Closes#37007
---
`ChatMistralAI` was POSTing `HumanMessage` content lists verbatim, so
canonical `ImageContentBlock` dicts (`{"type": "image", "url"/"base64":
...}`) reached the Mistral API unchanged and were rejected — the API
expects OpenAI-shape `{"type": "image_url", "image_url": {"url":
"..."}}`. Multimodal inputs failed for both URL and base64 images.
## Changes
- Introduce `_format_message_content` in
`langchain_mistralai.chat_models`, which delegates to
`is_data_content_block` and
`convert_to_openai_data_block(api="chat/completions")` from
`langchain-core`. Reuses the same translator `langchain-openai` and
`langchain-fireworks` (#37090) use, so v0 `source_type` blocks, v1
`url`/`base64` blocks, and `file_id` references are all handled by one
canonical path.
- Route `HumanMessage` content through `_format_message_content` in
`_convert_message_to_mistral_chat_message`. Strings, already-translated
`image_url` blocks, and Mistral-specific blocks (`document_url`,
`input_audio`) pass through unchanged; the API surfaces an error for
anything it doesn't recognize.
---------
Co-authored-by: Akash Choudhary <achoudhary@lenovo.com>
Co-authored-by: Mason Daugherty <github@mdrxy.com>
The async embed function does not properly handle HTTP errors.
For instance with large batches, Mistral AI returns `Too many inputs in
request, split into more batches.` in a 400 error.
This leads to a KeyError in `response.json()["data"]` l.288
This PR fixes the issue by:
- calling `response.raise_for_status()` before returning
- adding a retry similarly to what is done in the synchronous
counterpart `embed_documents`
I also added an integration test, but willing to move it to unit tests
if more relevant.
**partners: Enable max_retries in ChatMistralAI**
**Description**
- This pull request reactivates the retry logic in the
completion_with_retry method of the ChatMistralAI class, restoring the
intended functionality of the previously ineffective max_retries
parameter. New unit test that mocks failed/successful retry calls and an
integration test to confirm end-to-end functionality.
**Issue**
- Closes#30362
**Dependencies**
- No additional dependencies required
Co-authored-by: andrasfe <andrasf94@gmail.com>
We are implementing a token-counting callback handler in
`langchain-core` that is intended to work with all chat models
supporting usage metadata. The callback will aggregate usage metadata by
model. This requires responses to include the model name in its
metadata.
To support this, if a model `returns_usage_metadata`, we check that it
includes a string model name in its `response_metadata` in the
`"model_name"` key.
More context: https://github.com/langchain-ai/langchain/pull/30487
- Test if models support forcing tool calls via `tool_choice`. If they
do, they should support
- `"any"` to specify any tool
- the tool name as a string to force calling a particular tool
- Add `tool_choice` to signature of `BaseChatModel.bind_tools` in core
- Deprecate `tool_choice_value` in standard tests in favor of a boolean
`has_tool_choice`
Will follow up with PRs in external repos (tested in AWS and Google
already).
Took a "census" of models supported by init_chat_model-- of those that
return model names in response metadata, these were the only two that
had it keyed under `"model"` instead of `"model_name"`.
**Description:**
Added ability to set `prefix` attribute to prevent error :
```
httpx.HTTPStatusError: Error response 400 while fetching https://api.mistral.ai/v1/chat/completions: {"object":"error","message":"Expected last role User or Tool (or Assistant with prefix True) for serving but got assistant","type":"invalid_request_error","param":null,"code":null}
```
Co-authored-by: Sylvain DEPARTE <sylvain.departe@wizbii.com>
Thank you for contributing to LangChain!
**Description:**
Similar to other packages (`langchain_openai`, `langchain_anthropic`) it
would be beneficial if that `ChatMistralAI` model could fetch the API
base URL from the environment.
This PR allows this via the following order:
- provided value
- then whatever `MISTRAL_API_URL` is set to
- then whatever `MISTRAL_BASE_URL` is set to
- if `None`, then default is ` "https://api.mistral.com/v1"`
- [x] **Add tests and docs**:
Added unit tests, docs I feel are unnecessary, as this is just aligning
with other packages that do the same?
- [x] **Lint and test**:
Additional guidelines:
- Make sure optional dependencies are imported within a function.
- Please do not add dependencies to pyproject.toml files (even optional
ones) unless they are required for unit tests.
- Most PRs should not touch more than one package.
- Changes should be backwards compatible.
- If you are adding something to community, do not re-import it in
langchain.
If no one reviews your PR within a few days, please @-mention one of
baskaryan, efriis, eyurtsev, ccurme, vbarda, hwchase17.
---------
Co-authored-by: Chester Curme <chester.curme@gmail.com>
Here we allow standard tests to specify a value for `tool_choice` via a
`tool_choice_value` property, which defaults to None.
