id
stringlengths 14
15
| text
stringlengths 49
2.47k
| source
stringlengths 61
166
|
|---|---|---|
549a5b53223c-1
|
yield_keys(*[, prefix])
Get an iterator over keys that match the given prefix.
__init__(store: BaseStore[str, Any], key_encoder: Callable[[K], str], value_serializer: Callable[[V], bytes], value_deserializer: Callable[[Any], V]) → None[source]¶
Initialize an EncodedStore.
mdelete(keys: Sequence[K]) → None[source]¶
Delete the given keys and their associated values.
mget(keys: Sequence[K]) → List[Optional[V]][source]¶
Get the values associated with the given keys.
mset(key_value_pairs: Sequence[Tuple[K, V]]) → None[source]¶
Set the values for the given keys.
yield_keys(*, prefix: Optional[str] = None) → Union[Iterator[K], Iterator[str]][source]¶
Get an iterator over keys that match the given prefix.
|
https://api.python.langchain.com/en/latest/storage/langchain.storage.encoder_backed.EncoderBackedStore.html
|
ff9dd5d7964c-0
|
langchain.storage.file_system.LocalFileStore¶
class langchain.storage.file_system.LocalFileStore(root_path: Union[str, Path])[source]¶
BaseStore interface that works on the local file system.
Examples
Create a LocalFileStore instance and perform operations on it:
from langchain.storage import LocalFileStore
# Instantiate the LocalFileStore with the root path
file_store = LocalFileStore("/path/to/root")
# Set values for keys
file_store.mset([("key1", b"value1"), ("key2", b"value2")])
# Get values for keys
values = file_store.mget(["key1", "key2"]) # Returns [b"value1", b"value2"]
# Delete keys
file_store.mdelete(["key1"])
# Iterate over keys
for key in file_store.yield_keys():
print(key)
Implement the BaseStore interface for the local file system.
Parameters
root_path (Union[str, Path]) – The root path of the file store. All keys are
interpreted as paths relative to this root.
Methods
__init__(root_path)
Implement the BaseStore interface for the local file system.
mdelete(keys)
Delete the given keys and their associated values.
mget(keys)
Get the values associated with the given keys.
mset(key_value_pairs)
Set the values for the given keys.
yield_keys([prefix])
Get an iterator over keys that match the given prefix.
__init__(root_path: Union[str, Path]) → None[source]¶
Implement the BaseStore interface for the local file system.
Parameters
root_path (Union[str, Path]) – The root path of the file store. All keys are
interpreted as paths relative to this root.
mdelete(keys: Sequence[str]) → None[source]¶
Delete the given keys and their associated values.
Parameters
|
https://api.python.langchain.com/en/latest/storage/langchain.storage.file_system.LocalFileStore.html
|
ff9dd5d7964c-1
|
Delete the given keys and their associated values.
Parameters
keys (Sequence[str]) – A sequence of keys to delete.
Returns
None
mget(keys: Sequence[str]) → List[Optional[bytes]][source]¶
Get the values associated with the given keys.
Parameters
keys – A sequence of keys.
Returns
A sequence of optional values associated with the keys.
If a key is not found, the corresponding value will be None.
mset(key_value_pairs: Sequence[Tuple[str, bytes]]) → None[source]¶
Set the values for the given keys.
Parameters
key_value_pairs – A sequence of key-value pairs.
Returns
None
yield_keys(prefix: Optional[str] = None) → Iterator[str][source]¶
Get an iterator over keys that match the given prefix.
Parameters
prefix (Optional[str]) – The prefix to match.
Returns
An iterator over keys that match the given prefix.
Return type
Iterator[str]
|
https://api.python.langchain.com/en/latest/storage/langchain.storage.file_system.LocalFileStore.html
|
65bfb40833c0-0
|
langchain.evaluation.criteria.eval_chain.resolve_criteria¶
langchain.evaluation.criteria.eval_chain.resolve_criteria(criteria: Optional[Union[Mapping[str, str], Criteria, ConstitutionalPrinciple, str]]) → Dict[str, str][source]¶
Resolve the criteria to evaluate.
Parameters
criteria (CRITERIA_TYPE) –
The criteria to evaluate the runs against. It can be:
a mapping of a criterion name to its description
a single criterion name present in one of the default criteria
a single ConstitutionalPrinciple instance
Returns
A dictionary mapping criterion names to descriptions.
Return type
Dict[str, str]
Examples
>>> criterion = "relevance"
>>> CriteriaEvalChain.resolve_criteria(criteria)
{'relevance': 'Is the submission referring to a real quote from the text?'}
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.criteria.eval_chain.resolve_criteria.html
|
04b1d360fbb8-0
|
langchain.evaluation.criteria.eval_chain.CriteriaResultOutputParser¶
class langchain.evaluation.criteria.eval_chain.CriteriaResultOutputParser[source]¶
Bases: BaseOutputParser[dict]
A parser for the output of the CriteriaEvalChain.
Create a new model by parsing and validating input data from keyword arguments.
Raises ValidationError if the input data cannot be parsed to form a valid model.
async abatch(inputs: List[Input], config: Optional[Union[RunnableConfig, List[RunnableConfig]]] = None, *, max_concurrency: Optional[int] = None) → List[Output]¶
async ainvoke(input: str | langchain.schema.messages.BaseMessage, config: langchain.schema.runnable.RunnableConfig | None = None) → T¶
async aparse(text: str) → T¶
Parse a single string model output into some structure.
Parameters
text – String output of a language model.
Returns
Structured output.
async aparse_result(result: List[Generation]) → T¶
Parse a list of candidate model Generations into a specific format.
The return value is parsed from only the first Generation in the result, whichis assumed to be the highest-likelihood Generation.
Parameters
result – A list of Generations to be parsed. The Generations are assumed
to be different candidate outputs for a single model input.
Returns
Structured output.
async astream(input: Input, config: Optional[RunnableConfig] = None) → AsyncIterator[Output]¶
batch(inputs: List[Input], config: Optional[Union[RunnableConfig, List[RunnableConfig]]] = None, *, max_concurrency: Optional[int] = None) → List[Output]¶
bind(**kwargs: Any) → Runnable[Input, Output]¶
Bind arguments to a Runnable, returning a new Runnable.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.criteria.eval_chain.CriteriaResultOutputParser.html
|
04b1d360fbb8-1
|
Bind arguments to a Runnable, returning a new Runnable.
classmethod construct(_fields_set: Optional[SetStr] = None, **values: Any) → Model¶
Creates a new model setting __dict__ and __fields_set__ from trusted or pre-validated data.
Default values are respected, but no other validation is performed.
Behaves as if Config.extra = ‘allow’ was set since it adds all passed values
copy(*, include: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, exclude: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, update: Optional[DictStrAny] = None, deep: bool = False) → Model¶
Duplicate a model, optionally choose which fields to include, exclude and change.
Parameters
include – fields to include in new model
exclude – fields to exclude from new model, as with values this takes precedence over include
update – values to change/add in the new model. Note: the data is not validated before creating
the new model: you should trust this data
deep – set to True to make a deep copy of the model
Returns
new model instance
dict(**kwargs: Any) → Dict¶
Return dictionary representation of output parser.
classmethod from_orm(obj: Any) → Model¶
get_format_instructions() → str¶
Instructions on how the LLM output should be formatted.
invoke(input: str | langchain.schema.messages.BaseMessage, config: langchain.schema.runnable.RunnableConfig | None = None) → T¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.criteria.eval_chain.CriteriaResultOutputParser.html
|
04b1d360fbb8-2
|
json(*, include: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, exclude: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, by_alias: bool = False, skip_defaults: Optional[bool] = None, exclude_unset: bool = False, exclude_defaults: bool = False, exclude_none: bool = False, encoder: Optional[Callable[[Any], Any]] = None, models_as_dict: bool = True, **dumps_kwargs: Any) → unicode¶
Generate a JSON representation of the model, include and exclude arguments as per dict().
encoder is an optional function to supply as default to json.dumps(), other arguments as per json.dumps().
parse(text: str) → Dict[str, Any][source]¶
Parse the output text.
Parameters
text (str) – The output text to parse.
Returns
The parsed output.
Return type
Dict
classmethod parse_file(path: Union[str, Path], *, content_type: unicode = None, encoding: unicode = 'utf8', proto: Protocol = None, allow_pickle: bool = False) → Model¶
classmethod parse_obj(obj: Any) → Model¶
classmethod parse_raw(b: Union[str, bytes], *, content_type: unicode = None, encoding: unicode = 'utf8', proto: Protocol = None, allow_pickle: bool = False) → Model¶
parse_result(result: List[Generation]) → T¶
Parse a list of candidate model Generations into a specific format.
The return value is parsed from only the first Generation in the result, whichis assumed to be the highest-likelihood Generation.
Parameters
result – A list of Generations to be parsed. The Generations are assumed
to be different candidate outputs for a single model input.
Returns
Structured output.
parse_with_prompt(completion: str, prompt: PromptValue) → Any¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.criteria.eval_chain.CriteriaResultOutputParser.html
|
04b1d360fbb8-3
|
Structured output.
parse_with_prompt(completion: str, prompt: PromptValue) → Any¶
Parse the output of an LLM call with the input prompt for context.
The prompt is largely provided in the event the OutputParser wants
to retry or fix the output in some way, and needs information from
the prompt to do so.
Parameters
completion – String output of a language model.
prompt – Input PromptValue.
Returns
Structured output
classmethod schema(by_alias: bool = True, ref_template: unicode = '#/definitions/{model}') → DictStrAny¶
classmethod schema_json(*, by_alias: bool = True, ref_template: unicode = '#/definitions/{model}', **dumps_kwargs: Any) → unicode¶
stream(input: Input, config: Optional[RunnableConfig] = None) → Iterator[Output]¶
to_json() → Union[SerializedConstructor, SerializedNotImplemented]¶
to_json_not_implemented() → SerializedNotImplemented¶
classmethod update_forward_refs(**localns: Any) → None¶
Try to update ForwardRefs on fields based on this Model, globalns and localns.
classmethod validate(value: Any) → Model¶
with_fallbacks(fallbacks: ~typing.Sequence[~langchain.schema.runnable.Runnable[~langchain.schema.runnable.Input, ~langchain.schema.runnable.Output]], *, exceptions_to_handle: ~typing.Tuple[~typing.Type[BaseException]] = (<class 'Exception'>,)) → RunnableWithFallbacks[Input, Output]¶
property lc_attributes: Dict¶
Return a list of attribute names that should be included in the
serialized kwargs. These attributes must be accepted by the
constructor.
property lc_namespace: List[str]¶
Return the namespace of the langchain object.
eg. [“langchain”, “llms”, “openai”]
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.criteria.eval_chain.CriteriaResultOutputParser.html
|
04b1d360fbb8-4
|
eg. [“langchain”, “llms”, “openai”]
property lc_secrets: Dict[str, str]¶
Return a map of constructor argument names to secret ids.
eg. {“openai_api_key”: “OPENAI_API_KEY”}
property lc_serializable: bool¶
Return whether or not the class is serializable.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.criteria.eval_chain.CriteriaResultOutputParser.html
|
e02b7b494139-0
|
langchain.evaluation.comparison.eval_chain.LabeledPairwiseStringEvalChain¶
class langchain.evaluation.comparison.eval_chain.LabeledPairwiseStringEvalChain[source]¶
Bases: PairwiseStringEvalChain
A chain for comparing two outputs, such as the outputsof two models, prompts, or outputs of a single model on similar inputs,
with labeled preferences.
output_parser¶
The output parser for the chain.
Type
BaseOutputParser
param callback_manager: Optional[BaseCallbackManager] = None¶
Deprecated, use callbacks instead.
param callbacks: Callbacks = None¶
Optional list of callback handlers (or callback manager). Defaults to None.
Callback handlers are called throughout the lifecycle of a call to a chain,
starting with on_chain_start, ending with on_chain_end or on_chain_error.
Each custom chain can optionally call additional callback methods, see Callback docs
for full details.
param llm: BaseLanguageModel [Required]¶
Language model to call.
param llm_kwargs: dict [Optional]¶
param memory: Optional[BaseMemory] = None¶
Optional memory object. Defaults to None.
Memory is a class that gets called at the start
and at the end of every chain. At the start, memory loads variables and passes
them along in the chain. At the end, it saves any returned variables.
There are many different types of memory - please see memory docs
for the full catalog.
param metadata: Optional[Dict[str, Any]] = None¶
Optional metadata associated with the chain. Defaults to None.
This metadata will be associated with each call to this chain,
and passed as arguments to the handlers defined in callbacks.
You can use these to eg identify a specific instance of a chain with its use case.
param output_key: str = 'results'¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.comparison.eval_chain.LabeledPairwiseStringEvalChain.html
|
e02b7b494139-1
|
param output_key: str = 'results'¶
param output_parser: BaseOutputParser [Optional]¶
Output parser to use.
Defaults to one that takes the most likely string but does not change it
otherwise.
param prompt: BasePromptTemplate [Required]¶
Prompt object to use.
param return_final_only: bool = True¶
Whether to return only the final parsed result. Defaults to True.
If false, will return a bunch of extra information about the generation.
param tags: Optional[List[str]] = None¶
Optional list of tags associated with the chain. Defaults to None.
These tags will be associated with each call to this chain,
and passed as arguments to the handlers defined in callbacks.
You can use these to eg identify a specific instance of a chain with its use case.
param verbose: bool [Optional]¶
Whether or not run in verbose mode. In verbose mode, some intermediate logs
will be printed to the console. Defaults to langchain.verbose value.
__call__(inputs: Union[Dict[str, Any], Any], return_only_outputs: bool = False, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, *, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, include_run_info: bool = False) → Dict[str, Any]¶
Execute the chain.
Parameters
inputs – Dictionary of inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
return_only_outputs – Whether to return only outputs in the
response. If True, only new keys generated by this chain will be
returned. If False, both input keys and new keys generated by this
chain will be returned. Defaults to False.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.comparison.eval_chain.LabeledPairwiseStringEvalChain.html
|
e02b7b494139-2
|
chain will be returned. Defaults to False.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
metadata – Optional metadata associated with the chain. Defaults to None
include_run_info – Whether to include run info in the response. Defaults
to False.
Returns
A dict of named outputs. Should contain all outputs specified inChain.output_keys.
async aapply(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → List[Dict[str, str]]¶
Utilize the LLM generate method for speed gains.
async aapply_and_parse(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → Sequence[Union[str, List[str], Dict[str, str]]]¶
Call apply and then parse the results.
async abatch(inputs: List[Input], config: Optional[Union[RunnableConfig, List[RunnableConfig]]] = None, *, max_concurrency: Optional[int] = None) → List[Output]¶
async acall(inputs: Union[Dict[str, Any], Any], return_only_outputs: bool = False, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, *, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, include_run_info: bool = False) → Dict[str, Any]¶
Asynchronously execute the chain.
