feat(api): Define OpenAI-compatible models

groq · May 22, 2024 · bc1bb7b · bc1bb7b
1 parent d48e365
commit bc1bb7b
Show file tree

Hide file tree

Showing 48 changed files with 929 additions and 827 deletions.
diff --git a/.stats.yml b/.stats.yml
@@ -1,2 +1,2 @@
 configured_endpoints: 7
-openapi_spec_url: https://storage.googleapis.com/stainless-sdk-openapi-specs/groqcloud%2Fgroqcloud-45aff5d28bf67a962a928a38a6cf90ae3034bc824d2e65036d8784643e4995db.yml
+openapi_spec_url: https://storage.googleapis.com/stainless-sdk-openapi-specs/groqcloud%2Fgroqcloud-36fcf453a77cbc8279361577a1e785a3a86ef7bcbde2195270a83e93cbc4b8b3.yml
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,42 +1,5 @@
 # Changelog
 
-## 0.7.0 (2024-05-22)
-
-Full Changelog: [v0.6.0...v0.7.0](https://github.com/groq/groq-python/compare/v0.6.0...v0.7.0)
-
-### Features
-
-* **api:** Improve types ([e46d3ed](https://github.com/groq/groq-python/commit/e46d3ed8591ef543a6f50c4443825fd547b74ab4))
-
-
-### Chores
-
-* fix array indices in README examples ([bacc106](https://github.com/groq/groq-python/commit/bacc10611d74ae9369093bbf44a451182b82e253))
-* patch streaming ([9287ee7](https://github.com/groq/groq-python/commit/9287ee77e505187bb504729fad95edde41663ad0))
-
-## 0.6.0 (2024-05-16)
-
-Full Changelog: [v0.5.0...v0.6.0](https://github.com/groq/groq-python/compare/v0.5.0...v0.6.0)
-
-### Features
-
-* **api:** OpenAPI spec update via Stainless API ([#39](https://github.com/groq/groq-python/issues/39)) ([53a8b27](https://github.com/groq/groq-python/commit/53a8b27365c085227e449812779c5bafadada9fb))
-* **api:** OpenAPI spec update via Stainless API ([#45](https://github.com/groq/groq-python/issues/45)) ([18bce01](https://github.com/groq/groq-python/commit/18bce01c835fbc0485ef78f12b28caead37d15c4))
-* **api:** update via SDK Studio ([#35](https://github.com/groq/groq-python/issues/35)) ([5ca3b78](https://github.com/groq/groq-python/commit/5ca3b785626145a87a27d64a4a4c3eae1b1fa8de))
-* **api:** update via SDK Studio ([#38](https://github.com/groq/groq-python/issues/38)) ([dc73eeb](https://github.com/groq/groq-python/commit/dc73eeb92a441123596b3f3420a757074ee62345))
-* **api:** update via SDK Studio ([#40](https://github.com/groq/groq-python/issues/40)) ([7227816](https://github.com/groq/groq-python/commit/7227816b92397d913f8aeceda64127e2db99f720))
-* **api:** update via SDK Studio ([#41](https://github.com/groq/groq-python/issues/41)) ([3fdb317](https://github.com/groq/groq-python/commit/3fdb31786699d3298d5064faac77e26dbc7852d8))
-* **api:** update via SDK Studio ([#42](https://github.com/groq/groq-python/issues/42)) ([d270e51](https://github.com/groq/groq-python/commit/d270e51d63edc138b0f0afa0b53c2f84a69b1f73))
-* **api:** update via SDK Studio ([#43](https://github.com/groq/groq-python/issues/43)) ([0e04381](https://github.com/groq/groq-python/commit/0e043816e5cbcdf5def7583d20312c6f5e29808f))
-* **api:** update via SDK Studio ([#44](https://github.com/groq/groq-python/issues/44)) ([3807b4c](https://github.com/groq/groq-python/commit/3807b4c67981aec711d3e60e57ec0e5e9493f8a0))
-* **api:** update via SDK Studio ([#46](https://github.com/groq/groq-python/issues/46)) ([ed1ea2a](https://github.com/groq/groq-python/commit/ed1ea2a99f95d4d353c3dd9baf7fc1c5d37cc20a))
-
-
-### Chores
-
-* **examples:** embedding example ([961159c](https://github.com/groq/groq-python/commit/961159c9c83a21cd9631af4802e4b538ebbbb186))
-* patch streaming ([c8eaad3](https://github.com/groq/groq-python/commit/c8eaad35abe33b692e029495157948db3a09af3c))
-
 ## 0.5.0 (2024-04-11)
 