Chat models [available in
Together](https://docs.together.ai/docs/chat-models) have issues passing
standard tool calling tests:
- llama 3.1 models currently [appear to rely on user-side
parsing](https://docs.together.ai/docs/llama-3-function-calling) in
Together;
- Mixtral-8x7B and Mistral-7B (currently tested) consistently do not
call tools in some tests.
Specifying tool_choice also lets us remove an existing `xfail` and use a
smaller model in Groq tests.
Mistral appears to have added validation for the format of its tool call
IDs:
`{"object":"error","message":"Tool call id was abc123 but must be a-z,
A-Z, 0-9, with a length of
9.","type":"invalid_request_error","param":null,"code":null}`
This breaks compatibility of messages from other providers. Here we add
a function that converts any string to a Mistral-valid tool call ID, and
apply it to incoming messages.
- Refactor standard test classes to make them easier to configure
- Update openai to support stop_sequences init param
- Update groq to support stop_sequences init param
- Update fireworks to support max_retries init param
- Update ChatModel.bind_tools to type tool_choice
- Update groq to handle tool_choice="any". **this may be controversial**
---------
Co-authored-by: Chester Curme <chester.curme@gmail.com>
```python
class UsageMetadata(TypedDict):
"""Usage metadata for a message, such as token counts.
Attributes:
input_tokens: (int) count of input (or prompt) tokens
output_tokens: (int) count of output (or completion) tokens
total_tokens: (int) total token count
"""
input_tokens: int
output_tokens: int
total_tokens: int
```
```python
class AIMessage(BaseMessage):
...
usage_metadata: Optional[UsageMetadata] = None
"""If provided, token usage information associated with the message."""
...
```
This PR fixes#21196.
The error was occurring when calling chat completion API with a chat
history. Indeed, the Mistral API does not accept both `content` and
`tool_calls` in the same body.
This PR removes one of theses variables depending on the necessity.
---------
Co-authored-by: Maxime Perrin <mperrin@doing.fr>
Co-authored-by: Chester Curme <chester.curme@gmail.com>
Mistral gives us one ID per response, no individual IDs for tool calls.
```python
from langchain.agents import AgentExecutor, create_tool_calling_agent, tool
from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
from langchain_mistralai import ChatMistralAI
prompt = ChatPromptTemplate.from_messages(
[
("system", "You are a helpful assistant"),
("human", "{input}"),
MessagesPlaceholder("agent_scratchpad"),
]
)
model = ChatMistralAI(model="mistral-large-latest", temperature=0)
@tool
def magic_function(input: int) -> int:
"""Applies a magic function to an input."""
return input + 2
tools = [magic_function]
agent = create_tool_calling_agent(model, tools, prompt)
agent_executor = AgentExecutor(agent=agent, tools=tools, verbose=True)
agent_executor.invoke({"input": "what is the value of magic_function(3)?"})
```
---------
Co-authored-by: Eugene Yurtsev <eyurtsev@gmail.com>
core[minor], langchain[patch], openai[minor], anthropic[minor], fireworks[minor], groq[minor], mistralai[minor]
```python
class ToolCall(TypedDict):
name: str
args: Dict[str, Any]
id: Optional[str]
class InvalidToolCall(TypedDict):
name: Optional[str]
args: Optional[str]
id: Optional[str]
error: Optional[str]
class ToolCallChunk(TypedDict):
name: Optional[str]
args: Optional[str]
id: Optional[str]
index: Optional[int]
class AIMessage(BaseMessage):
...
tool_calls: List[ToolCall] = []
invalid_tool_calls: List[InvalidToolCall] = []
...
class AIMessageChunk(AIMessage, BaseMessageChunk):
...
tool_call_chunks: Optional[List[ToolCallChunk]] = None
...
```
Important considerations:
- Parsing logic occurs within different providers;
- ~Changing output type is a breaking change for anyone doing explicit
type checking;~
- ~Langsmith rendering will need to be updated:
https://github.com/langchain-ai/langchainplus/pull/3561~
- ~Langserve will need to be updated~
- Adding chunks:
- ~AIMessage + ToolCallsMessage = ToolCallsMessage if either has
non-null .tool_calls.~
- Tool call chunks are appended, merging when having equal values of
`index`.
- additional_kwargs accumulate the normal way.
- During streaming:
- ~Messages can change types (e.g., from AIMessageChunk to
AIToolCallsMessageChunk)~
- Output parsers parse additional_kwargs (during .invoke they read off
tool calls).
Packages outside of `partners/`:
- https://github.com/langchain-ai/langchain-cohere/pull/7
- https://github.com/langchain-ai/langchain-google/pull/123/files
---------
Co-authored-by: Chester Curme <chester.curme@gmail.com>