Parameters
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.comparison.eval_chain.LabeledPairwiseStringEvalChain.html
|
e02b7b494139-3
|
Asynchronously execute the chain.
Parameters
inputs – Dictionary of inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
return_only_outputs – Whether to return only outputs in the
response. If True, only new keys generated by this chain will be
returned. If False, both input keys and new keys generated by this
chain will be returned. Defaults to False.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
metadata – Optional metadata associated with the chain. Defaults to None
include_run_info – Whether to include run info in the response. Defaults
to False.
Returns
A dict of named outputs. Should contain all outputs specified inChain.output_keys.
async aevaluate_string_pairs(*, prediction: str, prediction_b: str, reference: Optional[str] = None, input: Optional[str] = None, **kwargs: Any) → dict¶
Asynchronously evaluate the output string pairs.
Parameters
prediction (str) – The output string from the first model.
prediction_b (str) – The output string from the second model.
reference (Optional[str], optional) – The expected output / reference string.
input (Optional[str], optional) – The input string.
**kwargs – Additional keyword arguments, such as callbacks and optional reference strings.
Returns
A dictionary containing the preference, scores, and/or other information.
Return type
dict
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.comparison.eval_chain.LabeledPairwiseStringEvalChain.html
|
e02b7b494139-4
|
A dictionary containing the preference, scores, and/or other information.
Return type
dict
async agenerate(input_list: List[Dict[str, Any]], run_manager: Optional[AsyncCallbackManagerForChainRun] = None) → LLMResult¶
Generate LLM result from inputs.
async ainvoke(input: Dict[str, Any], config: Optional[RunnableConfig] = None) → Dict[str, Any]¶
apply(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → List[Dict[str, str]]¶
Utilize the LLM generate method for speed gains.
apply_and_parse(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → Sequence[Union[str, List[str], Dict[str, str]]]¶
Call apply and then parse the results.
async apredict(callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, **kwargs: Any) → str¶
Format prompt with kwargs and pass to LLM.
Parameters
callbacks – Callbacks to pass to LLMChain
**kwargs – Keys to pass to prompt template.
Returns
Completion from LLM.
Example
completion = llm.predict(adjective="funny")
async apredict_and_parse(callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, **kwargs: Any) → Union[str, List[str], Dict[str, str]]¶
Call apredict and then parse the results.
async aprep_prompts(input_list: List[Dict[str, Any]], run_manager: Optional[AsyncCallbackManagerForChainRun] = None) → Tuple[List[PromptValue], Optional[List[str]]]¶
Prepare prompts from inputs.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.comparison.eval_chain.LabeledPairwiseStringEvalChain.html
|
e02b7b494139-5
|
Prepare prompts from inputs.
async arun(*args: Any, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, **kwargs: Any) → Any¶
Convenience method for executing chain.
The main difference between this method and Chain.__call__ is that this
method expects inputs to be passed directly in as positional arguments or
keyword arguments, whereas Chain.__call__ expects a single input dictionary
with all the inputs
Parameters
*args – If the chain expects a single input, it can be passed in as the
sole positional argument.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
**kwargs – If the chain expects multiple inputs, they can be passed in
directly as keyword arguments.
Returns
The chain output.
Example
# Suppose we have a single-input chain that takes a 'question' string:
await chain.arun("What's the temperature in Boise, Idaho?")
# -> "The temperature in Boise is..."
# Suppose we have a multi-input chain that takes a 'question' string
# and 'context' string:
question = "What's the temperature in Boise, Idaho?"
context = "Weather report for Boise, Idaho on 07/03/23..."
await chain.arun(question=question, context=context)
# -> "The temperature in Boise is..."
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.comparison.eval_chain.LabeledPairwiseStringEvalChain.html
|
e02b7b494139-6
|
# -> "The temperature in Boise is..."
async astream(input: Input, config: Optional[RunnableConfig] = None) → AsyncIterator[Output]¶
batch(inputs: List[Input], config: Optional[Union[RunnableConfig, List[RunnableConfig]]] = None, *, max_concurrency: Optional[int] = None) → List[Output]¶
bind(**kwargs: Any) → Runnable[Input, Output]¶
Bind arguments to a Runnable, returning a new Runnable.
classmethod construct(_fields_set: Optional[SetStr] = None, **values: Any) → Model¶
Creates a new model setting __dict__ and __fields_set__ from trusted or pre-validated data.
Default values are respected, but no other validation is performed.
Behaves as if Config.extra = ‘allow’ was set since it adds all passed values
copy(*, include: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, exclude: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, update: Optional[DictStrAny] = None, deep: bool = False) → Model¶
Duplicate a model, optionally choose which fields to include, exclude and change.
Parameters
include – fields to include in new model
exclude – fields to exclude from new model, as with values this takes precedence over include
update – values to change/add in the new model. Note: the data is not validated before creating
the new model: you should trust this data
deep – set to True to make a deep copy of the model
Returns
new model instance
create_outputs(llm_result: LLMResult) → List[Dict[str, Any]]¶
Create outputs from response.
dict(**kwargs: Any) → Dict¶
Dictionary representation of chain.
Expects Chain._chain_type property to be implemented and for memory to benull.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.comparison.eval_chain.LabeledPairwiseStringEvalChain.html
|
e02b7b494139-7
|
Expects Chain._chain_type property to be implemented and for memory to benull.
Parameters
**kwargs – Keyword arguments passed to default pydantic.BaseModel.dict
method.
Returns
A dictionary representation of the chain.
Example
..code-block:: python
chain.dict(exclude_unset=True)
# -> {“_type”: “foo”, “verbose”: False, …}
evaluate_string_pairs(*, prediction: str, prediction_b: str, reference: Optional[str] = None, input: Optional[str] = None, **kwargs: Any) → dict¶
Evaluate the output string pairs.
Parameters
prediction (str) – The output string from the first model.
prediction_b (str) – The output string from the second model.
reference (Optional[str], optional) – The expected output / reference string.
input (Optional[str], optional) – The input string.
**kwargs – Additional keyword arguments, such as callbacks and optional reference strings.
Returns
A dictionary containing the preference, scores, and/or other information.
Return type
dict
classmethod from_llm(llm: BaseLanguageModel, *, prompt: Optional[PromptTemplate] = None, criteria: Optional[Union[Mapping[str, str], Criteria, ConstitutionalPrinciple, str]] = None, **kwargs: Any) → PairwiseStringEvalChain[source]¶
Initialize the LabeledPairwiseStringEvalChain from an LLM.
Parameters
llm (BaseLanguageModel) – The LLM to use.
prompt (PromptTemplate, optional) – The prompt to use.
criteria (Union[CRITERIA_TYPE, str], optional) – The criteria to use.
**kwargs (Any) – Additional keyword arguments.
Returns
The initialized LabeledPairwiseStringEvalChain.
Return type
LabeledPairwiseStringEvalChain
Raises
ValueError – If the input variables are not as expected.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.comparison.eval_chain.LabeledPairwiseStringEvalChain.html
|
e02b7b494139-8
|
Raises
ValueError – If the input variables are not as expected.
classmethod from_orm(obj: Any) → Model¶
classmethod from_string(llm: BaseLanguageModel, template: str) → LLMChain¶
Create LLMChain from LLM and template.
generate(input_list: List[Dict[str, Any]], run_manager: Optional[CallbackManagerForChainRun] = None) → LLMResult¶
Generate LLM result from inputs.
invoke(input: Dict[str, Any], config: Optional[RunnableConfig] = None) → Dict[str, Any]¶
json(*, include: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, exclude: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, by_alias: bool = False, skip_defaults: Optional[bool] = None, exclude_unset: bool = False, exclude_defaults: bool = False, exclude_none: bool = False, encoder: Optional[Callable[[Any], Any]] = None, models_as_dict: bool = True, **dumps_kwargs: Any) → unicode¶
Generate a JSON representation of the model, include and exclude arguments as per dict().
encoder is an optional function to supply as default to json.dumps(), other arguments as per json.dumps().
classmethod parse_file(path: Union[str, Path], *, content_type: unicode = None, encoding: unicode = 'utf8', proto: Protocol = None, allow_pickle: bool = False) → Model¶
classmethod parse_obj(obj: Any) → Model¶
classmethod parse_raw(b: Union[str, bytes], *, content_type: unicode = None, encoding: unicode = 'utf8', proto: Protocol = None, allow_pickle: bool = False) → Model¶
predict(callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, **kwargs: Any) → str¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.comparison.eval_chain.LabeledPairwiseStringEvalChain.html
|
e02b7b494139-9
|
Format prompt with kwargs and pass to LLM.
Parameters
callbacks – Callbacks to pass to LLMChain
**kwargs – Keys to pass to prompt template.
Returns
Completion from LLM.
Example
completion = llm.predict(adjective="funny")
predict_and_parse(callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, **kwargs: Any) → Union[str, List[str], Dict[str, Any]]¶
Call predict and then parse the results.
prep_inputs(inputs: Union[Dict[str, Any], Any]) → Dict[str, str]¶
Validate and prepare chain inputs, including adding inputs from memory.
Parameters
inputs – Dictionary of raw inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
Returns
A dictionary of all inputs, including those added by the chain’s memory.
prep_outputs(inputs: Dict[str, str], outputs: Dict[str, str], return_only_outputs: bool = False) → Dict[str, str]¶
Validate and prepare chain outputs, and save info about this run to memory.
Parameters
inputs – Dictionary of chain inputs, including any inputs added by chain
memory.
outputs – Dictionary of initial chain outputs.
return_only_outputs – Whether to only return the chain outputs. If False,
inputs are also added to the final outputs.
Returns
A dict of the final chain outputs.
prep_prompts(input_list: List[Dict[str, Any]], run_manager: Optional[CallbackManagerForChainRun] = None) → Tuple[List[PromptValue], Optional[List[str]]]¶
Prepare prompts from inputs.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.comparison.eval_chain.LabeledPairwiseStringEvalChain.html
|
e02b7b494139-10
|
Prepare prompts from inputs.
run(*args: Any, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, **kwargs: Any) → Any¶
Convenience method for executing chain.
The main difference between this method and Chain.__call__ is that this
method expects inputs to be passed directly in as positional arguments or
keyword arguments, whereas Chain.__call__ expects a single input dictionary
with all the inputs
Parameters
*args – If the chain expects a single input, it can be passed in as the
sole positional argument.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
**kwargs – If the chain expects multiple inputs, they can be passed in
directly as keyword arguments.
Returns
The chain output.
Example
# Suppose we have a single-input chain that takes a 'question' string:
chain.run("What's the temperature in Boise, Idaho?")
# -> "The temperature in Boise is..."
# Suppose we have a multi-input chain that takes a 'question' string
# and 'context' string:
question = "What's the temperature in Boise, Idaho?"
context = "Weather report for Boise, Idaho on 07/03/23..."
chain.run(question=question, context=context)
# -> "The temperature in Boise is..."
save(file_path: Union[Path, str]) → None¶
Save the chain.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.comparison.eval_chain.LabeledPairwiseStringEvalChain.html
|
e02b7b494139-11
|
save(file_path: Union[Path, str]) → None¶
Save the chain.
Expects Chain._chain_type property to be implemented and for memory to benull.
Parameters
file_path – Path to file to save the chain to.
Example
chain.save(file_path="path/chain.yaml")
classmethod schema(by_alias: bool = True, ref_template: unicode = '#/definitions/{model}') → DictStrAny¶
classmethod schema_json(*, by_alias: bool = True, ref_template: unicode = '#/definitions/{model}', **dumps_kwargs: Any) → unicode¶
stream(input: Input, config: Optional[RunnableConfig] = None) → Iterator[Output]¶
to_json() → Union[SerializedConstructor, SerializedNotImplemented]¶
to_json_not_implemented() → SerializedNotImplemented¶
classmethod update_forward_refs(**localns: Any) → None¶
Try to update ForwardRefs on fields based on this Model, globalns and localns.
classmethod validate(value: Any) → Model¶
with_fallbacks(fallbacks: ~typing.Sequence[~langchain.schema.runnable.Runnable[~langchain.schema.runnable.Input, ~langchain.schema.runnable.Output]], *, exceptions_to_handle: ~typing.Tuple[~typing.Type[BaseException]] = (<class 'Exception'>,)) → RunnableWithFallbacks[Input, Output]¶
property lc_attributes: Dict¶
Return a list of attribute names that should be included in the
serialized kwargs. These attributes must be accepted by the
constructor.
property lc_namespace: List[str]¶
Return the namespace of the langchain object.
eg. [“langchain”, “llms”, “openai”]
property lc_secrets: Dict[str, str]¶
Return a map of constructor argument names to secret ids.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.comparison.eval_chain.LabeledPairwiseStringEvalChain.html
|
e02b7b494139-12
|
Return a map of constructor argument names to secret ids.
eg. {“openai_api_key”: “OPENAI_API_KEY”}
property lc_serializable: bool¶
Return whether or not the class is serializable.
property requires_input: bool¶
Return whether the chain requires an input.
Returns
True if the chain requires an input, False otherwise.
Return type
bool
property requires_reference: bool¶
Return whether the chain requires a reference.
Returns
True if the chain requires a reference, False otherwise.
Return type
bool
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.comparison.eval_chain.LabeledPairwiseStringEvalChain.html
|
dd7471652bf5-0
|
langchain.evaluation.qa.eval_chain.CotQAEvalChain¶
class langchain.evaluation.qa.eval_chain.CotQAEvalChain[source]¶
Bases: ContextQAEvalChain
LLM Chain for evaluating QA using chain of thought reasoning.
param callback_manager: Optional[BaseCallbackManager] = None¶
Deprecated, use callbacks instead.
param callbacks: Callbacks = None¶
Optional list of callback handlers (or callback manager). Defaults to None.
Callback handlers are called throughout the lifecycle of a call to a chain,
starting with on_chain_start, ending with on_chain_end or on_chain_error.
Each custom chain can optionally call additional callback methods, see Callback docs
for full details.
param llm: BaseLanguageModel [Required]¶
Language model to call.
param llm_kwargs: dict [Optional]¶
param memory: Optional[BaseMemory] = None¶
Optional memory object. Defaults to None.
Memory is a class that gets called at the start
and at the end of every chain. At the start, memory loads variables and passes
them along in the chain. At the end, it saves any returned variables.
There are many different types of memory - please see memory docs
for the full catalog.
param metadata: Optional[Dict[str, Any]] = None¶
Optional metadata associated with the chain. Defaults to None.
This metadata will be associated with each call to this chain,
and passed as arguments to the handlers defined in callbacks.
You can use these to eg identify a specific instance of a chain with its use case.
param output_key: str = 'text'¶
param output_parser: BaseLLMOutputParser [Optional]¶
Output parser to use.
