Commit Graph

57 Commits

Author SHA1 Message Date
ccurme
936aedd10c
mistral[patch]: add usage_metadata to (a)invoke and (a)stream (#22781) 2024-06-11 15:34:50 -04:00
Bagatur
8e86080def
mistralai[patch]: Release 0.1.8 (#22494) 2024-06-04 11:33:06 -07:00
Mohammad Mohtashim
577ed68b59
mistralai[patch]: Added Json Mode for ChatMistralAI (#22213)
- **Description:** Powered
[ChatMistralAI.with_structured_output](fbfed65fb1/libs/partners/mistralai/langchain_mistralai/chat_models.py (L609))
via json mode
 

-  **Issue:** #22081

---------

Co-authored-by: Bagatur <baskaryan@gmail.com>
2024-05-27 21:16:52 +00:00
ccurme
fbfed65fb1
core, partners: add token usage attribute to AIMessage (#21944)
```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."""
    ...
```
2024-05-23 14:21:58 -04:00
ccurme
9c76739425
mistral: implement ls_params (#21867) 2024-05-20 11:49:48 -07:00
ccurme
181dfef118
core, standard tests, partner packages: add test for model params (#21677)
1. Adds `.get_ls_params` to BaseChatModel which returns
```python
class LangSmithParams(TypedDict, total=False):
    ls_provider: str
    ls_model_name: str
    ls_model_type: Literal["chat"]
    ls_temperature: Optional[float]
    ls_max_tokens: Optional[int]
    ls_stop: Optional[List[str]]
```
by default it will only return
```python
{ls_model_type="chat", ls_stop=stop}
```

2. Add these params to inheritable metadata in
`CallbackManager.configure`

3. Implement `.get_ls_params` and populate all params for Anthropic +
all subclasses of BaseChatOpenAI

Sample trace:
https://smith.langchain.com/public/d2962673-4c83-47c7-b51e-61d07aaffb1b/r

**OpenAI**:
<img width="984" alt="Screenshot 2024-05-17 at 10 03 35 AM"
src="https://github.com/langchain-ai/langchain/assets/26529506/2ef41f74-a9df-4e0e-905d-da74fa82a910">

**Anthropic**:
<img width="978" alt="Screenshot 2024-05-17 at 10 06 07 AM"
src="https://github.com/langchain-ai/langchain/assets/26529506/39701c9f-7da5-4f1a-ab14-84e9169d63e7">

**Mistral** (and all others for which params are not yet populated):
<img width="977" alt="Screenshot 2024-05-17 at 10 08 43 AM"
src="https://github.com/langchain-ai/langchain/assets/26529506/37d7d894-fec2-4300-986f-49a5f0191b03">
2024-05-17 13:51:26 -04:00
Erick Friis
aca98fd150
multiple: releases with relaxed core dep (#21724) 2024-05-15 19:29:35 +00:00
Erick Friis
c77d2f2b06
multiple: core 0.2 nonbreaking dep, check_diff community->langchain dep (#21646)
0.2 is not a breaking release for core (but it is for langchain and
community)

To keep the core+langchain+community packages in sync at 0.2, we will
relax deps throughout the ecosystem to tolerate `langchain-core` 0.2
2024-05-13 19:50:36 -07:00
Erick Friis
36c2ca3c8b
mistralai: relax tokenizers dep (#21277) 2024-05-03 14:16:22 -07:00
ccurme
eb0a2fd53a
mistral: release 0.1.6 (#21214) 2024-05-02 13:59:19 -04:00
Maxime Perrin
1ebb5a70ad
partners(mistralai): Removing unused variable in completion request (using tool_calls or content) (#21201)
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>
2024-05-02 13:20:14 -04:00
Erick Friis
854ae3e1de
mistralai: release 0.1.5, allow client passing in (#21034) 2024-04-29 17:14:26 +00:00
ccurme
84b8e67c9c
mistral: release 0.1.4 (#20940) 2024-04-26 13:06:02 -04:00
ccurme
fdabd3cdf5
mistral, openai: support custom tokenizers in chat models (#20901) 2024-04-25 15:23:29 -04:00
Erick Friis
5da9dd1195
mistral: comment batching param (#20868)
Addresses #20523
2024-04-25 00:38:21 +00:00
Erick Friis
8c95ac3145
docs, multiple: de-beta with_structured_output (#20850) 2024-04-24 19:34:57 +00:00
ccurme
6622829c67
mistral: catch GatedRepoError, release 0.1.3 (#20802)
https://github.com/langchain-ai/langchain/issues/20618

