string,
-    parentEvent?: Event,
-  ): AsyncGenerator<string, void, unknown> {
-    return this.streamChat([{ content: query, role: "user" }], parentEvent);
-  }
 }

packages/core/src/llm/mistral.ts

@@ -43,6 +43,7 @@ export class MistralAISession {
  */
 export class MistralAI implements LLM {
   hasStreaming: boolean = true;
+  
 
   // Per completion MistralAI params
   model: keyof typeof ALL_AVAILABLE_MISTRAL_MODELS;
@@ -94,16 +95,13 @@ export class MistralAI implements LLM {
     };
   }
 
-  async chat<
-    T extends boolean | undefined = undefined,
-    R = T extends true ? AsyncGenerator<string, void, unknown> : ChatResponse,
-  >(messages: ChatMessage[], parentEvent?: Event, streaming?: T): Promise<R> {
+  async chat(messages: ChatMessage[], parentEvent?: Event, streaming?: boolean): Promise<any> {
     // Streaming
     if (streaming) {
       if (!this.hasStreaming) {
         throw Error("No streaming support for this LLM.");
       }
-      return this.streamChat(messages, parentEvent) as R;
+      return this.streamChat(messages, parentEvent) as any;
     }
     // Non-streaming
     const client = await this.session.getClient();
@@ -114,10 +112,7 @@ export class MistralAI implements LLM {
     } as R;
   }
 
-  async complete<
-    T extends boolean | undefined = undefined,
-    R = T extends true ? AsyncGenerator<string, void, unknown> : ChatResponse,
-  >(prompt: string, parentEvent?: Event, streaming?: T): Promise<R> {
+  async complete(prompt: string, parentEvent?: Event, streaming?: boolean): Promise<ChatResponse> {
     return this.chat(
       [{ content: prompt, role: "user" }],
       parentEvent,
@@ -148,7 +143,6 @@ export class MistralAI implements LLM {
     var idx_counter: number = 0;
     for await (const part of chunkStream) {
       if (!part.choices.length) continue;
-
       part.choices[0].index = idx_counter;
       const isDone: boolean =
         part.choices[0].finish_reason === "stop" ? true : false;

packages/core/src/llm/openai.ts

@@ -1,6 +1,8 @@
 import _ from "lodash";
 import OpenAI, { ClientOptions } from "openai";
 
+export { OpenAI };
+
 export class AzureOpenAI extends OpenAI {
   protected override authHeaders() {
     return { "api-key": this.apiKey };

packages/core/src/tests/CallbackManager.test.ts

@@ -1,5 +1,5 @@
-import { Document } from "../Node";
-import { ServiceContext, serviceContextFromDefaults } from "../ServiceContext";
+
+
 import {
   CallbackManager,
   RetrievalCallbackResponse,
@@ -8,9 +8,18 @@ import {
 import { OpenAIEmbedding } from "../embeddings";
 import { SummaryIndex } from "../indices/summary";
 import { VectorStoreIndex } from "../indices/vectorStore/VectorStoreIndex";
-import { OpenAI } from "../llm/LLM";
+import { OpenAI, Event, ClientOptions } from "../llm/openai";
+
+
+
+// Mock the OpenAI getOpenAISession function during testing
+jest.mock("../llm/openai", () => {
+  return {
+    getOpenAISession: jest.fn().mockImplementation(() => null),
+  };
+});
 import { ResponseSynthesizer, SimpleResponseBuilder } from "../synthesizers";
-import { mockEmbeddingModel, mockLlmGeneration } from "./utility/mockOpenAI";
+
 
 // Mock the OpenAI getOpenAISession function during testing
 jest.mock("../llm/openai", () => {
@@ -37,18 +46,18 @@ describe("CallbackManager: onLLMStream and onRetrieve", () => {
     });
 
     const languageModel = new OpenAI({
-      model: "gpt-3.5-turbo",
+      model: "davinci",
       callbackManager,
     });
     mockLlmGeneration({ languageModel, callbackManager });
 