Defaults to one that takes the most likely string but does not change it
otherwise.
param prompt: BasePromptTemplate [Required]¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.CotQAEvalChain.html
|
dd7471652bf5-1
|
otherwise.
param prompt: BasePromptTemplate [Required]¶
Prompt object to use.
param return_final_only: bool = True¶
Whether to return only the final parsed result. Defaults to True.
If false, will return a bunch of extra information about the generation.
param tags: Optional[List[str]] = None¶
Optional list of tags associated with the chain. Defaults to None.
These tags will be associated with each call to this chain,
and passed as arguments to the handlers defined in callbacks.
You can use these to eg identify a specific instance of a chain with its use case.
param verbose: bool [Optional]¶
Whether or not run in verbose mode. In verbose mode, some intermediate logs
will be printed to the console. Defaults to langchain.verbose value.
__call__(inputs: Union[Dict[str, Any], Any], return_only_outputs: bool = False, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, *, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, include_run_info: bool = False) → Dict[str, Any]¶
Execute the chain.
Parameters
inputs – Dictionary of inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
return_only_outputs – Whether to return only outputs in the
response. If True, only new keys generated by this chain will be
returned. If False, both input keys and new keys generated by this
chain will be returned. Defaults to False.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.CotQAEvalChain.html
|
dd7471652bf5-2
|
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
metadata – Optional metadata associated with the chain. Defaults to None
include_run_info – Whether to include run info in the response. Defaults
to False.
Returns
A dict of named outputs. Should contain all outputs specified inChain.output_keys.
async aapply(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → List[Dict[str, str]]¶
Utilize the LLM generate method for speed gains.
async aapply_and_parse(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → Sequence[Union[str, List[str], Dict[str, str]]]¶
Call apply and then parse the results.
async abatch(inputs: List[Input], config: Optional[Union[RunnableConfig, List[RunnableConfig]]] = None, *, max_concurrency: Optional[int] = None) → List[Output]¶
async acall(inputs: Union[Dict[str, Any], Any], return_only_outputs: bool = False, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, *, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, include_run_info: bool = False) → Dict[str, Any]¶
Asynchronously execute the chain.
Parameters
inputs – Dictionary of inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.CotQAEvalChain.html
|
dd7471652bf5-3
|
Chain.input_keys except for inputs that will be set by the chain’s
memory.
return_only_outputs – Whether to return only outputs in the
response. If True, only new keys generated by this chain will be
returned. If False, both input keys and new keys generated by this
chain will be returned. Defaults to False.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
metadata – Optional metadata associated with the chain. Defaults to None
include_run_info – Whether to include run info in the response. Defaults
to False.
Returns
A dict of named outputs. Should contain all outputs specified inChain.output_keys.
async aevaluate_strings(*, prediction: str, reference: Optional[str] = None, input: Optional[str] = None, **kwargs: Any) → dict¶
Asynchronously evaluate Chain or LLM output, based on optional input and label.
Parameters
prediction (str) – The LLM or chain prediction to evaluate.
reference (Optional[str], optional) – The reference label to evaluate against.
input (Optional[str], optional) – The input to consider during evaluation.
**kwargs – Additional keyword arguments, including callbacks, tags, etc.
Returns
The evaluation results containing the score or value.
Return type
dict
async agenerate(input_list: List[Dict[str, Any]], run_manager: Optional[AsyncCallbackManagerForChainRun] = None) → LLMResult¶
Generate LLM result from inputs.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.CotQAEvalChain.html
|
dd7471652bf5-4
|
Generate LLM result from inputs.
async ainvoke(input: Dict[str, Any], config: Optional[RunnableConfig] = None) → Dict[str, Any]¶
apply(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → List[Dict[str, str]]¶
Utilize the LLM generate method for speed gains.
apply_and_parse(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → Sequence[Union[str, List[str], Dict[str, str]]]¶
Call apply and then parse the results.
async apredict(callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, **kwargs: Any) → str¶
Format prompt with kwargs and pass to LLM.
Parameters
callbacks – Callbacks to pass to LLMChain
**kwargs – Keys to pass to prompt template.
Returns
Completion from LLM.
Example
completion = llm.predict(adjective="funny")
async apredict_and_parse(callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, **kwargs: Any) → Union[str, List[str], Dict[str, str]]¶
Call apredict and then parse the results.
async aprep_prompts(input_list: List[Dict[str, Any]], run_manager: Optional[AsyncCallbackManagerForChainRun] = None) → Tuple[List[PromptValue], Optional[List[str]]]¶
Prepare prompts from inputs.
async arun(*args: Any, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, **kwargs: Any) → Any¶
Convenience method for executing chain.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.CotQAEvalChain.html
|
dd7471652bf5-5
|
Convenience method for executing chain.
The main difference between this method and Chain.__call__ is that this
method expects inputs to be passed directly in as positional arguments or
keyword arguments, whereas Chain.__call__ expects a single input dictionary
with all the inputs
Parameters
*args – If the chain expects a single input, it can be passed in as the
sole positional argument.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
**kwargs – If the chain expects multiple inputs, they can be passed in
directly as keyword arguments.
Returns
The chain output.
Example
# Suppose we have a single-input chain that takes a 'question' string:
await chain.arun("What's the temperature in Boise, Idaho?")
# -> "The temperature in Boise is..."
# Suppose we have a multi-input chain that takes a 'question' string
# and 'context' string:
question = "What's the temperature in Boise, Idaho?"
context = "Weather report for Boise, Idaho on 07/03/23..."
await chain.arun(question=question, context=context)
# -> "The temperature in Boise is..."
async astream(input: Input, config: Optional[RunnableConfig] = None) → AsyncIterator[Output]¶
batch(inputs: List[Input], config: Optional[Union[RunnableConfig, List[RunnableConfig]]] = None, *, max_concurrency: Optional[int] = None) → List[Output]¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.CotQAEvalChain.html
|
dd7471652bf5-6
|
bind(**kwargs: Any) → Runnable[Input, Output]¶
Bind arguments to a Runnable, returning a new Runnable.
classmethod construct(_fields_set: Optional[SetStr] = None, **values: Any) → Model¶
Creates a new model setting __dict__ and __fields_set__ from trusted or pre-validated data.
Default values are respected, but no other validation is performed.
Behaves as if Config.extra = ‘allow’ was set since it adds all passed values
copy(*, include: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, exclude: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, update: Optional[DictStrAny] = None, deep: bool = False) → Model¶
Duplicate a model, optionally choose which fields to include, exclude and change.
Parameters
include – fields to include in new model
exclude – fields to exclude from new model, as with values this takes precedence over include
update – values to change/add in the new model. Note: the data is not validated before creating
the new model: you should trust this data
deep – set to True to make a deep copy of the model
Returns
new model instance
create_outputs(llm_result: LLMResult) → List[Dict[str, Any]]¶
Create outputs from response.
dict(**kwargs: Any) → Dict¶
Dictionary representation of chain.
Expects Chain._chain_type property to be implemented and for memory to benull.
Parameters
**kwargs – Keyword arguments passed to default pydantic.BaseModel.dict
method.
Returns
A dictionary representation of the chain.
Example
..code-block:: python
chain.dict(exclude_unset=True)
# -> {“_type”: “foo”, “verbose”: False, …}
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.CotQAEvalChain.html
|
dd7471652bf5-7
|
# -> {“_type”: “foo”, “verbose”: False, …}
evaluate(examples: List[dict], predictions: List[dict], question_key: str = 'query', context_key: str = 'context', prediction_key: str = 'result', *, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → List[dict]¶
Evaluate question answering examples and predictions.
evaluate_strings(*, prediction: str, reference: Optional[str] = None, input: Optional[str] = None, **kwargs: Any) → dict¶
Evaluate Chain or LLM output, based on optional input and label.
Parameters
prediction (str) – The LLM or chain prediction to evaluate.
reference (Optional[str], optional) – The reference label to evaluate against.
input (Optional[str], optional) – The input to consider during evaluation.
**kwargs – Additional keyword arguments, including callbacks, tags, etc.
Returns
The evaluation results containing the score or value.
Return type
dict
classmethod from_llm(llm: BaseLanguageModel, prompt: Optional[PromptTemplate] = None, **kwargs: Any) → CotQAEvalChain[source]¶
Load QA Eval Chain from LLM.
classmethod from_orm(obj: Any) → Model¶
classmethod from_string(llm: BaseLanguageModel, template: str) → LLMChain¶
Create LLMChain from LLM and template.
generate(input_list: List[Dict[str, Any]], run_manager: Optional[CallbackManagerForChainRun] = None) → LLMResult¶
Generate LLM result from inputs.
invoke(input: Dict[str, Any], config: Optional[RunnableConfig] = None) → Dict[str, Any]¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.CotQAEvalChain.html
|
dd7471652bf5-8
|
json(*, include: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, exclude: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, by_alias: bool = False, skip_defaults: Optional[bool] = None, exclude_unset: bool = False, exclude_defaults: bool = False, exclude_none: bool = False, encoder: Optional[Callable[[Any], Any]] = None, models_as_dict: bool = True, **dumps_kwargs: Any) → unicode¶
Generate a JSON representation of the model, include and exclude arguments as per dict().
encoder is an optional function to supply as default to json.dumps(), other arguments as per json.dumps().
classmethod parse_file(path: Union[str, Path], *, content_type: unicode = None, encoding: unicode = 'utf8', proto: Protocol = None, allow_pickle: bool = False) → Model¶
classmethod parse_obj(obj: Any) → Model¶
classmethod parse_raw(b: Union[str, bytes], *, content_type: unicode = None, encoding: unicode = 'utf8', proto: Protocol = None, allow_pickle: bool = False) → Model¶
predict(callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, **kwargs: Any) → str¶
Format prompt with kwargs and pass to LLM.
Parameters
callbacks – Callbacks to pass to LLMChain
**kwargs – Keys to pass to prompt template.
Returns
Completion from LLM.
Example
completion = llm.predict(adjective="funny")
predict_and_parse(callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, **kwargs: Any) → Union[str, List[str], Dict[str, Any]]¶
Call predict and then parse the results.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.CotQAEvalChain.html
|
dd7471652bf5-9
|
Call predict and then parse the results.
prep_inputs(inputs: Union[Dict[str, Any], Any]) → Dict[str, str]¶
Validate and prepare chain inputs, including adding inputs from memory.
Parameters
inputs – Dictionary of raw inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
Returns
A dictionary of all inputs, including those added by the chain’s memory.
prep_outputs(inputs: Dict[str, str], outputs: Dict[str, str], return_only_outputs: bool = False) → Dict[str, str]¶
Validate and prepare chain outputs, and save info about this run to memory.
Parameters
inputs – Dictionary of chain inputs, including any inputs added by chain
memory.
outputs – Dictionary of initial chain outputs.
return_only_outputs – Whether to only return the chain outputs. If False,
inputs are also added to the final outputs.
Returns
A dict of the final chain outputs.
prep_prompts(input_list: List[Dict[str, Any]], run_manager: Optional[CallbackManagerForChainRun] = None) → Tuple[List[PromptValue], Optional[List[str]]]¶
Prepare prompts from inputs.
run(*args: Any, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, **kwargs: Any) → Any¶
Convenience method for executing chain.
The main difference between this method and Chain.__call__ is that this
method expects inputs to be passed directly in as positional arguments or
keyword arguments, whereas Chain.__call__ expects a single input dictionary
with all the inputs
Parameters
*args – If the chain expects a single input, it can be passed in as the
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.CotQAEvalChain.html
|
dd7471652bf5-10
|
*args – If the chain expects a single input, it can be passed in as the
sole positional argument.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
**kwargs – If the chain expects multiple inputs, they can be passed in
directly as keyword arguments.
Returns
The chain output.
Example
# Suppose we have a single-input chain that takes a 'question' string:
chain.run("What's the temperature in Boise, Idaho?")
# -> "The temperature in Boise is..."
# Suppose we have a multi-input chain that takes a 'question' string
# and 'context' string:
question = "What's the temperature in Boise, Idaho?"
context = "Weather report for Boise, Idaho on 07/03/23..."
chain.run(question=question, context=context)
# -> "The temperature in Boise is..."
save(file_path: Union[Path, str]) → None¶
Save the chain.
Expects Chain._chain_type property to be implemented and for memory to benull.
Parameters
file_path – Path to file to save the chain to.
Example
chain.save(file_path="path/chain.yaml")
classmethod schema(by_alias: bool = True, ref_template: unicode = '#/definitions/{model}') → DictStrAny¶
classmethod schema_json(*, by_alias: bool = True, ref_template: unicode = '#/definitions/{model}', **dumps_kwargs: Any) → unicode¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.CotQAEvalChain.html
|
dd7471652bf5-11
|
stream(input: Input, config: Optional[RunnableConfig] = None) → Iterator[Output]¶
to_json() → Union[SerializedConstructor, SerializedNotImplemented]¶
to_json_not_implemented() → SerializedNotImplemented¶
classmethod update_forward_refs(**localns: Any) → None¶
Try to update ForwardRefs on fields based on this Model, globalns and localns.
classmethod validate(value: Any) → Model¶
with_fallbacks(fallbacks: ~typing.Sequence[~langchain.schema.runnable.Runnable[~langchain.schema.runnable.Input, ~langchain.schema.runnable.Output]], *, exceptions_to_handle: ~typing.Tuple[~typing.Type[BaseException]] = (<class 'Exception'>,)) → RunnableWithFallbacks[Input, Output]¶
property evaluation_name: str¶
The name of the evaluation.
property lc_attributes: Dict¶
Return a list of attribute names that should be included in the
serialized kwargs. These attributes must be accepted by the
constructor.
property lc_namespace: List[str]¶
Return the namespace of the langchain object.
eg. [“langchain”, “llms”, “openai”]
property lc_secrets: Dict[str, str]¶
Return a map of constructor argument names to secret ids.
eg. {“openai_api_key”: “OPENAI_API_KEY”}
property lc_serializable: bool¶
Return whether or not the class is serializable.
property requires_input: bool¶
Whether the chain requires an input string.
property requires_reference: bool¶
Whether the chain requires a reference string.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.CotQAEvalChain.html
|
b0b846b9b799-0
|
langchain.evaluation.loading.load_dataset¶
langchain.evaluation.loading.load_dataset(uri: str) → List[Dict][source]¶
Load a dataset from the LangChainDatasets HuggingFace org.
Parameters
uri – The uri of the dataset to load.
Returns
A list of dictionaries, each representing a row in the dataset.
