feat!: Migrate internal client from googleai_dart to google_generativ…

…e_ai (#407)

docs/modules/model_io/models/chat_models/integrations/googleai.md

@@ -4,11 +4,21 @@ Wrapper around [Google AI for Developers](https://ai.google.dev/) API (aka Gemin
 
 ## Setup
 
-To use `ChatGoogleGenerativeAI` you need to have an API key. You can get one [here](https://makersuite.google.com/app/apikey).
-
-The following models are available at the moment:
-- `gemini-pro`: text -> text model
-- `gemini-pro-vision`: text / image -> text model
+To use `ChatGoogleGenerativeAI` you need to have an API key. You can get one [here](https://aistudio.google.com/app/apikey).
+
+The following models are available:
+- `gemini-1.0-pro` (or `gemini-pro`):
+    * text -> text model
+    * Max input token: 30720
+    * Max output tokens: 2048
+- `gemini-pro-vision`:
+    * text / image -> text model
+    * Max input token: 12288
+    * Max output tokens: 4096
+- `gemini-1.5-pro-latest`: text / image -> text model
+    * text / image -> text model
+    * Max input token: 1048576
+    * Max output tokens: 8192
 
 Mind that this list may not be up-to-date. Refer to the [documentation](https://ai.google.dev/models) for the updated list.
 
@@ -20,17 +30,15 @@ final apiKey = Platform.environment['GOOGLEAI_API_KEY'];
 final chatModel = ChatGoogleGenerativeAI(
   apiKey: apiKey,
   defaultOptions: ChatGoogleGenerativeAIOptions(
+    model: 'gemini-1.5-pro-latest',
     temperature: 0,
   ),
 );
 
-const template = '''
-You are a helpful assistant that translates {input_language} to {output_language}. 
-
-Text to translate: 
-{text}''';
-final humanMessagePrompt = HumanChatMessagePromptTemplate.fromTemplate(template);
-final chatPrompt = ChatPromptTemplate.fromPromptMessages([humanMessagePrompt]);
+final chatPrompt = ChatPromptTemplate.fromTemplates([
+  (ChatMessageType.system, 'You are a helpful assistant that translates {input_language} to {output_language}.'),
+  (ChatMessageType.human, 'Text to translate:\n{text}'),
+]);
 
 final chain = chatPrompt | chatModel | StringOutputParser();
 
@@ -40,7 +48,7 @@ final res = await chain.invoke({
   'text': 'I love programming.',
 });
 print(res);
-// -> 'J'adore la programmation.'final
+// -> 'J'adore programmer.'
 ```
 
 ## Multimodal support
@@ -51,7 +59,7 @@ final apiKey = Platform.environment['GOOGLEAI_API_KEY'];
 final chatModel = ChatGoogleGenerativeAI(
   apiKey: apiKey,
   defaultOptions: ChatGoogleGenerativeAIOptions(
-    model: 'gemini-pro-vision',
+    model: 'gemini-1.5-pro-latest',
     temperature: 0,
   ),
 );
@@ -71,35 +79,35 @@ final res = await chatModel.invoke(
   ]),
 );
 print(res.output.content);
-// -> 'A Red and Green Apple'
+// -> 'That is an apple.'
 ```
 