-    const embedModel = new OpenAIEmbedding();
-    mockEmbeddingModel(embedModel);
+    
+    
 
     serviceContext = serviceContextFromDefaults({
       callbackManager,
-      llm: languageModel,
-      embedModel,
+      llm: newLanguageModel,
+      embedModel: new OpenAIEmbedding(),
     });
   });
 
@@ -67,7 +76,7 @@ describe("CallbackManager: onLLMStream and onRetrieve", () => {
     });
     const queryEngine = vectorStoreIndex.asQueryEngine();
     const query = "What is the author's name?";
-    const response = await queryEngine.query(query);
+    const response = await queryEngine.query(query, serviceContext);
     expect(response.toString()).toBe("MOCK_TOKEN_1-MOCK_TOKEN_2");
     expect(streamCallbackData).toEqual([
       {

packages/create-llama/templates/types/simple/fastapi/README-template.md

@@ -1,13 +1,15 @@
-This is a [LlamaIndex](https://www.llamaindex.ai/) project using [FastAPI](https://fastapi.tiangolo.com/) bootstrapped with [`create-llama`](https://github.com/run-llama/LlamaIndexTS/tree/main/packages/create-llama).
+This is a [LlamaIndex](https://www.llamaindex.ai/) project using [FastAPI](https://fastapi.tiangolo.com/) bootstrapped with [`create-llama`](https://github.com/run-llama/create-llama-ts).
 
 ## Getting Started
 
-First, setup the environment:
+First, setup the environment by running the following commands:
 
 ```
+```bash
 poetry install
 poetry shell
 ```
+```
 
 By default, we use the OpenAI LLM (though you can customize, see app/api/routers/chat.py). As a result you need to specify an `OPENAI_API_KEY` in an .env file in this directory.
 
@@ -17,29 +19,33 @@ Example `backend/.env` file:
 OPENAI_API_KEY=<openai_api_key>
 ```
 
-Second, generate the embeddings of the documents in the `./data` directory (if this folder exists - otherwise, skip this step):
+Second, generate the embeddings of the documents by following the steps below:
 
 ```
-python app/engine/generate.py
+```bash
+python app/data/generate.py
+```
 ```
 
-Third, run the development server:
+Next, run the development server with the command below:
 
 ```
 python main.py
 ```
 
-Then call the API endpoint `/api/chat` to see the result:
+After running the server, call the API endpoint `/api/chat` to see the result using the following command:
 
 ```
-curl --location 'localhost:8000/api/chat' \
+```bash
+curl --location '127.0.0.1:8000/api/chat' \
 --header 'Content-Type: application/json' \
 --data '{ "messages": [{ "role": "user", "content": "Hello" }] }'
 ```
+```
 
 You can start editing the API by modifying `app/api/routers/chat.py`. The endpoint auto-updates as you save the file.
 
-Open [http://localhost:8000/docs](http://localhost:8000/docs) with your browser to see the Swagger UI of the API.
+Now, open [http://localhost:8000/docs](http://localhost:8000/docs) with your browser to see the Swagger UI of the API.
 
 The API allows CORS for all origins to simplify development. You can change this behavior by setting the `ENVIRONMENT` environment variable to `prod`:
 
@@ -47,10 +53,24 @@ The API allows CORS for all origins to simplify development. You can change this
 ENVIRONMENT=prod uvicorn main:app
 ```
 
-## Learn More
+## Troubleshooting and Error Logs
+
+### Troubleshooting Common GitHub Actions Issues
+
+If the GitHub Actions run fails, please consider the following troubleshooting steps:
 
 To learn more about LlamaIndex, take a look at the following resources:
 
-- [LlamaIndex Documentation](https://docs.llamaindex.ai) - learn about LlamaIndex.
+- [LlamaIndex Documentation](https://docs.llamaindex.ai) - learn more about LlamaIndex.
 