Prerequisites
pip install datasets
Examples
from langchain.evaluation import load_dataset
ds = load_dataset("llm-math")
Examples using load_dataset¶
Question Answering Benchmarking: State of the Union Address
Question Answering Benchmarking: Paul Graham Essay
Evaluating an OpenAPI Chain
Comparing Chain Outputs
SQL Question Answering Benchmarking: Chinook
Agent VectorDB Question Answering Benchmarking
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.loading.load_dataset.html
|
185db86f559c-0
|
langchain.evaluation.schema.AgentTrajectoryEvaluator¶
class langchain.evaluation.schema.AgentTrajectoryEvaluator[source]¶
Interface for evaluating agent trajectories.
Attributes
requires_input
Whether this evaluator requires an input string.
requires_reference
Whether this evaluator requires a reference label.
Methods
__init__()
aevaluate_agent_trajectory(*, prediction, ...)
Asynchronously evaluate a trajectory.
evaluate_agent_trajectory(*, prediction, ...)
Evaluate a trajectory.
__init__()¶
async aevaluate_agent_trajectory(*, prediction: str, agent_trajectory: Sequence[Tuple[AgentAction, str]], input: str, reference: Optional[str] = None, **kwargs: Any) → dict[source]¶
Asynchronously evaluate a trajectory.
Parameters
prediction (str) – The final predicted response.
agent_trajectory (List[Tuple[AgentAction, str]]) – The intermediate steps forming the agent trajectory.
input (str) – The input to the agent.
reference (Optional[str]) – The reference answer.
Returns
The evaluation result.
Return type
dict
evaluate_agent_trajectory(*, prediction: str, agent_trajectory: Sequence[Tuple[AgentAction, str]], input: str, reference: Optional[str] = None, **kwargs: Any) → dict[source]¶
Evaluate a trajectory.
Parameters
prediction (str) – The final predicted response.
agent_trajectory (List[Tuple[AgentAction, str]]) – The intermediate steps forming the agent trajectory.
input (str) – The input to the agent.
reference (Optional[str]) – The reference answer.
Returns
The evaluation result.
Return type
dict
Examples using AgentTrajectoryEvaluator¶
Custom Trajectory Evaluator
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.schema.AgentTrajectoryEvaluator.html
|
c38e00f542f3-0
|
langchain.evaluation.loading.load_evaluator¶
langchain.evaluation.loading.load_evaluator(evaluator: EvaluatorType, *, llm: Optional[BaseLanguageModel] = None, **kwargs: Any) → Chain[source]¶
Load the requested evaluation chain specified by a string.
Parameters
evaluator (EvaluatorType) – The type of evaluator to load.
llm (BaseLanguageModel, optional) – The language model to use for evaluation, by default None
**kwargs (Any) – Additional keyword arguments to pass to the evaluator.
Returns
The loaded evaluation chain.
Return type
Chain
Examples
>>> from langchain.evaluation import load_evaluator, EvaluatorType
>>> evaluator = load_evaluator(EvaluatorType.QA)
Examples using load_evaluator¶
Comparing Chain Outputs
Agent Trajectory
Pairwise Embedding Distance
Pairwise String Comparison
Criteria Evaluation
QA Correctness
String Distance
Embedding Distance
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.loading.load_evaluator.html
|
ac9943e02bb3-0
|
langchain.evaluation.agents.trajectory_eval_chain.TrajectoryEval¶
class langchain.evaluation.agents.trajectory_eval_chain.TrajectoryEval[source]¶
A named tuple containing the score and reasoning for a trajectory.
score: float¶
The score for the trajectory, normalized from 0 to 1.
reasoning: str¶
The reasoning for the score.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.agents.trajectory_eval_chain.TrajectoryEval.html
|
a72d448a20e3-0
|
langchain.evaluation.embedding_distance.base.EmbeddingDistance¶
class langchain.evaluation.embedding_distance.base.EmbeddingDistance(value, names=None, *, module=None, qualname=None, type=None, start=1, boundary=None)[source]¶
Embedding Distance Metric.
COSINE¶
Cosine distance metric.
EUCLIDEAN¶
Euclidean distance metric.
MANHATTAN¶
Manhattan distance metric.
CHEBYSHEV¶
Chebyshev distance metric.
HAMMING¶
Hamming distance metric.
COSINE = 'cosine'¶
EUCLIDEAN = 'euclidean'¶
MANHATTAN = 'manhattan'¶
CHEBYSHEV = 'chebyshev'¶
HAMMING = 'hamming'¶
Examples using EmbeddingDistance¶
Pairwise Embedding Distance
Embedding Distance
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.embedding_distance.base.EmbeddingDistance.html
|
b561e3b25fec-0
|
langchain.evaluation.criteria.eval_chain.Criteria¶
class langchain.evaluation.criteria.eval_chain.Criteria(value, names=None, *, module=None, qualname=None, type=None, start=1, boundary=None)[source]¶
A Criteria to evaluate.
CONCISENESS = 'conciseness'¶
RELEVANCE = 'relevance'¶
CORRECTNESS = 'correctness'¶
COHERENCE = 'coherence'¶
HARMFULNESS = 'harmfulness'¶
MALICIOUSNESS = 'maliciousness'¶
HELPFULNESS = 'helpfulness'¶
CONTROVERSIALITY = 'controversiality'¶
MISOGYNY = 'misogyny'¶
CRIMINALITY = 'criminality'¶
INSENSITIVITY = 'insensitivity'¶
DEPTH = 'depth'¶
CREATIVITY = 'creativity'¶
DETAIL = 'detail'¶
Examples using Criteria¶
Criteria Evaluation
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.criteria.eval_chain.Criteria.html
|
6d135c7b8837-0
|
langchain.evaluation.string_distance.base.StringDistance¶
class langchain.evaluation.string_distance.base.StringDistance(value, names=None, *, module=None, qualname=None, type=None, start=1, boundary=None)[source]¶
Distance metric to use.
DAMERAU_LEVENSHTEIN¶
The Damerau-Levenshtein distance.
LEVENSHTEIN¶
The Levenshtein distance.
JARO¶
The Jaro distance.
JARO_WINKLER¶
The Jaro-Winkler distance.
HAMMING¶
The Hamming distance.
INDEL¶
The Indel distance.
DAMERAU_LEVENSHTEIN = 'damerau_levenshtein'¶
LEVENSHTEIN = 'levenshtein'¶
JARO = 'jaro'¶
JARO_WINKLER = 'jaro_winkler'¶
HAMMING = 'hamming'¶
INDEL = 'indel'¶
Examples using StringDistance¶
String Distance
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.string_distance.base.StringDistance.html
|
4fe041a40323-0
|
langchain.evaluation.embedding_distance.base.PairwiseEmbeddingDistanceEvalChain¶
class langchain.evaluation.embedding_distance.base.PairwiseEmbeddingDistanceEvalChain[source]¶
Bases: _EmbeddingDistanceChainMixin, PairwiseStringEvaluator
Use embedding distances to score semantic difference between two predictions.
Examples:
>>> chain = PairwiseEmbeddingDistanceEvalChain()
>>> result = chain.evaluate_string_pairs(prediction=”Hello”, prediction_b=”Hi”)
>>> print(result)
{‘score’: 0.5}
Create a new model by parsing and validating input data from keyword arguments.
Raises ValidationError if the input data cannot be parsed to form a valid model.
param callback_manager: Optional[BaseCallbackManager] = None¶
Deprecated, use callbacks instead.
param callbacks: Callbacks = None¶
Optional list of callback handlers (or callback manager). Defaults to None.
Callback handlers are called throughout the lifecycle of a call to a chain,
starting with on_chain_start, ending with on_chain_end or on_chain_error.
Each custom chain can optionally call additional callback methods, see Callback docs
for full details.
param distance_metric: langchain.evaluation.embedding_distance.base.EmbeddingDistance = EmbeddingDistance.COSINE¶
param embeddings: langchain.embeddings.base.Embeddings [Optional]¶
param memory: Optional[BaseMemory] = None¶
Optional memory object. Defaults to None.
Memory is a class that gets called at the start
and at the end of every chain. At the start, memory loads variables and passes
them along in the chain. At the end, it saves any returned variables.
There are many different types of memory - please see memory docs
for the full catalog.
param metadata: Optional[Dict[str, Any]] = None¶
Optional metadata associated with the chain. Defaults to None.
This metadata will be associated with each call to this chain,
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.embedding_distance.base.PairwiseEmbeddingDistanceEvalChain.html
|
4fe041a40323-1
|
This metadata will be associated with each call to this chain,
and passed as arguments to the handlers defined in callbacks.
You can use these to eg identify a specific instance of a chain with its use case.
param tags: Optional[List[str]] = None¶
Optional list of tags associated with the chain. Defaults to None.
These tags will be associated with each call to this chain,
and passed as arguments to the handlers defined in callbacks.
You can use these to eg identify a specific instance of a chain with its use case.
param verbose: bool [Optional]¶
Whether or not run in verbose mode. In verbose mode, some intermediate logs
will be printed to the console. Defaults to langchain.verbose value.
__call__(inputs: Union[Dict[str, Any], Any], return_only_outputs: bool = False, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, *, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, include_run_info: bool = False) → Dict[str, Any]¶
Execute the chain.
Parameters
inputs – Dictionary of inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
return_only_outputs – Whether to return only outputs in the
response. If True, only new keys generated by this chain will be
returned. If False, both input keys and new keys generated by this
chain will be returned. Defaults to False.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.embedding_distance.base.PairwiseEmbeddingDistanceEvalChain.html
|
4fe041a40323-2
|
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
metadata – Optional metadata associated with the chain. Defaults to None
include_run_info – Whether to include run info in the response. Defaults
to False.
Returns
A dict of named outputs. Should contain all outputs specified inChain.output_keys.
async abatch(inputs: List[Input], config: Optional[Union[RunnableConfig, List[RunnableConfig]]] = None, *, max_concurrency: Optional[int] = None) → List[Output]¶
async acall(inputs: Union[Dict[str, Any], Any], return_only_outputs: bool = False, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, *, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, include_run_info: bool = False) → Dict[str, Any]¶
Asynchronously execute the chain.
Parameters
inputs – Dictionary of inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
return_only_outputs – Whether to return only outputs in the
response. If True, only new keys generated by this chain will be
returned. If False, both input keys and new keys generated by this
chain will be returned. Defaults to False.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.embedding_distance.base.PairwiseEmbeddingDistanceEvalChain.html
|
4fe041a40323-3
|
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
metadata – Optional metadata associated with the chain. Defaults to None
include_run_info – Whether to include run info in the response. Defaults
to False.
Returns
A dict of named outputs. Should contain all outputs specified inChain.output_keys.
async aevaluate_string_pairs(*, prediction: str, prediction_b: str, reference: Optional[str] = None, input: Optional[str] = None, **kwargs: Any) → dict¶
Asynchronously evaluate the output string pairs.
Parameters
prediction (str) – The output string from the first model.
prediction_b (str) – The output string from the second model.
reference (Optional[str], optional) – The expected output / reference string.
input (Optional[str], optional) – The input string.
**kwargs – Additional keyword arguments, such as callbacks and optional reference strings.
Returns
A dictionary containing the preference, scores, and/or other information.
Return type
dict
async ainvoke(input: Dict[str, Any], config: Optional[RunnableConfig] = None) → Dict[str, Any]¶
apply(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → List[Dict[str, str]]¶
Call the chain on all inputs in the list.
async arun(*args: Any, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, **kwargs: Any) → Any¶
Convenience method for executing chain.
The main difference between this method and Chain.__call__ is that this
method expects inputs to be passed directly in as positional arguments or
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.embedding_distance.base.PairwiseEmbeddingDistanceEvalChain.html
|
4fe041a40323-4
|
method expects inputs to be passed directly in as positional arguments or
keyword arguments, whereas Chain.__call__ expects a single input dictionary
with all the inputs
Parameters
*args – If the chain expects a single input, it can be passed in as the
sole positional argument.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
**kwargs – If the chain expects multiple inputs, they can be passed in
directly as keyword arguments.
Returns
The chain output.
Example
# Suppose we have a single-input chain that takes a 'question' string:
await chain.arun("What's the temperature in Boise, Idaho?")
# -> "The temperature in Boise is..."
# Suppose we have a multi-input chain that takes a 'question' string
# and 'context' string:
question = "What's the temperature in Boise, Idaho?"
context = "Weather report for Boise, Idaho on 07/03/23..."
await chain.arun(question=question, context=context)
# -> "The temperature in Boise is..."
async astream(input: Input, config: Optional[RunnableConfig] = None) → AsyncIterator[Output]¶
batch(inputs: List[Input], config: Optional[Union[RunnableConfig, List[RunnableConfig]]] = None, *, max_concurrency: Optional[int] = None) → List[Output]¶
bind(**kwargs: Any) → Runnable[Input, Output]¶
Bind arguments to a Runnable, returning a new Runnable.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.embedding_distance.base.PairwiseEmbeddingDistanceEvalChain.html
|
4fe041a40323-5
|
Bind arguments to a Runnable, returning a new Runnable.
classmethod construct(_fields_set: Optional[SetStr] = None, **values: Any) → Model¶
Creates a new model setting __dict__ and __fields_set__ from trusted or pre-validated data.
Default values are respected, but no other validation is performed.
Behaves as if Config.extra = ‘allow’ was set since it adds all passed values
copy(*, include: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, exclude: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, update: Optional[DictStrAny] = None, deep: bool = False) → Model¶
Duplicate a model, optionally choose which fields to include, exclude and change.
Parameters
include – fields to include in new model
exclude – fields to exclude from new model, as with values this takes precedence over include
update – values to change/add in the new model. Note: the data is not validated before creating
the new model: you should trust this data
deep – set to True to make a deep copy of the model
Returns
new model instance
dict(**kwargs: Any) → Dict¶
Dictionary representation of chain.
Expects Chain._chain_type property to be implemented and for memory to benull.
Parameters
**kwargs – Keyword arguments passed to default pydantic.BaseModel.dict
method.
Returns
A dictionary representation of the chain.
Example
..code-block:: python
chain.dict(exclude_unset=True)
# -> {“_type”: “foo”, “verbose”: False, …}
evaluate_string_pairs(*, prediction: str, prediction_b: str, reference: Optional[str] = None, input: Optional[str] = None, **kwargs: Any) → dict¶
Evaluate the output string pairs.
Parameters
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.embedding_distance.base.PairwiseEmbeddingDistanceEvalChain.html
|
4fe041a40323-6
|
Evaluate the output string pairs.
Parameters
prediction (str) – The output string from the first model.
prediction_b (str) – The output string from the second model.
reference (Optional[str], optional) – The expected output / reference string.
input (Optional[str], optional) – The input string.
**kwargs – Additional keyword arguments, such as callbacks and optional reference strings.
Returns
A dictionary containing the preference, scores, and/or other information.