 Full Changelog: [v0.4.2...v0.5.0](https://github.com/groq/groq-python/compare/v0.4.2...v0.5.0)

diff --git a/README.md b/README.md
@@ -24,9 +24,13 @@ pip install groq
 The full API of this library can be found in [api.md](api.md).
 
 ```python
+import os
 from groq import Groq
 
-client = Groq()
+client = Groq(
+ # This is the default and can be omitted
+ api_key=os.environ.get("GROQ_API_KEY"),
+)
 
 chat_completion = client.chat.completions.create(
  messages=[
@@ -35,9 +39,9 @@ chat_completion = client.chat.completions.create(
  "content": "Explain the importance of low latency LLMs",
  }
  ],
- model="mixtral-8x7b-32768",
+ model="llama3-8b-8192",
 )
-print(chat_completion.choices[0].message.content)
+print(chat_completion.choices_0.message.content)
 ```
 
 While you can provide an `api_key` keyword argument,
@@ -50,10 +54,14 @@ so that your API Key is not stored in source control.
 Simply import `AsyncGroq` instead of `Groq` and use `await` with each API call:
 
 ```python
+import os
 import asyncio
 from groq import AsyncGroq
 
-client = AsyncGroq()
+client = AsyncGroq(
+ # This is the default and can be omitted
+ api_key=os.environ.get("GROQ_API_KEY"),
+)
 
 
 async def main() -> None:
@@ -64,9 +72,9 @@ async def main() -> None:
  "content": "Explain the importance of low latency LLMs",
  }
  ],
- model="mixtral-8x7b-32768",
+ model="llama3-8b-8192",
  )
- print(chat_completion.choices[0].message.content)
+ print(chat_completion.choices_0.message.content)
 
 
 asyncio.run(main())
@@ -110,7 +118,7 @@ try:
  "content": "Explain the importance of low latency LLMs",
  },
  ],
- model="mixtral-8x7b-32768",
+ model="llama3-8b-8192",
  )
 except groq.APIConnectionError as e:
  print("The server could not be reached")
@@ -165,7 +173,7 @@ client.with_options(max_retries=5).chat.completions.create(
  "content": "Explain the importance of low latency LLMs",
  },
  ],
- model="mixtral-8x7b-32768",
+ model="llama3-8b-8192",
 )
 ```
 
@@ -200,7 +208,7 @@ client.with_options(timeout=5.0).chat.completions.create(
  "content": "Explain the importance of low latency LLMs",
  },
  ],
- model="mixtral-8x7b-32768",
+ model="llama3-8b-8192",
 )
 ```
 
@@ -248,7 +256,7 @@ response = client.chat.completions.with_raw_response.create(
  "role": "user",
  "content": "Explain the importance of low latency LLMs",
  }],
- model="mixtral-8x7b-32768",
+ model="llama3-8b-8192",
 )
 print(response.headers.get('X-My-Header'))
 
@@ -278,7 +286,7 @@ with client.chat.completions.with_streaming_response.create(
  "content": "Explain the importance of low latency LLMs",
  },
  ],
- model="mixtral-8x7b-32768",
+ model="llama3-8b-8192",
 ) as response:
  print(response.headers.get("X-My-Header"))
 

diff --git a/api.md b/api.md
@@ -1,14 +1,16 @@
-# Embeddings
-
-Types:
+# Shared Types
 
 ```python
-from groq.types import CreateEmbeddingResponse, Embedding
+from groq.types import ErrorObject, FunctionDefinition, FunctionParameters
 ```
 
-Methods:
+# Completions
 
-- <code title="post /openai/v1/embeddings">client.embeddings.<a href="./src/groq/resources/embeddings.py">create</a>(\*\*<a href="src/groq/types/embedding_create_params.py">params</a>) -> <a href="./src/groq/types/create_embedding_response.py">CreateEmbeddingResponse</a></code>
+Types:
+
+```python
+from groq.types import CompletionUsage
+```
 
 # Chat
 
@@ -17,13 +19,45 @@ Methods:
 Types:
 