 ## Streaming
 
 ```dart
 final apiKey = Platform.environment['GOOGLEAI_API_KEY'];
 
-final promptTemplate = ChatPromptTemplate.fromTemplate(
-    'You are a helpful assistant that replies only with numbers '
-    'in order without any spaces or commas '
-    'List the numbers from 1 to {max_num}');
+final promptTemplate = ChatPromptTemplate.fromTemplates(const [
+  (ChatMessageType.system, 'You are a helpful assistant that replies only with numbers in order without any spaces or commas.'),
+  (ChatMessageType.human, 'List the numbers from 1 to {max_num}'),
+]);
 
-final chatModel = ChatGoogleGenerativeAI(apiKey: apiKey);
+final chatModel = ChatGoogleGenerativeAI(
+  apiKey: apiKey,
+  defaultOptions: const ChatGoogleGenerativeAIOptions(
+    model: 'gemini-1.5-pro-latest',
+    temperature: 0,
+  ),
+);
 
 final chain = promptTemplate.pipe(chatModel).pipe(StringOutputParser());
 
 final stream = chain.stream({'max_num': '30'});
 await stream.forEach(print);
-// 1234567891011121
-// 31415161718192021222324252627282
-// 930
-```
-
-## Limitations
+// 1
+// 2345678910111213
+// 1415161718192021
+// 222324252627282930 
 
-As of the time this doc was written (15/12/23), Gemini has some restrictions on the types and structure of prompts it accepts. Specifically:
-
-1. When providing multimodal (image) inputs, you are restricted to at most 1 message of “human” (user) type. You cannot pass multiple messages (though the single human message may have multiple content entries).
-2. System messages are not accepted.
-3. For regular chat conversations, messages must follow the human/ai/human/ai alternating pattern. You may not provide 2 AI or human messages in sequence.
-4. Message may be blocked if they violate the safety checks of the LLM. In this case, the model will return an empty response.
+chatModel.close();
+```

docs/modules/retrieval/text_embedding/integrations/google_ai.md

@@ -4,8 +4,10 @@ The embedding service in the [Gemini API](https://ai.google.dev/docs/embeddings_
 
 ## Available models
 
-- `embedding-001` (default)
-  * Optimized for creating embeddings for text of up to 2048 tokens
+- `text-embedding-004`
+  * Dimensions: 768 (with support for reduced dimensionality)
+- `embedding-001`
+  * Dimensions: 768
 
 The previous list of models may not be exhaustive or up-to-date. Check out the [Google AI documentation](https://ai.google.dev/models/gemini) for the latest list of available models.
 

examples/browser_summarizer/pubspec.lock

@@ -246,7 +246,7 @@ packages:
       path: "../../packages/langchain_openai"
       relative: true
     source: path
-    version: "0.6.0+1"
+    version: "0.6.0+2"
   langchain_tiktoken:
     dependency: transitive
     description:
@@ -309,7 +309,7 @@ packages:
       path: "../../packages/openai_dart"
       relative: true
     source: path
-    version: "0.3.0"
+    version: "0.3.1"
   path:
     dependency: transitive
     description:

examples/docs_examples/bin/modules/model_io/models/chat_models/integrations/googleai.dart

@@ -17,19 +17,18 @@ Future<void> _chatGoogleGenerativeAI() async {
   final chatModel = ChatGoogleGenerativeAI(
     apiKey: apiKey,
     defaultOptions: const ChatGoogleGenerativeAIOptions(
+      model: 'gemini-1.5-pro-latest',
       temperature: 0,
     ),
   );
 
-  const template = '''
-You are a helpful assistant that translates {input_language} to {output_language}. 
-
-Text to translate: 
-{text}''';
-  final humanMessagePrompt =
-      HumanChatMessagePromptTemplate.fromTemplate(template);
-  final chatPrompt =
-      ChatPromptTemplate.fromPromptMessages([humanMessagePrompt]);
+  final chatPrompt = ChatPromptTemplate.fromTemplates(const [
+    (
+      ChatMessageType.system,
+      'You are a helpful assistant that translates {input_language} to {output_language}.'
+    ),
+    (ChatMessageType.human, 'Text to translate:\n{text}'),
+  ]);
 