Return type
dict
classmethod from_orm(obj: Any) → Model¶
invoke(input: Dict[str, Any], config: Optional[RunnableConfig] = None) → Dict[str, Any]¶
json(*, include: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, exclude: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, by_alias: bool = False, skip_defaults: Optional[bool] = None, exclude_unset: bool = False, exclude_defaults: bool = False, exclude_none: bool = False, encoder: Optional[Callable[[Any], Any]] = None, models_as_dict: bool = True, **dumps_kwargs: Any) → unicode¶
Generate a JSON representation of the model, include and exclude arguments as per dict().
encoder is an optional function to supply as default to json.dumps(), other arguments as per json.dumps().
classmethod parse_file(path: Union[str, Path], *, content_type: unicode = None, encoding: unicode = 'utf8', proto: Protocol = None, allow_pickle: bool = False) → Model¶
classmethod parse_obj(obj: Any) → Model¶
classmethod parse_raw(b: Union[str, bytes], *, content_type: unicode = None, encoding: unicode = 'utf8', proto: Protocol = None, allow_pickle: bool = False) → Model¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.embedding_distance.base.PairwiseEmbeddingDistanceEvalChain.html
|
4fe041a40323-7
|
prep_inputs(inputs: Union[Dict[str, Any], Any]) → Dict[str, str]¶
Validate and prepare chain inputs, including adding inputs from memory.
Parameters
inputs – Dictionary of raw inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
Returns
A dictionary of all inputs, including those added by the chain’s memory.
prep_outputs(inputs: Dict[str, str], outputs: Dict[str, str], return_only_outputs: bool = False) → Dict[str, str]¶
Validate and prepare chain outputs, and save info about this run to memory.
Parameters
inputs – Dictionary of chain inputs, including any inputs added by chain
memory.
outputs – Dictionary of initial chain outputs.
return_only_outputs – Whether to only return the chain outputs. If False,
inputs are also added to the final outputs.
Returns
A dict of the final chain outputs.
run(*args: Any, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, **kwargs: Any) → Any¶
Convenience method for executing chain.
The main difference between this method and Chain.__call__ is that this
method expects inputs to be passed directly in as positional arguments or
keyword arguments, whereas Chain.__call__ expects a single input dictionary
with all the inputs
Parameters
*args – If the chain expects a single input, it can be passed in as the
sole positional argument.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.embedding_distance.base.PairwiseEmbeddingDistanceEvalChain.html
|
4fe041a40323-8
|
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
**kwargs – If the chain expects multiple inputs, they can be passed in
directly as keyword arguments.
Returns
The chain output.
Example
# Suppose we have a single-input chain that takes a 'question' string:
chain.run("What's the temperature in Boise, Idaho?")
# -> "The temperature in Boise is..."
# Suppose we have a multi-input chain that takes a 'question' string
# and 'context' string:
question = "What's the temperature in Boise, Idaho?"
context = "Weather report for Boise, Idaho on 07/03/23..."
chain.run(question=question, context=context)
# -> "The temperature in Boise is..."
save(file_path: Union[Path, str]) → None¶
Save the chain.
Expects Chain._chain_type property to be implemented and for memory to benull.
Parameters
file_path – Path to file to save the chain to.
Example
chain.save(file_path="path/chain.yaml")
classmethod schema(by_alias: bool = True, ref_template: unicode = '#/definitions/{model}') → DictStrAny¶
classmethod schema_json(*, by_alias: bool = True, ref_template: unicode = '#/definitions/{model}', **dumps_kwargs: Any) → unicode¶
stream(input: Input, config: Optional[RunnableConfig] = None) → Iterator[Output]¶
to_json() → Union[SerializedConstructor, SerializedNotImplemented]¶
to_json_not_implemented() → SerializedNotImplemented¶
classmethod update_forward_refs(**localns: Any) → None¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.embedding_distance.base.PairwiseEmbeddingDistanceEvalChain.html
|
4fe041a40323-9
|
classmethod update_forward_refs(**localns: Any) → None¶
Try to update ForwardRefs on fields based on this Model, globalns and localns.
classmethod validate(value: Any) → Model¶
with_fallbacks(fallbacks: ~typing.Sequence[~langchain.schema.runnable.Runnable[~langchain.schema.runnable.Input, ~langchain.schema.runnable.Output]], *, exceptions_to_handle: ~typing.Tuple[~typing.Type[BaseException]] = (<class 'Exception'>,)) → RunnableWithFallbacks[Input, Output]¶
property evaluation_name: str¶
property input_keys: List[str]¶
Return the input keys of the chain.
Returns
The input keys.
Return type
List[str]
property lc_attributes: Dict¶
Return a list of attribute names that should be included in the
serialized kwargs. These attributes must be accepted by the
constructor.
property lc_namespace: List[str]¶
Return the namespace of the langchain object.
eg. [“langchain”, “llms”, “openai”]
property lc_secrets: Dict[str, str]¶
Return a map of constructor argument names to secret ids.
eg. {“openai_api_key”: “OPENAI_API_KEY”}
property lc_serializable: bool¶
Return whether or not the class is serializable.
property output_keys: List[str]¶
Return the output keys of the chain.
Returns
The output keys.
Return type
List[str]
property requires_input: bool¶
Whether this evaluator requires an input string.
property requires_reference: bool¶
Whether this evaluator requires a reference label.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.embedding_distance.base.PairwiseEmbeddingDistanceEvalChain.html
|
ee9315a0d90d-0
|
langchain.evaluation.qa.eval_chain.QAEvalChain¶
class langchain.evaluation.qa.eval_chain.QAEvalChain[source]¶
Bases: LLMChain, StringEvaluator, LLMEvalChain
LLM Chain for evaluating question answering.
param callback_manager: Optional[BaseCallbackManager] = None¶
Deprecated, use callbacks instead.
param callbacks: Callbacks = None¶
Optional list of callback handlers (or callback manager). Defaults to None.
Callback handlers are called throughout the lifecycle of a call to a chain,
starting with on_chain_start, ending with on_chain_end or on_chain_error.
Each custom chain can optionally call additional callback methods, see Callback docs
for full details.
param llm: BaseLanguageModel [Required]¶
Language model to call.
param llm_kwargs: dict [Optional]¶
param memory: Optional[BaseMemory] = None¶
Optional memory object. Defaults to None.
Memory is a class that gets called at the start
and at the end of every chain. At the start, memory loads variables and passes
them along in the chain. At the end, it saves any returned variables.
There are many different types of memory - please see memory docs
for the full catalog.
param metadata: Optional[Dict[str, Any]] = None¶
Optional metadata associated with the chain. Defaults to None.
This metadata will be associated with each call to this chain,
and passed as arguments to the handlers defined in callbacks.
You can use these to eg identify a specific instance of a chain with its use case.
param output_parser: BaseLLMOutputParser [Optional]¶
Output parser to use.
Defaults to one that takes the most likely string but does not change it
otherwise.
param prompt: BasePromptTemplate [Required]¶
Prompt object to use.
param return_final_only: bool = True¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.QAEvalChain.html
|
ee9315a0d90d-1
|
Prompt object to use.
param return_final_only: bool = True¶
Whether to return only the final parsed result. Defaults to True.
If false, will return a bunch of extra information about the generation.
param tags: Optional[List[str]] = None¶
Optional list of tags associated with the chain. Defaults to None.
These tags will be associated with each call to this chain,
and passed as arguments to the handlers defined in callbacks.
You can use these to eg identify a specific instance of a chain with its use case.
param verbose: bool [Optional]¶
Whether or not run in verbose mode. In verbose mode, some intermediate logs
will be printed to the console. Defaults to langchain.verbose value.
__call__(inputs: Union[Dict[str, Any], Any], return_only_outputs: bool = False, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, *, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, include_run_info: bool = False) → Dict[str, Any]¶
Execute the chain.
Parameters
inputs – Dictionary of inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
return_only_outputs – Whether to return only outputs in the
response. If True, only new keys generated by this chain will be
returned. If False, both input keys and new keys generated by this
chain will be returned. Defaults to False.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.QAEvalChain.html
|
ee9315a0d90d-2
|
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
metadata – Optional metadata associated with the chain. Defaults to None
include_run_info – Whether to include run info in the response. Defaults
to False.
Returns
A dict of named outputs. Should contain all outputs specified inChain.output_keys.
async aapply(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → List[Dict[str, str]]¶
Utilize the LLM generate method for speed gains.
async aapply_and_parse(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → Sequence[Union[str, List[str], Dict[str, str]]]¶
Call apply and then parse the results.
async abatch(inputs: List[Input], config: Optional[Union[RunnableConfig, List[RunnableConfig]]] = None, *, max_concurrency: Optional[int] = None) → List[Output]¶
async acall(inputs: Union[Dict[str, Any], Any], return_only_outputs: bool = False, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, *, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, include_run_info: bool = False) → Dict[str, Any]¶
Asynchronously execute the chain.
Parameters
inputs – Dictionary of inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
return_only_outputs – Whether to return only outputs in the
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.QAEvalChain.html
|
ee9315a0d90d-3
|
memory.
return_only_outputs – Whether to return only outputs in the
response. If True, only new keys generated by this chain will be
returned. If False, both input keys and new keys generated by this
chain will be returned. Defaults to False.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
metadata – Optional metadata associated with the chain. Defaults to None
include_run_info – Whether to include run info in the response. Defaults
to False.
Returns
A dict of named outputs. Should contain all outputs specified inChain.output_keys.
async aevaluate_strings(*, prediction: str, reference: Optional[str] = None, input: Optional[str] = None, **kwargs: Any) → dict¶
Asynchronously evaluate Chain or LLM output, based on optional input and label.
Parameters
prediction (str) – The LLM or chain prediction to evaluate.
reference (Optional[str], optional) – The reference label to evaluate against.
input (Optional[str], optional) – The input to consider during evaluation.
**kwargs – Additional keyword arguments, including callbacks, tags, etc.
Returns
The evaluation results containing the score or value.
Return type
dict
async agenerate(input_list: List[Dict[str, Any]], run_manager: Optional[AsyncCallbackManagerForChainRun] = None) → LLMResult¶
Generate LLM result from inputs.
async ainvoke(input: Dict[str, Any], config: Optional[RunnableConfig] = None) → Dict[str, Any]¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.QAEvalChain.html
|
ee9315a0d90d-4
|
apply(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → List[Dict[str, str]]¶
Utilize the LLM generate method for speed gains.
apply_and_parse(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → Sequence[Union[str, List[str], Dict[str, str]]]¶
Call apply and then parse the results.
async apredict(callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, **kwargs: Any) → str¶
Format prompt with kwargs and pass to LLM.
Parameters
callbacks – Callbacks to pass to LLMChain
**kwargs – Keys to pass to prompt template.
Returns
Completion from LLM.
Example
completion = llm.predict(adjective="funny")
async apredict_and_parse(callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, **kwargs: Any) → Union[str, List[str], Dict[str, str]]¶
Call apredict and then parse the results.
async aprep_prompts(input_list: List[Dict[str, Any]], run_manager: Optional[AsyncCallbackManagerForChainRun] = None) → Tuple[List[PromptValue], Optional[List[str]]]¶
Prepare prompts from inputs.
async arun(*args: Any, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, **kwargs: Any) → Any¶
Convenience method for executing chain.
The main difference between this method and Chain.__call__ is that this
method expects inputs to be passed directly in as positional arguments or
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.QAEvalChain.html
|
ee9315a0d90d-5
|
method expects inputs to be passed directly in as positional arguments or
keyword arguments, whereas Chain.__call__ expects a single input dictionary
with all the inputs
Parameters
*args – If the chain expects a single input, it can be passed in as the
sole positional argument.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
**kwargs – If the chain expects multiple inputs, they can be passed in
directly as keyword arguments.
Returns
The chain output.
Example
# Suppose we have a single-input chain that takes a 'question' string:
await chain.arun("What's the temperature in Boise, Idaho?")
# -> "The temperature in Boise is..."
# Suppose we have a multi-input chain that takes a 'question' string
# and 'context' string:
question = "What's the temperature in Boise, Idaho?"
context = "Weather report for Boise, Idaho on 07/03/23..."
await chain.arun(question=question, context=context)
# -> "The temperature in Boise is..."
async astream(input: Input, config: Optional[RunnableConfig] = None) → AsyncIterator[Output]¶
batch(inputs: List[Input], config: Optional[Union[RunnableConfig, List[RunnableConfig]]] = None, *, max_concurrency: Optional[int] = None) → List[Output]¶
bind(**kwargs: Any) → Runnable[Input, Output]¶
Bind arguments to a Runnable, returning a new Runnable.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.QAEvalChain.html
|
ee9315a0d90d-6
|
Bind arguments to a Runnable, returning a new Runnable.
classmethod construct(_fields_set: Optional[SetStr] = None, **values: Any) → Model¶
Creates a new model setting __dict__ and __fields_set__ from trusted or pre-validated data.
Default values are respected, but no other validation is performed.
Behaves as if Config.extra = ‘allow’ was set since it adds all passed values
copy(*, include: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, exclude: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, update: Optional[DictStrAny] = None, deep: bool = False) → Model¶
Duplicate a model, optionally choose which fields to include, exclude and change.
Parameters
include – fields to include in new model
exclude – fields to exclude from new model, as with values this takes precedence over include
update – values to change/add in the new model. Note: the data is not validated before creating
the new model: you should trust this data
deep – set to True to make a deep copy of the model
Returns
new model instance
create_outputs(llm_result: LLMResult) → List[Dict[str, Any]]¶
Create outputs from response.
dict(**kwargs: Any) → Dict¶
Dictionary representation of chain.
Expects Chain._chain_type property to be implemented and for memory to benull.
Parameters
**kwargs – Keyword arguments passed to default pydantic.BaseModel.dict
method.
Returns
A dictionary representation of the chain.
Example
..code-block:: python
chain.dict(exclude_unset=True)
# -> {“_type”: “foo”, “verbose”: False, …}
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.QAEvalChain.html
|
ee9315a0d90d-7
|
# -> {“_type”: “foo”, “verbose”: False, …}
evaluate(examples: Sequence[dict], predictions: Sequence[dict], question_key: str = 'query', answer_key: str = 'answer', prediction_key: str = 'result', *, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → List[dict][source]¶
Evaluate question answering examples and predictions.
evaluate_strings(*, prediction: str, reference: Optional[str] = None, input: Optional[str] = None, **kwargs: Any) → dict¶
Evaluate Chain or LLM output, based on optional input and label.
Parameters
prediction (str) – The LLM or chain prediction to evaluate.
reference (Optional[str], optional) – The reference label to evaluate against.
input (Optional[str], optional) – The input to consider during evaluation.
**kwargs – Additional keyword arguments, including callbacks, tags, etc.
Returns
The evaluation results containing the score or value.
Return type
dict
classmethod from_llm(llm: BaseLanguageModel, prompt: Optional[PromptTemplate] = None, **kwargs: Any) → QAEvalChain[source]¶
Load QA Eval Chain from LLM.