---------

Co-authored-by: Erick Friis <erick@langchain.dev>
2024-04-23 20:56:42 +00:00
Eugene Yurtsev
718c9cbe3a
mistral[patch]: Support both model and model_name (#20557) 2024-04-18 16:12:33 -04:00
Eugene Yurtsev
8c29b7bf35
mistralai[patch]: Use public attribute for eventsource.response (#20580)
Minor change, use the public attribute instead of the protected one.
2024-04-18 14:12:12 -04:00
ccurme
2238490069
mistral, openai: allow anthropic-style messages in message histories (#20565) 2024-04-17 15:55:45 -04:00
ccurme
4a17951900
mistral: read tool calls from AIMessage (#20554)
Co-authored-by: Eugene Yurtsev <eyurtsev@gmail.com>
2024-04-17 13:38:24 -04:00
Eugene Yurtsev
f257909699
mistralai[patch]: Surface http errors (#20555)
Do not swallow errors when streaming with httpx.

Update affected code if this PR gets merged to httpx:
https://github.com/florimondmanca/httpx-sse/pull/25/files
2024-04-17 10:47:56 -04:00
Bagatur
799714c629
release anthropic, fireworks, openai, groq, mistral (#20333) 2024-04-11 09:19:52 -07:00
ccurme
795c728f71
mistral[patch]: add IDs to tool calls (#20299)
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>
2024-04-11 11:09:30 -04:00
Erick Friis
9eb6f538f0
infra, multiple: rc release versions (#20252) 2024-04-09 17:54:58 -07:00
Bagatur
0d0458d1a7
mistralai[patch]: Pre-release 0.1.2-rc.1 (#20251) 2024-04-10 00:25:38 +00:00
Bagatur
9514bc4d67
core[minor], ...: add tool calls message (#18947)
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>
2024-04-09 18:41:42 -05:00
Bagatur
74d04a4e80
mistralai[patch]: Release 0.1.1 (#20239) 2024-04-09 21:53:01 +00:00
Erick Friis
855ba46f80
standard-tests: a standard unit and integration test set (#20182)
just chat models for now
2024-04-09 12:43:00 -07:00
Bagatur
3490d70238
mistralai[patch]: standardize model params (#20163)
Related to #20085
2024-04-08 11:48:38 -05:00
Bagatur
2f5606a318
mistralai[patch]: correct integration_test (#19774) 2024-03-29 21:47:35 +00:00
Pierre Véron
ace7b66261
mistralai[patch]: add missing _combine_llm_outputs implementation in ChatMistralAI (#18603)
# Description
Implementing `_combine_llm_outputs` to `ChatMistralAI` to override the
default implementation in `BaseChatModel` returning `{}`. The
implementation is inspired by the one in `ChatOpenAI` from package
`langchain-openai`.
# Issue
None
# Dependencies
None
# Twitter handle
None

---------

Co-authored-by: Bagatur <baskaryan@gmail.com>
2024-03-29 14:43:20 -07:00
Erick Friis
441a8012b3
mistralai[patch]: release 0.1.0 (#19540) 2024-03-25 17:29:40 -07:00
Erick Friis
b617085af0
mistralai[patch]: streaming tool calls (#19469) 2024-03-23 19:24:53 +00:00
aditya thomas
bc028294d0
docs: delete mistralai embeddings doc from incorrect location (#19432)
**Description:** Delete MistralAIEmbeddings usage document from folder
partners/mistralai/docs
**Issue:** The document is present in the folder docs/docs
**Dependencies:** None
2024-03-22 14:02:59 -07:00
Erick Friis
11e37943ed
mistralai[patch]: fix core version (#19454) 2024-03-22 20:48:13 +00:00
Erick Friis
3b093160c4
mistralai[patch]: release 0.1.0rc1 (#19453) 2024-03-22 20:34:36 +00:00
ccurme
c4599444ee
mistralai: update tool calling (#19451)
```python
from langchain.agents import tool
from langchain_mistralai import ChatMistralAI


llm = ChatMistralAI(model="mistral-large-latest", temperature=0)

@tool
def get_word_length(word: str) -> int:
    """Returns the length of a word."""
    return len(word)


tools = [get_word_length]
llm_with_tools = llm.bind_tools(tools)

llm_with_tools.invoke("how long is the word chrysanthemum")
```
currently raises
```
AttributeError: 'dict' object has no attribute 'model_dump'
```