 ```python
-from groq.types.chat import ChatCompletion
+from groq.types.chat import (
+ ChatCompletion,
+ ChatCompletionAssistantMessageParam,
+ ChatCompletionChunk,
+ ChatCompletionContentPart,
+ ChatCompletionContentPartImage,
+ ChatCompletionContentPartText,
+ ChatCompletionFunctionCallOption,
+ ChatCompletionFunctionMessageParam,
+ ChatCompletionMessage,
+ ChatCompletionMessageParam,
+ ChatCompletionMessageToolCall,
+ ChatCompletionNamedToolChoice,
+ ChatCompletionRole,
+ ChatCompletionSystemMessageParam,
+ ChatCompletionTokenLogprob,
+ ChatCompletionTool,
+ ChatCompletionToolChoiceOption,
+ ChatCompletionToolMessageParam,
+ ChatCompletionUserMessageParam,
+)
 ```
 
 Methods:
 
 - <code title="post /openai/v1/chat/completions">client.chat.completions.<a href="./src/groq/resources/chat/completions.py">create</a>(\*\*<a href="src/groq/types/chat/completion_create_params.py">params</a>) -> <a href="./src/groq/types/chat/chat_completion.py">ChatCompletion</a></code>
 
+# Embeddings
+
+Types:
+
+```python
+from groq.types import CreateEmbeddingResponse, Embedding
+```
+
+Methods:
+
+- <code title="post /openai/v1/embeddings">client.embeddings.<a href="./src/groq/resources/embeddings.py">create</a>(\*\*<a href="src/groq/types/embedding_create_params.py">params</a>) -> <a href="./src/groq/types/create_embedding_response.py">CreateEmbeddingResponse</a></code>
+
 # Audio
 
 ## Transcriptions
@@ -55,11 +89,11 @@ Methods:
 Types:
 