Parameters
llm (BaseLanguageModel) – the base language model to use.
prompt ('answer' and 'result' that will be used as the) – A prompt template containing the input_variables:
'input' –
prompt –
evaluation. (for) –
PROMPT. (Defaults to) –
**kwargs – additional keyword arguments.
Returns
the loaded QA eval chain.
Return type
QAEvalChain
classmethod from_orm(obj: Any) → Model¶
classmethod from_string(llm: BaseLanguageModel, template: str) → LLMChain¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.QAEvalChain.html
|
ee9315a0d90d-8
|
classmethod from_string(llm: BaseLanguageModel, template: str) → LLMChain¶
Create LLMChain from LLM and template.
generate(input_list: List[Dict[str, Any]], run_manager: Optional[CallbackManagerForChainRun] = None) → LLMResult¶
Generate LLM result from inputs.
invoke(input: Dict[str, Any], config: Optional[RunnableConfig] = None) → Dict[str, Any]¶
json(*, include: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, exclude: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, by_alias: bool = False, skip_defaults: Optional[bool] = None, exclude_unset: bool = False, exclude_defaults: bool = False, exclude_none: bool = False, encoder: Optional[Callable[[Any], Any]] = None, models_as_dict: bool = True, **dumps_kwargs: Any) → unicode¶
Generate a JSON representation of the model, include and exclude arguments as per dict().
encoder is an optional function to supply as default to json.dumps(), other arguments as per json.dumps().
classmethod parse_file(path: Union[str, Path], *, content_type: unicode = None, encoding: unicode = 'utf8', proto: Protocol = None, allow_pickle: bool = False) → Model¶
classmethod parse_obj(obj: Any) → Model¶
classmethod parse_raw(b: Union[str, bytes], *, content_type: unicode = None, encoding: unicode = 'utf8', proto: Protocol = None, allow_pickle: bool = False) → Model¶
predict(callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, **kwargs: Any) → str¶
Format prompt with kwargs and pass to LLM.
Parameters
callbacks – Callbacks to pass to LLMChain
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.QAEvalChain.html
|
ee9315a0d90d-9
|
Parameters
callbacks – Callbacks to pass to LLMChain
**kwargs – Keys to pass to prompt template.
Returns
Completion from LLM.
Example
completion = llm.predict(adjective="funny")
predict_and_parse(callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, **kwargs: Any) → Union[str, List[str], Dict[str, Any]]¶
Call predict and then parse the results.
prep_inputs(inputs: Union[Dict[str, Any], Any]) → Dict[str, str]¶
Validate and prepare chain inputs, including adding inputs from memory.
Parameters
inputs – Dictionary of raw inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
Returns
A dictionary of all inputs, including those added by the chain’s memory.
prep_outputs(inputs: Dict[str, str], outputs: Dict[str, str], return_only_outputs: bool = False) → Dict[str, str]¶
Validate and prepare chain outputs, and save info about this run to memory.
Parameters
inputs – Dictionary of chain inputs, including any inputs added by chain
memory.
outputs – Dictionary of initial chain outputs.
return_only_outputs – Whether to only return the chain outputs. If False,
inputs are also added to the final outputs.
Returns
A dict of the final chain outputs.
prep_prompts(input_list: List[Dict[str, Any]], run_manager: Optional[CallbackManagerForChainRun] = None) → Tuple[List[PromptValue], Optional[List[str]]]¶
Prepare prompts from inputs.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.QAEvalChain.html
|
ee9315a0d90d-10
|
Prepare prompts from inputs.
run(*args: Any, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, **kwargs: Any) → Any¶
Convenience method for executing chain.
The main difference between this method and Chain.__call__ is that this
method expects inputs to be passed directly in as positional arguments or
keyword arguments, whereas Chain.__call__ expects a single input dictionary
with all the inputs
Parameters
*args – If the chain expects a single input, it can be passed in as the
sole positional argument.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
**kwargs – If the chain expects multiple inputs, they can be passed in
directly as keyword arguments.
Returns
The chain output.
Example
# Suppose we have a single-input chain that takes a 'question' string:
chain.run("What's the temperature in Boise, Idaho?")
# -> "The temperature in Boise is..."
# Suppose we have a multi-input chain that takes a 'question' string
# and 'context' string:
question = "What's the temperature in Boise, Idaho?"
context = "Weather report for Boise, Idaho on 07/03/23..."
chain.run(question=question, context=context)
# -> "The temperature in Boise is..."
save(file_path: Union[Path, str]) → None¶
Save the chain.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.QAEvalChain.html
|
ee9315a0d90d-11
|
save(file_path: Union[Path, str]) → None¶
Save the chain.
Expects Chain._chain_type property to be implemented and for memory to benull.
Parameters
file_path – Path to file to save the chain to.
Example
chain.save(file_path="path/chain.yaml")
classmethod schema(by_alias: bool = True, ref_template: unicode = '#/definitions/{model}') → DictStrAny¶
classmethod schema_json(*, by_alias: bool = True, ref_template: unicode = '#/definitions/{model}', **dumps_kwargs: Any) → unicode¶
stream(input: Input, config: Optional[RunnableConfig] = None) → Iterator[Output]¶
to_json() → Union[SerializedConstructor, SerializedNotImplemented]¶
to_json_not_implemented() → SerializedNotImplemented¶
classmethod update_forward_refs(**localns: Any) → None¶
Try to update ForwardRefs on fields based on this Model, globalns and localns.
classmethod validate(value: Any) → Model¶
with_fallbacks(fallbacks: ~typing.Sequence[~langchain.schema.runnable.Runnable[~langchain.schema.runnable.Input, ~langchain.schema.runnable.Output]], *, exceptions_to_handle: ~typing.Tuple[~typing.Type[BaseException]] = (<class 'Exception'>,)) → RunnableWithFallbacks[Input, Output]¶
property evaluation_name: str¶
The name of the evaluation.
property lc_attributes: Dict¶
Return a list of attribute names that should be included in the
serialized kwargs. These attributes must be accepted by the
constructor.
property lc_namespace: List[str]¶
Return the namespace of the langchain object.
eg. [“langchain”, “llms”, “openai”]
property lc_secrets: Dict[str, str]¶
Return a map of constructor argument names to secret ids.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.QAEvalChain.html
|
ee9315a0d90d-12
|
Return a map of constructor argument names to secret ids.
eg. {“openai_api_key”: “OPENAI_API_KEY”}
property lc_serializable: bool¶
Return whether or not the class is serializable.
property requires_input: bool¶
Whether this evaluator requires an input string.
property requires_reference: bool¶
Whether this evaluator requires a reference label.
Examples using QAEvalChain¶
Question Answering Benchmarking: State of the Union Address
Question Answering Benchmarking: Paul Graham Essay
Data Augmented Question Answering
SQL Question Answering Benchmarking: Chinook
Question Answering
Agent VectorDB Question Answering Benchmarking
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.QAEvalChain.html
|
e04c80130c32-0
|
langchain.evaluation.agents.trajectory_eval_chain.TrajectoryEvalChain¶
class langchain.evaluation.agents.trajectory_eval_chain.TrajectoryEvalChain[source]¶
Bases: AgentTrajectoryEvaluator, LLMEvalChain
A chain for evaluating ReAct style agents.
This chain is used to evaluate ReAct style agents by reasoning about
the sequence of actions taken and their outcomes.
Example:
from langchain.agents import AgentType, initialize_agent
from langchain.chat_models import ChatOpenAI
from langchain.evaluation import TrajectoryEvalChain
from langchain.tools import tool
@tool
def geography_answers(country: str, question: str) -> str:
"""Very helpful answers to geography questions."""
return f"{country}? IDK - We may never know {question}."
llm = ChatOpenAI(model="gpt-3.5-turbo", temperature=0)
agent = initialize_agent(
tools=[geography_answers],
llm=llm,
agent=AgentType.OPENAI_FUNCTIONS,
return_intermediate_steps=True,
)
question = "How many dwell in the largest minor region in Argentina?"
response = agent(question)
eval_chain = TrajectoryEvalChain.from_llm(
llm=llm, agent_tools=[geography_answers], return_reasoning=True
)
result = eval_chain.evaluate_agent_trajectory(
input=question,
agent_trajectory=response["intermediate_steps"],
prediction=response["output"],
reference="Paris",
)
print(result["score"])
# 0
param agent_tools: Optional[List[langchain.tools.base.BaseTool]] = None¶
A list of tools available to the agent.
param callback_manager: Optional[BaseCallbackManager] = None¶
Deprecated, use callbacks instead.
param callbacks: Callbacks = None¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.agents.trajectory_eval_chain.TrajectoryEvalChain.html
|
e04c80130c32-1
|
Deprecated, use callbacks instead.
param callbacks: Callbacks = None¶
Optional list of callback handlers (or callback manager). Defaults to None.
Callback handlers are called throughout the lifecycle of a call to a chain,
starting with on_chain_start, ending with on_chain_end or on_chain_error.
Each custom chain can optionally call additional callback methods, see Callback docs
for full details.
param eval_chain: langchain.chains.llm.LLMChain [Required]¶
The language model chain used for evaluation.
param memory: Optional[BaseMemory] = None¶
Optional memory object. Defaults to None.
Memory is a class that gets called at the start
and at the end of every chain. At the start, memory loads variables and passes
them along in the chain. At the end, it saves any returned variables.
There are many different types of memory - please see memory docs
for the full catalog.
param metadata: Optional[Dict[str, Any]] = None¶
Optional metadata associated with the chain. Defaults to None.
This metadata will be associated with each call to this chain,
and passed as arguments to the handlers defined in callbacks.
You can use these to eg identify a specific instance of a chain with its use case.
param output_parser: langchain.evaluation.agents.trajectory_eval_chain.TrajectoryOutputParser [Optional]¶
The output parser used to parse the output.
param return_reasoning: bool = False¶
DEPRECATED. Reasoning always returned.
param tags: Optional[List[str]] = None¶
Optional list of tags associated with the chain. Defaults to None.
These tags will be associated with each call to this chain,
and passed as arguments to the handlers defined in callbacks.
You can use these to eg identify a specific instance of a chain with its use case.
param verbose: bool [Optional]¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.agents.trajectory_eval_chain.TrajectoryEvalChain.html
|
e04c80130c32-2
|
param verbose: bool [Optional]¶
Whether or not run in verbose mode. In verbose mode, some intermediate logs
will be printed to the console. Defaults to langchain.verbose value.
__call__(inputs: Union[Dict[str, Any], Any], return_only_outputs: bool = False, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, *, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, include_run_info: bool = False) → Dict[str, Any]¶
Execute the chain.
Parameters
inputs – Dictionary of inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
return_only_outputs – Whether to return only outputs in the
response. If True, only new keys generated by this chain will be
returned. If False, both input keys and new keys generated by this
chain will be returned. Defaults to False.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
metadata – Optional metadata associated with the chain. Defaults to None
include_run_info – Whether to include run info in the response. Defaults
to False.
Returns
A dict of named outputs. Should contain all outputs specified inChain.output_keys.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.agents.trajectory_eval_chain.TrajectoryEvalChain.html
|
e04c80130c32-3
|
Returns
A dict of named outputs. Should contain all outputs specified inChain.output_keys.
async abatch(inputs: List[Input], config: Optional[Union[RunnableConfig, List[RunnableConfig]]] = None, *, max_concurrency: Optional[int] = None) → List[Output]¶
async acall(inputs: Union[Dict[str, Any], Any], return_only_outputs: bool = False, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, *, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, include_run_info: bool = False) → Dict[str, Any]¶
Asynchronously execute the chain.
Parameters
inputs – Dictionary of inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
return_only_outputs – Whether to return only outputs in the
response. If True, only new keys generated by this chain will be
returned. If False, both input keys and new keys generated by this
chain will be returned. Defaults to False.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
metadata – Optional metadata associated with the chain. Defaults to None
include_run_info – Whether to include run info in the response. Defaults
to False.
Returns
A dict of named outputs. Should contain all outputs specified inChain.output_keys.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.agents.trajectory_eval_chain.TrajectoryEvalChain.html
|
e04c80130c32-4
|
Returns
A dict of named outputs. Should contain all outputs specified inChain.output_keys.
async aevaluate_agent_trajectory(*, prediction: str, agent_trajectory: Sequence[Tuple[AgentAction, str]], input: str, reference: Optional[str] = None, **kwargs: Any) → dict¶
Asynchronously evaluate a trajectory.
Parameters
prediction (str) – The final predicted response.
agent_trajectory (List[Tuple[AgentAction, str]]) – The intermediate steps forming the agent trajectory.
input (str) – The input to the agent.
reference (Optional[str]) – The reference answer.
Returns
The evaluation result.
Return type
dict
async ainvoke(input: Dict[str, Any], config: Optional[RunnableConfig] = None) → Dict[str, Any]¶
apply(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → List[Dict[str, str]]¶
Call the chain on all inputs in the list.
async arun(*args: Any, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, **kwargs: Any) → Any¶
Convenience method for executing chain.
The main difference between this method and Chain.__call__ is that this
method expects inputs to be passed directly in as positional arguments or
keyword arguments, whereas Chain.__call__ expects a single input dictionary
with all the inputs
Parameters
*args – If the chain expects a single input, it can be passed in as the
sole positional argument.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.agents.trajectory_eval_chain.TrajectoryEvalChain.html
|
e04c80130c32-5
|
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
**kwargs – If the chain expects multiple inputs, they can be passed in
directly as keyword arguments.
Returns
The chain output.
Example
# Suppose we have a single-input chain that takes a 'question' string:
await chain.arun("What's the temperature in Boise, Idaho?")
# -> "The temperature in Boise is..."
# Suppose we have a multi-input chain that takes a 'question' string
# and 'context' string:
question = "What's the temperature in Boise, Idaho?"
context = "Weather report for Boise, Idaho on 07/03/23..."
await chain.arun(question=question, context=context)
# -> "The temperature in Boise is..."
async astream(input: Input, config: Optional[RunnableConfig] = None) → AsyncIterator[Output]¶
batch(inputs: List[Input], config: Optional[Union[RunnableConfig, List[RunnableConfig]]] = None, *, max_concurrency: Optional[int] = None) → List[Output]¶
bind(**kwargs: Any) → Runnable[Input, Output]¶
Bind arguments to a Runnable, returning a new Runnable.
classmethod construct(_fields_set: Optional[SetStr] = None, **values: Any) → Model¶
Creates a new model setting __dict__ and __fields_set__ from trusted or pre-validated data.
Default values are respected, but no other validation is performed.
