mlflow.langchain

The mlflow.langchain module provides an API for logging and loading LangChain models. This module exports multivariate LangChain models in the langchain flavor and univariate LangChain models in the pyfunc flavor:

LangChain (native) format: This is the main flavor that can be accessed with LangChain APIs.
mlflow.pyfunc: Produced for use by generic pyfunc-based deployment tools and for batch inference.

mlflow.langchain.autolog(log_input_examples=False, log_model_signatures=False, log_models=False, log_datasets=False, log_inputs_outputs=None, disable=False, exclusive=False, disable_for_unsupported_versions=False, silent=False, registered_model_name=None, extra_tags=None, extra_model_classes=None, log_traces=True)[source]

Note

Experimental: This function may change or be removed in a future release without warning.

Note

Autologging is known to be compatible with the following package versions: 0.1.0 <= langchain <= 0.3.3. Autologging may not succeed when used with package versions outside of this range.

Enables (or disables) and configures autologging from Langchain to MLflow.

Parameters

log_input_examples – If True, input examples from inference data are collected and logged along with Langchain model artifacts during inference. If False, input examples are not logged. Note: Input examples are MLflow model attributes and are only collected if log_models is also True.
log_model_signatures – If True, ModelSignatures describing model inputs and outputs are collected and logged along with Langchain model artifacts during inference. If False, signatures are not logged. Note: Model signatures are MLflow model attributes and are only collected if log_models is also True.
log_models – If True, langchain models are logged as MLflow model artifacts. If False, langchain models are not logged. Input examples and model signatures, which are attributes of MLflow models, are also omitted when log_models is False.
log_datasets – If True, dataset information is logged to MLflow Tracking if applicable. If False, dataset information is not logged.
log_inputs_outputs – Deprecated The legacy parameter used for logging inference inputs and outputs. This argument will be removed in a future version of MLflow. The alternative is to use log_traces which logs traces for Langchain models, including inputs and outputs for each stage. If True, inference data and results are combined into a single pandas DataFrame and logged to MLflow Tracking as an artifact. If False, inference data and results are not logged. Default to False.
disable – If True, disables the Langchain autologging integration. If False, enables the Langchain autologging integration.
exclusive – If True, autologged content is not logged to user-created fluent runs. If False, autologged content is logged to the active fluent run, which may be user-created.
disable_for_unsupported_versions – If True, disable autologging for versions of langchain that have not been tested against this version of the MLflow client or are incompatible.
silent – If True, suppress all event logs and warnings from MLflow during Langchain autologging. If False, show all events and warnings during Langchain autologging.
registered_model_name – If given, each time a model is trained, it is registered as a new model version of the registered model with this name. The registered model is created if it does not already exist.
extra_tags – A dictionary of extra tags to set on each managed run created by autologging.
extra_model_classes – A list of langchain classes to log in addition to the default classes. We do not guarantee classes specified in this list can be logged as a model, but tracing will be supported. Note that all classes within the list must be subclasses of Runnable, and we only patch invoke, batch, and stream methods for tracing.
log_traces – If True, traces are logged for Langchain models by using MlflowLangchainTracer as a callback during inference. If False, no traces are collected during inference. Default to True.

mlflow.langchain.get_default_conda_env()[source]

Returns: The default Conda environment for MLflow Models produced by calls to save_model() and log_model().

mlflow.langchain.get_default_pip_requirements()[source]

Returns: A list of default pip requirements for MLflow Models produced by this flavor. Calls to save_model() and log_model() produce a pip environment that, at a minimum, contains these requirements.

mlflow.langchain.load_model(model_uri, dst_path=None)[source]

Note

Experimental: This function may change or be removed in a future release without warning.

Note

The ‘langchain’ MLflow Models integration is known to be compatible with 0.0.354 <= langchain <= 0.3.3. MLflow Models integrations with langchain may not succeed when used with package versions outside of this range.

Load a LangChain model from a local file or a run.

Parameters

model_uri –
The location, in URI format, of the MLflow model. For example:
- /Users/me/path/to/local/model
- relative/path/to/local/model
- s3://my_bucket/path/to/model
- runs:/<mlflow_run_id>/run-relative/path/to/model
For more information about supported URI schemes, see Referencing Artifacts.
dst_path – The local filesystem path to which to download the model artifact. This directory must already exist. If unspecified, a local output path will be created.

Returns

A LangChain model instance.

mlflow.langchain.log_model(lc_model, artifact_path, conda_env=None, code_paths=None, registered_model_name=None, signature: mlflow.models.signature.ModelSignature = None, input_example: Union[pandas.core.frame.DataFrame, numpy.ndarray, dict, list, csr_matrix, csc_matrix, str, bytes, tuple] = None, await_registration_for=300, pip_requirements=None, extra_pip_requirements=None, metadata=None, loader_fn=None, persist_dir=None, example_no_conversion=None, run_id=None, model_config=None, streamable=None, resources=None)[source]

Note

Experimental: This function may change or be removed in a future release without warning.