   final chain = chatPrompt | chatModel | const StringOutputParser();
 
@@ -39,7 +38,7 @@ Text to translate:
     'text': 'I love programming.',
   });
   print(res);
-  // -> 'J'adore la programmation.'
+  // -> 'J'adore programmer.'
 
   chatModel.close();
 }
@@ -50,7 +49,7 @@ Future<void> _chatGoogleGenerativeAIMultiModal() async {
   final chatModel = ChatGoogleGenerativeAI(
     apiKey: apiKey,
     defaultOptions: const ChatGoogleGenerativeAIOptions(
-      model: 'gemini-pro-vision',
+      model: 'gemini-1.5-pro-latest',
       temperature: 0,
     ),
   );
@@ -70,28 +69,39 @@ Future<void> _chatGoogleGenerativeAIMultiModal() async {
     ]),
   );
   print(res.output.content);
-  // -> 'A Red and Green Apple'
+  // -> 'That is an apple.'
 
   chatModel.close();
 }
 
 Future<void> _chatOpenAIStreaming() async {
   final apiKey = Platform.environment['GOOGLEAI_API_KEY'];
 
-  final promptTemplate = ChatPromptTemplate.fromTemplate(
+  final promptTemplate = ChatPromptTemplate.fromTemplates(const [
+    (
+      ChatMessageType.system,
       'You are a helpful assistant that replies only with numbers '
-      'in order without any spaces or commas '
-      'List the numbers from 1 to {max_num}');
+          'in order without any spaces or commas.',
+    ),
+    (ChatMessageType.human, 'List the numbers from 1 to {max_num}'),
+  ]);
 
-  final chatModel = ChatGoogleGenerativeAI(apiKey: apiKey);
+  final chatModel = ChatGoogleGenerativeAI(
+    apiKey: apiKey,
+    defaultOptions: const ChatGoogleGenerativeAIOptions(
+      model: 'gemini-1.5-pro-latest',
+      temperature: 0,
+    ),
+  );
 
   final chain = promptTemplate.pipe(chatModel).pipe(const StringOutputParser());
 
   final stream = chain.stream({'max_num': '30'});
   await stream.forEach(print);
-  // 1234567891011121
-  // 31415161718192021222324252627282
-  // 930
+  // 1
+  // 2345678910111213
+  // 1415161718192021
+  // 222324252627282930
 
   chatModel.close();
 }

examples/docs_examples/pubspec.lock

@@ -128,6 +128,14 @@ packages:
       url: "https://pub.dev"
     source: hosted
     version: "0.8.12"
+  google_generative_ai:
+    dependency: transitive
+    description:
+      name: google_generative_ai
+      sha256: "3a51ed314e596ddee9654ab9f1372fff0b8f3917be937655bcadebd5fa063df0"
+      url: "https://pub.dev"
+    source: hosted
+    version: "0.3.3"
   google_identity_services_web:
     dependency: transitive
     description:
@@ -136,13 +144,6 @@ packages:
       url: "https://pub.dev"
     source: hosted
     version: "0.3.1+1"
-  googleai_dart:
-    dependency: "direct overridden"
-    description:
-      path: "../../packages/googleai_dart"
-      relative: true
-    source: path
-    version: "0.0.4"
   googleapis:
     dependency: transitive
     description:
@@ -270,7 +271,7 @@ packages:
       path: "../../packages/langchain_openai"
       relative: true
     source: path
-    version: "0.6.0+1"
+    version: "0.6.0+2"
   langchain_tiktoken:
     dependency: transitive
     description:
@@ -323,7 +324,7 @@ packages:
       path: "../../packages/openai_dart"
       relative: true
     source: path
-    version: "0.3.0"
+    version: "0.3.1"
   path:
     dependency: transitive
     description:

examples/docs_examples/pubspec_overrides.yaml

@@ -1,9 +1,7 @@
-# melos_managed_dependency_overrides: chromadb,googleai_dart,langchain,langchain_chroma,langchain_google,langchain_mistralai,langchain_ollama,langchain_openai,mistralai_dart,ollama_dart,openai_dart,vertex_ai,langchain_core,langchain_community
+# melos_managed_dependency_overrides: chromadb,langchain,langchain_chroma,langchain_google,langchain_mistralai,langchain_ollama,langchain_openai,mistralai_dart,ollama_dart,openai_dart,vertex_ai,langchain_core,langchain_community
 dependency_overrides:
   chromadb:
     path: ../../packages/chromadb
-  googleai_dart:
-    path: ../../packages/googleai_dart