Behaves as if Config.extra = ‘allow’ was set since it adds all passed values
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.agents.trajectory_eval_chain.TrajectoryEvalChain.html
|
e04c80130c32-6
|
Behaves as if Config.extra = ‘allow’ was set since it adds all passed values
copy(*, include: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, exclude: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, update: Optional[DictStrAny] = None, deep: bool = False) → Model¶
Duplicate a model, optionally choose which fields to include, exclude and change.
Parameters
include – fields to include in new model
exclude – fields to exclude from new model, as with values this takes precedence over include
update – values to change/add in the new model. Note: the data is not validated before creating
the new model: you should trust this data
deep – set to True to make a deep copy of the model
Returns
new model instance
dict(**kwargs: Any) → Dict¶
Dictionary representation of chain.
Expects Chain._chain_type property to be implemented and for memory to benull.
Parameters
**kwargs – Keyword arguments passed to default pydantic.BaseModel.dict
method.
Returns
A dictionary representation of the chain.
Example
..code-block:: python
chain.dict(exclude_unset=True)
# -> {“_type”: “foo”, “verbose”: False, …}
evaluate_agent_trajectory(*, prediction: str, agent_trajectory: Sequence[Tuple[AgentAction, str]], input: str, reference: Optional[str] = None, **kwargs: Any) → dict¶
Evaluate a trajectory.
Parameters
prediction (str) – The final predicted response.
agent_trajectory (List[Tuple[AgentAction, str]]) – The intermediate steps forming the agent trajectory.
input (str) – The input to the agent.
reference (Optional[str]) – The reference answer.
Returns
The evaluation result.
Return type
dict
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.agents.trajectory_eval_chain.TrajectoryEvalChain.html
|
e04c80130c32-7
|
Returns
The evaluation result.
Return type
dict
classmethod from_llm(llm: BaseLanguageModel, agent_tools: Optional[Sequence[BaseTool]] = None, output_parser: Optional[TrajectoryOutputParser] = None, **kwargs: Any) → TrajectoryEvalChain[source]¶
Create a TrajectoryEvalChain object from a language model chain.
Parameters
llm (BaseChatModel) – The language model chain.
agent_tools (Optional[Sequence[BaseTool]]) – A list of tools
available to the agent.
output_parser (Optional[TrajectoryOutputParser]) – The output parser
used to parse the chain output into a score.
Returns
The TrajectoryEvalChain object.
Return type
TrajectoryEvalChain
classmethod from_orm(obj: Any) → Model¶
static get_agent_trajectory(steps: Union[str, Sequence[Tuple[AgentAction, str]]]) → str[source]¶
Get the agent trajectory as a formatted string.
Parameters
steps (Union[str, List[Tuple[AgentAction, str]]]) – The agent trajectory.
Returns
The formatted agent trajectory.
Return type
str
invoke(input: Dict[str, Any], config: Optional[RunnableConfig] = None) → Dict[str, Any]¶
json(*, include: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, exclude: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, by_alias: bool = False, skip_defaults: Optional[bool] = None, exclude_unset: bool = False, exclude_defaults: bool = False, exclude_none: bool = False, encoder: Optional[Callable[[Any], Any]] = None, models_as_dict: bool = True, **dumps_kwargs: Any) → unicode¶
Generate a JSON representation of the model, include and exclude arguments as per dict().
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.agents.trajectory_eval_chain.TrajectoryEvalChain.html
|
e04c80130c32-8
|
Generate a JSON representation of the model, include and exclude arguments as per dict().
encoder is an optional function to supply as default to json.dumps(), other arguments as per json.dumps().
classmethod parse_file(path: Union[str, Path], *, content_type: unicode = None, encoding: unicode = 'utf8', proto: Protocol = None, allow_pickle: bool = False) → Model¶
classmethod parse_obj(obj: Any) → Model¶
classmethod parse_raw(b: Union[str, bytes], *, content_type: unicode = None, encoding: unicode = 'utf8', proto: Protocol = None, allow_pickle: bool = False) → Model¶
prep_inputs(inputs: Union[Dict[str, Any], Any]) → Dict[str, str][source]¶
Validate and prep inputs.
prep_outputs(inputs: Dict[str, str], outputs: Dict[str, str], return_only_outputs: bool = False) → Dict[str, str]¶
Validate and prepare chain outputs, and save info about this run to memory.
Parameters
inputs – Dictionary of chain inputs, including any inputs added by chain
memory.
outputs – Dictionary of initial chain outputs.
return_only_outputs – Whether to only return the chain outputs. If False,
inputs are also added to the final outputs.
Returns
A dict of the final chain outputs.
run(*args: Any, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, **kwargs: Any) → Any¶
Convenience method for executing chain.
The main difference between this method and Chain.__call__ is that this
method expects inputs to be passed directly in as positional arguments or
keyword arguments, whereas Chain.__call__ expects a single input dictionary
with all the inputs
Parameters
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.agents.trajectory_eval_chain.TrajectoryEvalChain.html
|
e04c80130c32-9
|
with all the inputs
Parameters
*args – If the chain expects a single input, it can be passed in as the
sole positional argument.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
**kwargs – If the chain expects multiple inputs, they can be passed in
directly as keyword arguments.
Returns
The chain output.
Example
# Suppose we have a single-input chain that takes a 'question' string:
chain.run("What's the temperature in Boise, Idaho?")
# -> "The temperature in Boise is..."
# Suppose we have a multi-input chain that takes a 'question' string
# and 'context' string:
question = "What's the temperature in Boise, Idaho?"
context = "Weather report for Boise, Idaho on 07/03/23..."
chain.run(question=question, context=context)
# -> "The temperature in Boise is..."
save(file_path: Union[Path, str]) → None¶
Save the chain.
Expects Chain._chain_type property to be implemented and for memory to benull.
Parameters
file_path – Path to file to save the chain to.
Example
chain.save(file_path="path/chain.yaml")
classmethod schema(by_alias: bool = True, ref_template: unicode = '#/definitions/{model}') → DictStrAny¶
classmethod schema_json(*, by_alias: bool = True, ref_template: unicode = '#/definitions/{model}', **dumps_kwargs: Any) → unicode¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.agents.trajectory_eval_chain.TrajectoryEvalChain.html
|
e04c80130c32-10
|
stream(input: Input, config: Optional[RunnableConfig] = None) → Iterator[Output]¶
to_json() → Union[SerializedConstructor, SerializedNotImplemented]¶
to_json_not_implemented() → SerializedNotImplemented¶
classmethod update_forward_refs(**localns: Any) → None¶
Try to update ForwardRefs on fields based on this Model, globalns and localns.
classmethod validate(value: Any) → Model¶
with_fallbacks(fallbacks: ~typing.Sequence[~langchain.schema.runnable.Runnable[~langchain.schema.runnable.Input, ~langchain.schema.runnable.Output]], *, exceptions_to_handle: ~typing.Tuple[~typing.Type[BaseException]] = (<class 'Exception'>,)) → RunnableWithFallbacks[Input, Output]¶
property input_keys: List[str]¶
Get the input keys for the chain.
Returns
The input keys.
Return type
List[str]
property lc_attributes: Dict¶
Return a list of attribute names that should be included in the
serialized kwargs. These attributes must be accepted by the
constructor.
property lc_namespace: List[str]¶
Return the namespace of the langchain object.
eg. [“langchain”, “llms”, “openai”]
property lc_secrets: Dict[str, str]¶
Return a map of constructor argument names to secret ids.
eg. {“openai_api_key”: “OPENAI_API_KEY”}
property lc_serializable: bool¶
Return whether or not the class is serializable.
property output_keys: List[str]¶
Get the output keys for the chain.
Returns
The output keys.
Return type
List[str]
property requires_input: bool¶
Whether this evaluator requires an input string.
property requires_reference: bool¶
Whether this evaluator requires a reference label.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.agents.trajectory_eval_chain.TrajectoryEvalChain.html
|
8c8019e02616-0
|
langchain.evaluation.schema.LLMEvalChain¶
class langchain.evaluation.schema.LLMEvalChain[source]¶
Bases: Chain
A base class for evaluators that use an LLM.
Create a new model by parsing and validating input data from keyword arguments.
Raises ValidationError if the input data cannot be parsed to form a valid model.
param callback_manager: Optional[langchain.callbacks.base.BaseCallbackManager] = None¶
Deprecated, use callbacks instead.
param callbacks: Optional[Union[List[langchain.callbacks.base.BaseCallbackHandler], langchain.callbacks.base.BaseCallbackManager]] = None¶
Optional list of callback handlers (or callback manager). Defaults to None.
Callback handlers are called throughout the lifecycle of a call to a chain,
starting with on_chain_start, ending with on_chain_end or on_chain_error.
Each custom chain can optionally call additional callback methods, see Callback docs
for full details.
param memory: Optional[langchain.schema.memory.BaseMemory] = None¶
Optional memory object. Defaults to None.
Memory is a class that gets called at the start
and at the end of every chain. At the start, memory loads variables and passes
them along in the chain. At the end, it saves any returned variables.
There are many different types of memory - please see memory docs
for the full catalog.
param metadata: Optional[Dict[str, Any]] = None¶
Optional metadata associated with the chain. Defaults to None.
This metadata will be associated with each call to this chain,
and passed as arguments to the handlers defined in callbacks.
You can use these to eg identify a specific instance of a chain with its use case.
param tags: Optional[List[str]] = None¶
Optional list of tags associated with the chain. Defaults to None.
These tags will be associated with each call to this chain,
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.schema.LLMEvalChain.html
|
8c8019e02616-1
|
These tags will be associated with each call to this chain,
and passed as arguments to the handlers defined in callbacks.
You can use these to eg identify a specific instance of a chain with its use case.
param verbose: bool [Optional]¶
Whether or not run in verbose mode. In verbose mode, some intermediate logs
will be printed to the console. Defaults to langchain.verbose value.
__call__(inputs: Union[Dict[str, Any], Any], return_only_outputs: bool = False, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, *, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, include_run_info: bool = False) → Dict[str, Any]¶
Execute the chain.
Parameters
inputs – Dictionary of inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
return_only_outputs – Whether to return only outputs in the
response. If True, only new keys generated by this chain will be
returned. If False, both input keys and new keys generated by this
chain will be returned. Defaults to False.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
metadata – Optional metadata associated with the chain. Defaults to None
include_run_info – Whether to include run info in the response. Defaults
to False.
Returns
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.schema.LLMEvalChain.html
|
8c8019e02616-2
|
to False.
Returns
A dict of named outputs. Should contain all outputs specified inChain.output_keys.
async abatch(inputs: List[Input], config: Optional[Union[RunnableConfig, List[RunnableConfig]]] = None, *, max_concurrency: Optional[int] = None) → List[Output]¶
async acall(inputs: Union[Dict[str, Any], Any], return_only_outputs: bool = False, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, *, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, include_run_info: bool = False) → Dict[str, Any]¶
Asynchronously execute the chain.
Parameters
inputs – Dictionary of inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
return_only_outputs – Whether to return only outputs in the
response. If True, only new keys generated by this chain will be
returned. If False, both input keys and new keys generated by this
chain will be returned. Defaults to False.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
metadata – Optional metadata associated with the chain. Defaults to None
include_run_info – Whether to include run info in the response. Defaults
to False.
Returns
A dict of named outputs. Should contain all outputs specified inChain.output_keys.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.schema.LLMEvalChain.html
|
8c8019e02616-3
|
Returns
A dict of named outputs. Should contain all outputs specified inChain.output_keys.
async ainvoke(input: Dict[str, Any], config: Optional[RunnableConfig] = None) → Dict[str, Any]¶
apply(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → List[Dict[str, str]]¶
Call the chain on all inputs in the list.
async arun(*args: Any, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, **kwargs: Any) → Any¶
Convenience method for executing chain.
The main difference between this method and Chain.__call__ is that this
method expects inputs to be passed directly in as positional arguments or
keyword arguments, whereas Chain.__call__ expects a single input dictionary
with all the inputs
Parameters
*args – If the chain expects a single input, it can be passed in as the
sole positional argument.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
**kwargs – If the chain expects multiple inputs, they can be passed in
directly as keyword arguments.
Returns
The chain output.
Example
# Suppose we have a single-input chain that takes a 'question' string:
await chain.arun("What's the temperature in Boise, Idaho?")
# -> "The temperature in Boise is..."
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.schema.LLMEvalChain.html
|
8c8019e02616-4
|
# -> "The temperature in Boise is..."
# Suppose we have a multi-input chain that takes a 'question' string
# and 'context' string:
question = "What's the temperature in Boise, Idaho?"
context = "Weather report for Boise, Idaho on 07/03/23..."
await chain.arun(question=question, context=context)
# -> "The temperature in Boise is..."
async astream(input: Input, config: Optional[RunnableConfig] = None) → AsyncIterator[Output]¶
batch(inputs: List[Input], config: Optional[Union[RunnableConfig, List[RunnableConfig]]] = None, *, max_concurrency: Optional[int] = None) → List[Output]¶
bind(**kwargs: Any) → Runnable[Input, Output]¶
Bind arguments to a Runnable, returning a new Runnable.
classmethod construct(_fields_set: Optional[SetStr] = None, **values: Any) → Model¶
Creates a new model setting __dict__ and __fields_set__ from trusted or pre-validated data.
Default values are respected, but no other validation is performed.
Behaves as if Config.extra = ‘allow’ was set since it adds all passed values
copy(*, include: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, exclude: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, update: Optional[DictStrAny] = None, deep: bool = False) → Model¶
Duplicate a model, optionally choose which fields to include, exclude and change.
Parameters
include – fields to include in new model
exclude – fields to exclude from new model, as with values this takes precedence over include
update – values to change/add in the new model. Note: the data is not validated before creating
the new model: you should trust this data
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.schema.LLMEvalChain.html
|
8c8019e02616-5
|
the new model: you should trust this data
deep – set to True to make a deep copy of the model
Returns
new model instance
dict(**kwargs: Any) → Dict¶
Dictionary representation of chain.
Expects Chain._chain_type property to be implemented and for memory to benull.
Parameters
**kwargs – Keyword arguments passed to default pydantic.BaseModel.dict
method.
Returns
A dictionary representation of the chain.
Example
..code-block:: python
chain.dict(exclude_unset=True)
# -> {“_type”: “foo”, “verbose”: False, …}
abstract classmethod from_llm(llm: BaseLanguageModel, **kwargs: Any) → LLMEvalChain[source]¶
Create a new evaluator from an LLM.
classmethod from_orm(obj: Any) → Model¶
invoke(input: Dict[str, Any], config: Optional[RunnableConfig] = None) → Dict[str, Any]¶
json(*, include: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, exclude: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, by_alias: bool = False, skip_defaults: Optional[bool] = None, exclude_unset: bool = False, exclude_defaults: bool = False, exclude_none: bool = False, encoder: Optional[Callable[[Any], Any]] = None, models_as_dict: bool = True, **dumps_kwargs: Any) → unicode¶
Generate a JSON representation of the model, include and exclude arguments as per dict().
encoder is an optional function to supply as default to json.dumps(), other arguments as per json.dumps().
classmethod parse_file(path: Union[str, Path], *, content_type: unicode = None, encoding: unicode = 'utf8', proto: Protocol = None, allow_pickle: bool = False) → Model¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.schema.LLMEvalChain.html
|
8c8019e02616-6
|
classmethod parse_obj(obj: Any) → Model¶
classmethod parse_raw(b: Union[str, bytes], *, content_type: unicode = None, encoding: unicode = 'utf8', proto: Protocol = None, allow_pickle: bool = False) → Model¶
prep_inputs(inputs: Union[Dict[str, Any], Any]) → Dict[str, str]¶
Validate and prepare chain inputs, including adding inputs from memory.