Note

The ‘langchain’ MLflow Models integration is known to be compatible with 0.0.354 <= langchain <= 0.3.3. MLflow Models integrations with langchain may not succeed when used with package versions outside of this range.

Log a LangChain model as an MLflow artifact for the current run.

Parameters

lc_model –
A LangChain model, which could be a Chain, Agent, or retriever or a path containing the LangChain model code <https://github.com/mlflow/mlflow/blob/master/examples/langchain/chain_as_code_driver.py> for the above types. When using model as path, make sure to set the model by using mlflow.models.set_model().

Note

Experimental: Using model as path may change or be removed in a future release without warning.
artifact_path – Run-relative artifact path.
conda_env –
Either a dictionary representation of a Conda environment or the path to a conda environment yaml file. If provided, this describes the environment this model should be run in. At a minimum, it should specify the dependencies contained in get_default_conda_env(). If None, a conda environment with pip requirements inferred by mlflow.models.infer_pip_requirements() is added to the model. If the requirement inference fails, it falls back to using get_default_pip_requirements(). pip requirements from conda_env are written to a pip requirements.txt file and the full conda environment is written to conda.yaml. The following is an example dictionary representation of a conda environment:
```
{
    "name": "mlflow-env",
    "channels": ["conda-forge"],
    "dependencies": [
        "python=3.8.15",
        {
            "pip": [
                "langchain==x.y.z"
            ],
        },
    ],
}
```
code_paths –
A list of local filesystem paths to Python file dependencies (or directories containing file dependencies). These files are prepended to the system path when the model is loaded. Files declared as dependencies for a given model should have relative imports declared from a common root path if multiple files are defined with import dependencies between them to avoid import errors when loading the model.

For a detailed explanation of code_paths functionality, recommended usage patterns and limitations, see the code_paths usage guide.
registered_model_name – This argument may change or be removed in a future release without warning. If given, create a model version under registered_model_name, also creating a registered model if one with the given name does not exist.
signature –
ModelSignature describes model input and output Schema. If not specified, the model signature would be set according to lc_model.input_keys and lc_model.output_keys as columns names, and DataType.string as the column type. Alternatively, you can explicitly specify the model signature. The model signature can be inferred from datasets with valid model input (e.g. the training dataset with target column omitted) and valid model output (e.g. model predictions generated on the training dataset), for example:
```
from mlflow.models import infer_signature

chain = LLMChain(llm=llm, prompt=prompt)
prediction = chain.run(input_str)
input_columns = [
    {"type": "string", "name": input_key} for input_key in chain.input_keys
]
signature = infer_signature(input_columns, predictions)
```
input_example – one or several instances of valid model input. The input example is used as a hint of what data to feed the model. It will be converted to a Pandas DataFrame and then serialized to json using the Pandas split-oriented format, or a numpy array where the example will be serialized to json by converting it to a list. Bytes are base64-encoded. When the signature parameter is None, the input example is used to infer a model signature.
await_registration_for – Number of seconds to wait for the model version to finish being created and is in READY status. By default, the function waits for five minutes. Specify 0 or None to skip waiting.
pip_requirements – Either an iterable of pip requirement strings (e.g. ["langchain", "-r requirements.txt", "-c constraints.txt"]) or the string path to a pip requirements file on the local filesystem (e.g. "requirements.txt"). If provided, this describes the environment this model should be run in. If None, a default list of requirements is inferred by mlflow.models.infer_pip_requirements() from the current software environment. If the requirement inference fails, it falls back to using get_default_pip_requirements(). Both requirements and constraints are automatically parsed and written to requirements.txt and constraints.txt files, respectively, and stored as part of the model. Requirements are also written to the pip section of the model’s conda environment (conda.yaml) file.
extra_pip_requirements –
Either an iterable of pip requirement strings (e.g. ["pandas", "-r requirements.txt", "-c constraints.txt"]) or the string path to a pip requirements file on the local filesystem (e.g. "requirements.txt"). If provided, this describes additional pip requirements that are appended to a default set of pip requirements generated automatically based on the user’s current software environment. Both requirements and constraints are automatically parsed and written to requirements.txt and constraints.txt files, respectively, and stored as part of the model. Requirements are also written to the pip section of the model’s conda environment (conda.yaml) file.
Warning