-You can check out [the LlamaIndex GitHub repository](https://github.com/run-llama/llama_index) - your feedback and contributions are welcome!
+### Accessing and Interpreting Error Logs
+
+When the GitHub Actions run fails, you can access error logs to investigate the cause. Follow these steps to access the logs:
+
+1. Navigate to the GitHub Actions page for the repository.
+2. Click on the failed workflow to view the details.
+3. Scroll down to find the error logs and examine the details.
+
+The error logs provide valuable information to help identify the root cause of the failure.
+
+You can check out [the LlamaIndex GitHub repository](https://github.com/run-llama/LlamaIndexTS) - your feedback and contributions are welcome!

packages/create-llama/templates/types/simple/fastapi/app/api/routers/chat.py

@@ -1,7 +1,7 @@
 from typing import List
 
 from fastapi import APIRouter, Depends, HTTPException, status
-from llama_index.chat_engine.types import BaseChatEngine
+from app.engine.index import BaseChatEngine
 from llama_index.llms.base import ChatMessage
 from llama_index.llms.types import MessageRole
 from pydantic import BaseModel

packages/create-llama/templates/types/simple/fastapi/app/context.py

@@ -2,6 +2,8 @@ import os
 
 from llama_index import ServiceContext
 from llama_index.llms import OpenAI
+from llama_index.engine import BaseChatEngine
+from dotenv import load_dotenv
 
 
 def create_base_context():

packages/create-llama/templates/types/simple/fastapi/app/engine/index.py

@@ -1,4 +1,4 @@
-from llama_index.chat_engine import SimpleChatEngine
+from app.engine.index import SimpleChatEngine
 
 from app.context import create_base_context
 

packages/create-llama/templates/types/simple/fastapi/main.py

@@ -1,4 +1,6 @@
 from dotenv import load_dotenv
+from fastapi import FastAPI, Request, status, HTTPException
+from fastapi.middleware.cors import CORSMiddleware
 
 load_dotenv()
 
@@ -11,7 +13,7 @@ from fastapi.middleware.cors import CORSMiddleware
 
 app = FastAPI()
 
-environment = os.getenv("ENVIRONMENT", "dev")  # Default to 'development' if not set
+environment = os.getenv("ENVIRONMENT", default="dev")  # Default to 'development' if not set
 
 
 if environment == "dev":
@@ -25,7 +27,7 @@ if environment == "dev":
         allow_headers=["*"],
     )
 
-app.include_router(chat_router, prefix="/api/chat")
+app.include_router(chat_router, prefix="/api/chat", tags=["chat"])
 
 
 if __name__ == "__main__":

packages/create-llama/templates/types/simple/fastapi/pyproject.toml

@@ -6,7 +6,7 @@ authors = ["Marcus Schiesser <mail@marcusschiesser.de>"]
 readme = "README.md"
 
 [tool.poetry.dependencies]
-python = "^3.11,<3.12"
+python = "^3.11"
 fastapi = "^0.104.1"
 uvicorn = { extras = ["standard"], version = "^0.23.2" }
 llama-index = "^0.9.19"

packages/create-llama/templates/types/streaming/fastapi/README-template.md

@@ -1,4 +1,4 @@
-This is a [LlamaIndex](https://www.llamaindex.ai/) project using [FastAPI](https://fastapi.tiangolo.com/) bootstrapped with [`create-llama`](https://github.com/run-llama/LlamaIndexTS/tree/main/packages/create-llama).
+This is a [LlamaIndex](https://www.llamaindex.ai/) project using [FastAPI](https://fastapi.tiangolo.com/) bootstrapped with [`create-llama`](https://github.com/run-llama/create-llama-ts).
 
 ## Getting Started
 
@@ -26,7 +26,7 @@ python app/engine/generate.py
 Third, run the development server:
 
 ```
-python main.py
+python app/main.py
 ```
 
 Then call the API endpoint `/api/chat` to see the result:
@@ -53,4 +53,4 @@ To learn more about LlamaIndex, take a look at the following resources:
 
 - [LlamaIndex Documentation](https://docs.llamaindex.ai) - learn about LlamaIndex.
 
-You can check out [the LlamaIndex GitHub repository](https://github.com/run-llama/llama_index) - your feedback and contributions are welcome!
+You can check out [the LlamaIndex GitHub repository](https://github.com/run-llama/LlamaIndexTS) - your feedback and contributions are welcome!