   langchain:
     path: ../../packages/langchain
   langchain_chroma:

examples/hello_world_backend/pubspec.lock

@@ -133,7 +133,7 @@ packages:
       path: "../../packages/langchain_openai"
       relative: true
     source: path
-    version: "0.6.0+1"
+    version: "0.6.0+2"
   langchain_tiktoken:
     dependency: transitive
     description:
@@ -156,7 +156,7 @@ packages:
       path: "../../packages/openai_dart"
       relative: true
     source: path
-    version: "0.3.0"
+    version: "0.3.1"
   path:
     dependency: transitive
     description:

examples/hello_world_cli/pubspec.lock

@@ -125,7 +125,7 @@ packages:
       path: "../../packages/langchain_openai"
       relative: true
     source: path
-    version: "0.6.0+1"
+    version: "0.6.0+2"
   langchain_tiktoken:
     dependency: transitive
     description:
@@ -148,7 +148,7 @@ packages:
       path: "../../packages/openai_dart"
       relative: true
     source: path
-    version: "0.3.0"
+    version: "0.3.1"
   path:
     dependency: transitive
     description:

examples/hello_world_flutter/pubspec.lock

@@ -154,7 +154,7 @@ packages:
       path: "../../packages/langchain_openai"
       relative: true
     source: path
-    version: "0.6.0+1"
+    version: "0.6.0+2"
   langchain_tiktoken:
     dependency: transitive
     description:
@@ -193,7 +193,7 @@ packages:
       path: "../../packages/openai_dart"
       relative: true
     source: path
-    version: "0.3.0"
+    version: "0.3.1"
   path:
     dependency: transitive
     description:

melos.yaml

@@ -35,6 +35,7 @@ command:
       flutter_markdown: ^0.6.22
       freezed_annotation: ^2.4.1
       gcloud: ^0.8.12
+      google_generative_ai: 0.3.3
       googleapis: ^12.0.0
       googleapis_auth: ^1.5.1
       http: ^1.1.0

packages/googleai_dart/README.md

@@ -5,7 +5,9 @@
 [![](https://dcbadge.vercel.app/api/server/x4qbhqecVR?style=flat)](https://discord.gg/x4qbhqecVR)
 [![MIT](https://img.shields.io/badge/license-MIT-purple.svg)](https://github.com/davidmigloz/langchain_dart/blob/main/LICENSE)
 
-Unofficial Dart client for [Google AI](https://ai.google.dev) for Developers (Gemini API).
+Unofficial Dart client for [Google AI](https://ai.google.dev) for Developers (Gemini API v1).
+
+> Note: The official [`google_generative_ai`](https://pub.dev/packages/google_generative_ai) now has feature parity with this package (except for the [Model info](https://github.com/google-gemini/generative-ai-dart/issues/93) endpoints). We plan to deprecate this package in the near future.
 
 ## Features
 
@@ -53,7 +55,7 @@ Refer to the [documentation](https://ai.google.dev/docs) for more information ab
 
 ### Authentication
 
-The Google AI API uses API keys for authentication. Visit [Google AI Studio dashboard](https://makersuite.google.com/app/apikey) page to retrieve the API key you'll use in your requests.
+The Google AI API uses API keys for authentication. Visit [Google AI Studio dashboard](https://aistudio.google.com/app/apikey) page to retrieve the API key you'll use in your requests.
 
 > **Remember that your API key is a secret!**  
 > Do not share it with others or expose it in any client-side code (browsers, apps). Production requests must be routed through your own backend server where your API key can be securely loaded from an environment variable or key management service.

packages/googleai_dart/lib/src/client.dart

@@ -16,7 +16,7 @@ class GoogleAIClient extends g.GoogleAIClient {
   ///
   /// Main configuration options:
   /// - `apiKey`: your Google AI API key. You can find your API key in the
-  ///   [Google AI Studio dashboard](https://makersuite.google.com/app/apikey).
+  ///   [Google AI Studio dashboard](https://aistudio.google.com/app/apikey).
   ///
   /// Advance configuration options:
   /// - `baseUrl`: the base URL to use.You can override this to use a different