The following arguments can’t be specified at the same time:
- conda_env
- pip_requirements
- extra_pip_requirements
This example demonstrates how to specify pip requirements using pip_requirements and extra_pip_requirements.
metadata – Custom metadata dictionary passed to the model and stored in the MLmodel file.
loader_fn –
A function that’s required for models containing objects that aren’t natively serialized by LangChain. This function takes a string persist_dir as an argument and returns the specific object that the model needs. Depending on the model, this could be a retriever, vectorstore, requests_wrapper, embeddings, or database. For RetrievalQA Chain and retriever models, the object is a (retriever). For APIChain models, it’s a (requests_wrapper). For HypotheticalDocumentEmbedder models, it’s an (embeddings). For SQLDatabaseChain models, it’s a (database).

persist_dir –

The directory where the object is stored. The loader_fn takes this string as the argument to load the object. This is optional for models containing objects that aren’t natively serialized by LangChain. MLflow logs the content in this directory as artifacts in the subdirectory named persist_dir_data.

Here is the code snippet for logging a RetrievalQA chain with loader_fn and persist_dir:

Note

In langchain_community >= 0.0.27, loading pickled data requires providing the allow_dangerous_deserialization argument.

qa = RetrievalQA.from_llm(llm=OpenAI(), retriever=db.as_retriever())


def load_retriever(persist_directory):
    embeddings = OpenAIEmbeddings()
    vectorstore = FAISS.load_local(
        persist_directory,
        embeddings,
        # you may need to add the line below
        # for langchain_community >= 0.0.27
        allow_dangerous_deserialization=True,
    )
    return vectorstore.as_retriever()


with mlflow.start_run() as run:
    logged_model = mlflow.langchain.log_model(
        qa,
        artifact_path="retrieval_qa",
        loader_fn=load_retriever,
        persist_dir=persist_dir,
    )

See a complete example in examples/langchain/retrieval_qa_chain.py.

example_no_conversion – This parameter is deprecated and will be removed in a future release. It’s no longer used and can be safely removed. Input examples are not converted anymore.
run_id – run_id to associate with this model version. If specified, we resume the run and log the model to that run. Otherwise, a new run is created. Default to None.
model_config –
The model configuration to apply to the model if saving model from code. This configuration is available during model loading.

Note

Experimental: This parameter may change or be removed in a future release without warning.
streamable – A boolean value indicating if the model supports streaming prediction. If True, the model must implement stream method. If None, If None, streamable is set to True if the model implements stream method. Default to None.
resources –

A list of model resources or a resources.yaml file containing a list of
resources required to serve the model.

Note

Experimental: This parameter may change or be removed in a future release without warning.

Returns

A ModelInfo instance that contains the metadata of the logged model.

mlflow.langchain.save_model(lc_model, path, conda_env=None, code_paths=None, mlflow_model=None, signature: mlflow.models.signature.ModelSignature = None, input_example: Union[pandas.core.frame.DataFrame, numpy.ndarray, dict, list, csr_matrix, csc_matrix, str, bytes, tuple] = None, pip_requirements=None, extra_pip_requirements=None, metadata=None, loader_fn=None, persist_dir=None, example_no_conversion=None, model_config=None, streamable: Optional[bool] = None)[source]

Note

Experimental: This function may change or be removed in a future release without warning.

Note

The ‘langchain’ MLflow Models integration is known to be compatible with 0.0.354 <= langchain <= 0.3.3. MLflow Models integrations with langchain may not succeed when used with package versions outside of this range.

Save a LangChain model to a path on the local file system.

Parameters

lc_model –
A LangChain model, which could be a Chain, Agent, retriever, or RunnableSequence, or a path containing the LangChain model code <https://github.com/mlflow/mlflow/blob/master/examples/langchain/chain_as_code_driver.py> for the above types. When using model as path, make sure to set the model by using mlflow.models.set_model().

Note

Experimental: Using model as path may change or be removed in a future release without warning.
path – Local path where the serialized model (as YAML) is to be saved.
conda_env –
Either a dictionary representation of a Conda environment or the path to a conda environment yaml file. If provided, this describes the environment this model should be run in. At a minimum, it should specify the dependencies contained in get_default_conda_env(). If None, a conda environment with pip requirements inferred by mlflow.models.infer_pip_requirements() is added to the model. If the requirement inference fails, it falls back to using get_default_pip_requirements(). pip requirements from conda_env are written to a pip requirements.txt file and the full conda environment is written to conda.yaml. The following is an example dictionary representation of a conda environment:
```
{
    "name": "mlflow-env",
    "channels": ["conda-forge"],
    "dependencies": [
        "python=3.8.15",
        {
            "pip": [
                "langchain==x.y.z"
            ],
        },
    ],
}
```
code_paths –
A list of local filesystem paths to Python file dependencies (or directories containing file dependencies). These files are prepended to the system path when the model is loaded. Files declared as dependencies for a given model should have relative imports declared from a common root path if multiple files are defined with import dependencies between them to avoid import errors when loading the model.