Same with `.with_structured_output`
```python
from langchain_mistralai import ChatMistralAI
from langchain_core.pydantic_v1 import BaseModel

class AnswerWithJustification(BaseModel):
    """An answer to the user question along with justification for the answer."""
    answer: str
    justification: str

llm = ChatMistralAI(model="mistral-large-latest", temperature=0)
structured_llm = llm.with_structured_output(AnswerWithJustification)

structured_llm.invoke("What weighs more a pound of bricks or a pound of feathers")
```

This appears to fix.
2024-03-22 16:03:48 -04:00
Erick Friis
53ac1ebbbc
mistralai[minor]: 0.1.0rc0, remove mistral sdk (#19420) 2024-03-22 01:24:58 +00:00
Bagatur
242af4b5a4
openai[patch], mistral[patch], fireworks[patch]: releases 0.0.8, 0.0.5, 0.0.2 (#18186) 2024-02-27 04:22:24 -08:00
Bagatur
b3f4de38ae
mistral[minor]: Function calling and with_structured_output (#18150)
![Screenshot 2024-02-26 at 2 07 06
PM](https://github.com/langchain-ai/langchain/assets/22008038/20cacb47-3b24-45b5-871b-dd169f1acd37)
2024-02-26 16:22:30 -08:00
Erick Friis
a99c667c22
partners: version constraints (#17492)
Core should be ^0.1 by default

Careful about 0.x.y and 0.0.z packages
2024-02-14 08:57:46 -08:00
Erick Friis
3a2eb6e12b
infra: add print rule to ruff (#16221)
Added noqa for existing prints. Can slowly remove / will prevent more
being intro'd
2024-02-09 16:13:30 -08:00
Charlie Marsh
24c0bab57b
infra, multiple: Upgrade configuration for Ruff v0.2.0 (#16905)
## Summary

This PR upgrades LangChain's Ruff configuration in preparation for
Ruff's v0.2.0 release. (The changes are compatible with Ruff v0.1.5,
which LangChain uses today.) Specifically, we're now warning when
linter-only options are specified under `[tool.ruff]` instead of
`[tool.ruff.lint]`.

---------

Co-authored-by: Erick Friis <erick@langchain.dev>
Co-authored-by: Bagatur <baskaryan@gmail.com>
2024-02-09 14:28:02 -08:00
Erick Friis
3e58df43c2
mistralai[patch]: release 0.0.4 (#17139) 2024-02-06 16:05:20 -08:00
Erick Friis
f881a3330c
mistralai[patch]: 16k token batching logic embed (#17136) 2024-02-06 15:59:08 -08:00
ccurme
0826d87ecd
langchain_mistralai[patch]: Invoke callback prior to yielding token (#16986)
- **Description:** Invoke callback prior to yielding token in stream and
astream methods for ChatMistralAI.
- **Issue:** https://github.com/langchain-ai/langchain/issues/16913
2024-02-03 16:30:50 -08:00
Erick Friis
65b231d40b
mistralai[patch]: async integration tests (#16214) 2024-01-18 09:45:44 -08:00
Erick Friis
06fe2f4fb0
partners: add license field (#16117)
- bumps package post versions for packages without current unreleased
updates
- will bump package version in release prs associated with packages that
do have changes (mistral, vertex)
2024-01-17 08:37:13 -08:00
Erick Friis
ce10fe0c2f
mistralai[patch]: release 0.0.3 (#16116)
embeddings
2024-01-17 08:36:05 -08:00