 ```python
-from groq.types import Model, ModelList
+from groq.types import Model, ModelDeleted, ModelListResponse
 ```
 
 Methods:
 
 - <code title="get /openai/v1/models/{model}">client.models.<a href="./src/groq/resources/models.py">retrieve</a>(model) -> <a href="./src/groq/types/model.py">Model</a></code>
-- <code title="get /openai/v1/models">client.models.<a href="./src/groq/resources/models.py">list</a>() -> <a href="./src/groq/types/model_list.py">ModelList</a></code>
-- <code title="delete /openai/v1/models/{model}">client.models.<a href="./src/groq/resources/models.py">delete</a>(model) -> None</code>
+- <code title="get /openai/v1/models">client.models.<a href="./src/groq/resources/models.py">list</a>() -> <a href="./src/groq/types/model_list_response.py">ModelListResponse</a></code>
+- <code title="delete /openai/v1/models/{model}">client.models.<a href="./src/groq/resources/models.py">delete</a>(model) -> <a href="./src/groq/types/model_deleted.py">ModelDeleted</a></code>
diff --git a/src/groq/_client.py b/src/groq/_client.py
@@ -46,8 +46,8 @@
 
 
 class Groq(SyncAPIClient):
- embeddings: resources.Embeddings
  chat: resources.Chat
+ embeddings: resources.Embeddings
  audio: resources.Audio
  models: resources.Models
  with_raw_response: GroqWithRawResponse
@@ -107,8 +107,8 @@ def __init__(
  _strict_response_validation=_strict_response_validation,
  )
 
- self.embeddings = resources.Embeddings(self)
  self.chat = resources.Chat(self)
+ self.embeddings = resources.Embeddings(self)
  self.audio = resources.Audio(self)
  self.models = resources.Models(self)
  self.with_raw_response = GroqWithRawResponse(self)
@@ -220,8 +220,8 @@ def _make_status_error(
 
 
 class AsyncGroq(AsyncAPIClient):
- embeddings: resources.AsyncEmbeddings
  chat: resources.AsyncChat
+ embeddings: resources.AsyncEmbeddings
  audio: resources.AsyncAudio
  models: resources.AsyncModels
  with_raw_response: AsyncGroqWithRawResponse
@@ -281,8 +281,8 @@ def __init__(
  _strict_response_validation=_strict_response_validation,
  )
 
- self.embeddings = resources.AsyncEmbeddings(self)
  self.chat = resources.AsyncChat(self)
+ self.embeddings = resources.AsyncEmbeddings(self)
  self.audio = resources.AsyncAudio(self)
  self.models = resources.AsyncModels(self)
  self.with_raw_response = AsyncGroqWithRawResponse(self)
@@ -395,32 +395,32 @@ def _make_status_error(
 
 class GroqWithRawResponse:
  def __init__(self, client: Groq) -> None:
- self.embeddings = resources.EmbeddingsWithRawResponse(client.embeddings)
  self.chat = resources.ChatWithRawResponse(client.chat)
+ self.embeddings = resources.EmbeddingsWithRawResponse(client.embeddings)
  self.audio = resources.AudioWithRawResponse(client.audio)
  self.models = resources.ModelsWithRawResponse(client.models)
 
 
 class AsyncGroqWithRawResponse:
  def __init__(self, client: AsyncGroq) -> None:
- self.embeddings = resources.AsyncEmbeddingsWithRawResponse(client.embeddings)
  self.chat = resources.AsyncChatWithRawResponse(client.chat)
+ self.embeddings = resources.AsyncEmbeddingsWithRawResponse(client.embeddings)
  self.audio = resources.AsyncAudioWithRawResponse(client.audio)
  self.models = resources.AsyncModelsWithRawResponse(client.models)
 
 
 class GroqWithStreamedResponse:
  def __init__(self, client: Groq) -> None:
- self.embeddings = resources.EmbeddingsWithStreamingResponse(client.embeddings)
  self.chat = resources.ChatWithStreamingResponse(client.chat)
+ self.embeddings = resources.EmbeddingsWithStreamingResponse(client.embeddings)
  self.audio = resources.AudioWithStreamingResponse(client.audio)
  self.models = resources.ModelsWithStreamingResponse(client.models)
 
 
 class AsyncGroqWithStreamedResponse:
  def __init__(self, client: AsyncGroq) -> None:
- self.embeddings = resources.AsyncEmbeddingsWithStreamingResponse(client.embeddings)
  self.chat = resources.AsyncChatWithStreamingResponse(client.chat)
+ self.embeddings = resources.AsyncEmbeddingsWithStreamingResponse(client.embeddings)
  self.audio = resources.AsyncAudioWithStreamingResponse(client.audio)
  self.models = resources.AsyncModelsWithStreamingResponse(client.models)
 

diff --git a/src/groq/_streaming.py b/src/groq/_streaming.py
@@ -55,8 +55,6 @@ def __stream__(self) -> Iterator[_T]:
  iterator = self._iter_events()
 
  for sse in iterator:
- if sse.data.startswith("[DONE]"):
- break
  yield process_data(data=sse.json(), cast_to=cast_to, response=response)
 
  # Ensure the entire stream is consumed
@@ -121,8 +119,6 @@ async def __stream__(self) -> AsyncIterator[_T]:
  iterator = self._iter_events()
 
  async for sse in iterator:
- if sse.data.startswith("[DONE]"):
- break
  yield process_data(data=sse.json(), cast_to=cast_to, response=response)
 
  # Ensure the entire stream is consumed

diff --git a/src/groq/resources/__init__.py b/src/groq/resources/__init__.py
@@ -34,18 +34,18 @@
 )
 
 __all__ = [
- "Embeddings",
- "AsyncEmbeddings",
- "EmbeddingsWithRawResponse",
- "AsyncEmbeddingsWithRawResponse",
- "EmbeddingsWithStreamingResponse",
- "AsyncEmbeddingsWithStreamingResponse",
  "Chat",
  "AsyncChat",
  "ChatWithRawResponse",
  "AsyncChatWithRawResponse",
  "ChatWithStreamingResponse",
  "AsyncChatWithStreamingResponse",
+ "Embeddings",
+ "AsyncEmbeddings",
+ "EmbeddingsWithRawResponse",
+ "AsyncEmbeddingsWithRawResponse",
+ "EmbeddingsWithStreamingResponse",
+ "AsyncEmbeddingsWithStreamingResponse",
  "Audio",
  "AsyncAudio",
  "AudioWithRawResponse",