For a detailed explanation of code_paths functionality, recommended usage patterns and limitations, see the code_paths usage guide.
mlflow_model – mlflow.models.Model this flavor is being added to.
signature –
ModelSignature describes model input and output Schema. If not specified, the model signature would be set according to lc_model.input_keys and lc_model.output_keys as columns names, and DataType.string as the column type. Alternatively, you can explicitly specify the model signature. The model signature can be inferred from datasets with valid model input (e.g. the training dataset with target column omitted) and valid model output (e.g. model predictions generated on the training dataset), for example:
```
from mlflow.models import infer_signature

chain = LLMChain(llm=llm, prompt=prompt)
prediction = chain.run(input_str)
input_columns = [
    {"type": "string", "name": input_key} for input_key in chain.input_keys
]
signature = infer_signature(input_columns, predictions)
```
input_example – one or several instances of valid model input. The input example is used as a hint of what data to feed the model. It will be converted to a Pandas DataFrame and then serialized to json using the Pandas split-oriented format, or a numpy array where the example will be serialized to json by converting it to a list. Bytes are base64-encoded. When the signature parameter is None, the input example is used to infer a model signature.
pip_requirements – Either an iterable of pip requirement strings (e.g. ["langchain", "-r requirements.txt", "-c constraints.txt"]) or the string path to a pip requirements file on the local filesystem (e.g. "requirements.txt"). If provided, this describes the environment this model should be run in. If None, a default list of requirements is inferred by mlflow.models.infer_pip_requirements() from the current software environment. If the requirement inference fails, it falls back to using get_default_pip_requirements(). Both requirements and constraints are automatically parsed and written to requirements.txt and constraints.txt files, respectively, and stored as part of the model. Requirements are also written to the pip section of the model’s conda environment (conda.yaml) file.
extra_pip_requirements –
Either an iterable of pip requirement strings (e.g. ["pandas", "-r requirements.txt", "-c constraints.txt"]) or the string path to a pip requirements file on the local filesystem (e.g. "requirements.txt"). If provided, this describes additional pip requirements that are appended to a default set of pip requirements generated automatically based on the user’s current software environment. Both requirements and constraints are automatically parsed and written to requirements.txt and constraints.txt files, respectively, and stored as part of the model. Requirements are also written to the pip section of the model’s conda environment (conda.yaml) file.
Warning

The following arguments can’t be specified at the same time:
- conda_env
- pip_requirements
- extra_pip_requirements
This example demonstrates how to specify pip requirements using pip_requirements and extra_pip_requirements.
metadata – Custom metadata dictionary passed to the model and stored in the MLmodel file.
loader_fn –
A function that’s required for models containing objects that aren’t natively serialized by LangChain. This function takes a string persist_dir as an argument and returns the specific object that the model needs. Depending on the model, this could be a retriever, vectorstore, requests_wrapper, embeddings, or database. For RetrievalQA Chain and retriever models, the object is a (retriever). For APIChain models, it’s a (requests_wrapper). For HypotheticalDocumentEmbedder models, it’s an (embeddings). For SQLDatabaseChain models, it’s a (database).

persist_dir –

The directory where the object is stored. The loader_fn takes this string as the argument to load the object. This is optional for models containing objects that aren’t natively serialized by LangChain. MLflow logs the content in this directory as artifacts in the subdirectory named persist_dir_data.

Here is the code snippet for logging a RetrievalQA chain with loader_fn and persist_dir:

Note

In langchain_community >= 0.0.27, loading pickled data requires providing the allow_dangerous_deserialization argument.

qa = RetrievalQA.from_llm(llm=OpenAI(), retriever=db.as_retriever())


def load_retriever(persist_directory):
    embeddings = OpenAIEmbeddings()
    vectorstore = FAISS.load_local(
        persist_directory,
        embeddings,
        # you may need to add the line below
        # for langchain_community >= 0.0.27
        allow_dangerous_deserialization=True,
    )
    return vectorstore.as_retriever()


with mlflow.start_run() as run:
    logged_model = mlflow.langchain.log_model(
        qa,
        artifact_path="retrieval_qa",
        loader_fn=load_retriever,
        persist_dir=persist_dir,
    )

See a complete example in examples/langchain/retrieval_qa_chain.py.

example_no_conversion – This parameter is deprecated and will be removed in a future release. It’s no longer used and can be safely removed. Input examples are not converted anymore.
model_config –
The model configuration to apply to the model if saving model from code. This configuration is available during model loading.

Note

Experimental: This parameter may change or be removed in a future release without warning.
streamable – A boolean value indicating if the model supports streaming prediction. If True, the model must implement stream method. If None, streamable is set to True if the model implements stream method. Default to None.