Parameters
inputs – Dictionary of raw inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
Returns
A dictionary of all inputs, including those added by the chain’s memory.
prep_outputs(inputs: Dict[str, str], outputs: Dict[str, str], return_only_outputs: bool = False) → Dict[str, str]¶
Validate and prepare chain outputs, and save info about this run to memory.
Parameters
inputs – Dictionary of chain inputs, including any inputs added by chain
memory.
outputs – Dictionary of initial chain outputs.
return_only_outputs – Whether to only return the chain outputs. If False,
inputs are also added to the final outputs.
Returns
A dict of the final chain outputs.
run(*args: Any, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, **kwargs: Any) → Any¶
Convenience method for executing chain.
The main difference between this method and Chain.__call__ is that this
method expects inputs to be passed directly in as positional arguments or
keyword arguments, whereas Chain.__call__ expects a single input dictionary
with all the inputs
Parameters
*args – If the chain expects a single input, it can be passed in as the
sole positional argument.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.schema.LLMEvalChain.html
|
8c8019e02616-7
|
sole positional argument.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
**kwargs – If the chain expects multiple inputs, they can be passed in
directly as keyword arguments.
Returns
The chain output.
Example
# Suppose we have a single-input chain that takes a 'question' string:
chain.run("What's the temperature in Boise, Idaho?")
# -> "The temperature in Boise is..."
# Suppose we have a multi-input chain that takes a 'question' string
# and 'context' string:
question = "What's the temperature in Boise, Idaho?"
context = "Weather report for Boise, Idaho on 07/03/23..."
chain.run(question=question, context=context)
# -> "The temperature in Boise is..."
save(file_path: Union[Path, str]) → None¶
Save the chain.
Expects Chain._chain_type property to be implemented and for memory to benull.
Parameters
file_path – Path to file to save the chain to.
Example
chain.save(file_path="path/chain.yaml")
classmethod schema(by_alias: bool = True, ref_template: unicode = '#/definitions/{model}') → DictStrAny¶
classmethod schema_json(*, by_alias: bool = True, ref_template: unicode = '#/definitions/{model}', **dumps_kwargs: Any) → unicode¶
stream(input: Input, config: Optional[RunnableConfig] = None) → Iterator[Output]¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.schema.LLMEvalChain.html
|
8c8019e02616-8
|
to_json() → Union[SerializedConstructor, SerializedNotImplemented]¶
to_json_not_implemented() → SerializedNotImplemented¶
classmethod update_forward_refs(**localns: Any) → None¶
Try to update ForwardRefs on fields based on this Model, globalns and localns.
classmethod validate(value: Any) → Model¶
with_fallbacks(fallbacks: ~typing.Sequence[~langchain.schema.runnable.Runnable[~langchain.schema.runnable.Input, ~langchain.schema.runnable.Output]], *, exceptions_to_handle: ~typing.Tuple[~typing.Type[BaseException]] = (<class 'Exception'>,)) → RunnableWithFallbacks[Input, Output]¶
abstract property input_keys: List[str]¶
Keys expected to be in the chain input.
property lc_attributes: Dict¶
Return a list of attribute names that should be included in the
serialized kwargs. These attributes must be accepted by the
constructor.
property lc_namespace: List[str]¶
Return the namespace of the langchain object.
eg. [“langchain”, “llms”, “openai”]
property lc_secrets: Dict[str, str]¶
Return a map of constructor argument names to secret ids.
eg. {“openai_api_key”: “OPENAI_API_KEY”}
property lc_serializable: bool¶
Return whether or not the class is serializable.
abstract property output_keys: List[str]¶
Keys expected to be in the chain output.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.schema.LLMEvalChain.html
|
b1a8a0631ca2-0
|
langchain.evaluation.qa.eval_chain.ContextQAEvalChain¶
class langchain.evaluation.qa.eval_chain.ContextQAEvalChain[source]¶
Bases: LLMChain, StringEvaluator, LLMEvalChain
LLM Chain for evaluating QA w/o GT based on context
param callback_manager: Optional[BaseCallbackManager] = None¶
Deprecated, use callbacks instead.
param callbacks: Callbacks = None¶
Optional list of callback handlers (or callback manager). Defaults to None.
Callback handlers are called throughout the lifecycle of a call to a chain,
starting with on_chain_start, ending with on_chain_end or on_chain_error.
Each custom chain can optionally call additional callback methods, see Callback docs
for full details.
param llm: BaseLanguageModel [Required]¶
Language model to call.
param llm_kwargs: dict [Optional]¶
param memory: Optional[BaseMemory] = None¶
Optional memory object. Defaults to None.
Memory is a class that gets called at the start
and at the end of every chain. At the start, memory loads variables and passes
them along in the chain. At the end, it saves any returned variables.
There are many different types of memory - please see memory docs
for the full catalog.
param metadata: Optional[Dict[str, Any]] = None¶
Optional metadata associated with the chain. Defaults to None.
This metadata will be associated with each call to this chain,
and passed as arguments to the handlers defined in callbacks.
You can use these to eg identify a specific instance of a chain with its use case.
param output_key: str = 'text'¶
param output_parser: BaseLLMOutputParser [Optional]¶
Output parser to use.
Defaults to one that takes the most likely string but does not change it
otherwise.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.ContextQAEvalChain.html
|
b1a8a0631ca2-1
|
Defaults to one that takes the most likely string but does not change it
otherwise.
param prompt: BasePromptTemplate [Required]¶
Prompt object to use.
param return_final_only: bool = True¶
Whether to return only the final parsed result. Defaults to True.
If false, will return a bunch of extra information about the generation.
param tags: Optional[List[str]] = None¶
Optional list of tags associated with the chain. Defaults to None.
These tags will be associated with each call to this chain,
and passed as arguments to the handlers defined in callbacks.
You can use these to eg identify a specific instance of a chain with its use case.
param verbose: bool [Optional]¶
Whether or not run in verbose mode. In verbose mode, some intermediate logs
will be printed to the console. Defaults to langchain.verbose value.
__call__(inputs: Union[Dict[str, Any], Any], return_only_outputs: bool = False, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, *, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, include_run_info: bool = False) → Dict[str, Any]¶
Execute the chain.
Parameters
inputs – Dictionary of inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
return_only_outputs – Whether to return only outputs in the
response. If True, only new keys generated by this chain will be
returned. If False, both input keys and new keys generated by this
chain will be returned. Defaults to False.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.ContextQAEvalChain.html
|
b1a8a0631ca2-2
|
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
metadata – Optional metadata associated with the chain. Defaults to None
include_run_info – Whether to include run info in the response. Defaults
to False.
Returns
A dict of named outputs. Should contain all outputs specified inChain.output_keys.
async aapply(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → List[Dict[str, str]]¶
Utilize the LLM generate method for speed gains.
async aapply_and_parse(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → Sequence[Union[str, List[str], Dict[str, str]]]¶
Call apply and then parse the results.
async abatch(inputs: List[Input], config: Optional[Union[RunnableConfig, List[RunnableConfig]]] = None, *, max_concurrency: Optional[int] = None) → List[Output]¶
async acall(inputs: Union[Dict[str, Any], Any], return_only_outputs: bool = False, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, *, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, include_run_info: bool = False) → Dict[str, Any]¶
Asynchronously execute the chain.
Parameters
inputs – Dictionary of inputs, or single input if chain expects
only one param. Should contain all inputs specified in
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.ContextQAEvalChain.html
|
b1a8a0631ca2-3
|
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
return_only_outputs – Whether to return only outputs in the
response. If True, only new keys generated by this chain will be
returned. If False, both input keys and new keys generated by this
chain will be returned. Defaults to False.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
metadata – Optional metadata associated with the chain. Defaults to None
include_run_info – Whether to include run info in the response. Defaults
to False.
Returns
A dict of named outputs. Should contain all outputs specified inChain.output_keys.
async aevaluate_strings(*, prediction: str, reference: Optional[str] = None, input: Optional[str] = None, **kwargs: Any) → dict¶
Asynchronously evaluate Chain or LLM output, based on optional input and label.
Parameters
prediction (str) – The LLM or chain prediction to evaluate.
reference (Optional[str], optional) – The reference label to evaluate against.
input (Optional[str], optional) – The input to consider during evaluation.
**kwargs – Additional keyword arguments, including callbacks, tags, etc.
Returns
The evaluation results containing the score or value.
Return type
dict
async agenerate(input_list: List[Dict[str, Any]], run_manager: Optional[AsyncCallbackManagerForChainRun] = None) → LLMResult¶
Generate LLM result from inputs.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.ContextQAEvalChain.html
|
b1a8a0631ca2-4
|
Generate LLM result from inputs.
async ainvoke(input: Dict[str, Any], config: Optional[RunnableConfig] = None) → Dict[str, Any]¶
apply(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → List[Dict[str, str]]¶
Utilize the LLM generate method for speed gains.
apply_and_parse(input_list: List[Dict[str, Any]], callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → Sequence[Union[str, List[str], Dict[str, str]]]¶
Call apply and then parse the results.
async apredict(callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, **kwargs: Any) → str¶
Format prompt with kwargs and pass to LLM.
Parameters
callbacks – Callbacks to pass to LLMChain
**kwargs – Keys to pass to prompt template.
Returns
Completion from LLM.
Example
completion = llm.predict(adjective="funny")
async apredict_and_parse(callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, **kwargs: Any) → Union[str, List[str], Dict[str, str]]¶
Call apredict and then parse the results.
async aprep_prompts(input_list: List[Dict[str, Any]], run_manager: Optional[AsyncCallbackManagerForChainRun] = None) → Tuple[List[PromptValue], Optional[List[str]]]¶
Prepare prompts from inputs.
async arun(*args: Any, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, **kwargs: Any) → Any¶
Convenience method for executing chain.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.ContextQAEvalChain.html
|
b1a8a0631ca2-5
|
Convenience method for executing chain.
The main difference between this method and Chain.__call__ is that this
method expects inputs to be passed directly in as positional arguments or
keyword arguments, whereas Chain.__call__ expects a single input dictionary
with all the inputs
Parameters
*args – If the chain expects a single input, it can be passed in as the
sole positional argument.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
**kwargs – If the chain expects multiple inputs, they can be passed in
directly as keyword arguments.
Returns
The chain output.
Example
# Suppose we have a single-input chain that takes a 'question' string:
await chain.arun("What's the temperature in Boise, Idaho?")
# -> "The temperature in Boise is..."
# Suppose we have a multi-input chain that takes a 'question' string
# and 'context' string:
question = "What's the temperature in Boise, Idaho?"
context = "Weather report for Boise, Idaho on 07/03/23..."
await chain.arun(question=question, context=context)
# -> "The temperature in Boise is..."
async astream(input: Input, config: Optional[RunnableConfig] = None) → AsyncIterator[Output]¶
batch(inputs: List[Input], config: Optional[Union[RunnableConfig, List[RunnableConfig]]] = None, *, max_concurrency: Optional[int] = None) → List[Output]¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.ContextQAEvalChain.html
|
b1a8a0631ca2-6
|
bind(**kwargs: Any) → Runnable[Input, Output]¶
Bind arguments to a Runnable, returning a new Runnable.
classmethod construct(_fields_set: Optional[SetStr] = None, **values: Any) → Model¶
Creates a new model setting __dict__ and __fields_set__ from trusted or pre-validated data.
Default values are respected, but no other validation is performed.
Behaves as if Config.extra = ‘allow’ was set since it adds all passed values
copy(*, include: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, exclude: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, update: Optional[DictStrAny] = None, deep: bool = False) → Model¶
Duplicate a model, optionally choose which fields to include, exclude and change.
Parameters
include – fields to include in new model
exclude – fields to exclude from new model, as with values this takes precedence over include
update – values to change/add in the new model. Note: the data is not validated before creating
the new model: you should trust this data
deep – set to True to make a deep copy of the model
Returns
new model instance
create_outputs(llm_result: LLMResult) → List[Dict[str, Any]]¶
Create outputs from response.
dict(**kwargs: Any) → Dict¶
Dictionary representation of chain.
Expects Chain._chain_type property to be implemented and for memory to benull.
Parameters
**kwargs – Keyword arguments passed to default pydantic.BaseModel.dict
method.
Returns
A dictionary representation of the chain.
Example
..code-block:: python
chain.dict(exclude_unset=True)
# -> {“_type”: “foo”, “verbose”: False, …}
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.ContextQAEvalChain.html
|
b1a8a0631ca2-7
|
# -> {“_type”: “foo”, “verbose”: False, …}
evaluate(examples: List[dict], predictions: List[dict], question_key: str = 'query', context_key: str = 'context', prediction_key: str = 'result', *, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None) → List[dict][source]¶
Evaluate question answering examples and predictions.
evaluate_strings(*, prediction: str, reference: Optional[str] = None, input: Optional[str] = None, **kwargs: Any) → dict¶
Evaluate Chain or LLM output, based on optional input and label.
Parameters
prediction (str) – The LLM or chain prediction to evaluate.
reference (Optional[str], optional) – The reference label to evaluate against.
input (Optional[str], optional) – The input to consider during evaluation.
**kwargs – Additional keyword arguments, including callbacks, tags, etc.
Returns
The evaluation results containing the score or value.
Return type
dict
classmethod from_llm(llm: BaseLanguageModel, prompt: Optional[PromptTemplate] = None, **kwargs: Any) → ContextQAEvalChain[source]¶
Load QA Eval Chain from LLM.
Parameters
llm (BaseLanguageModel) – the base language model to use.
prompt ('context' and 'result' that will be used as the) – A prompt template containing the input_variables:
'query' –
prompt –
evaluation. (for) –
PROMPT. (Defaults to) –
**kwargs – additional keyword arguments.
Returns
the loaded QA eval chain.
Return type
ContextQAEvalChain
classmethod from_orm(obj: Any) → Model¶
classmethod from_string(llm: BaseLanguageModel, template: str) → LLMChain¶
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.ContextQAEvalChain.html
|
b1a8a0631ca2-8
|
classmethod from_string(llm: BaseLanguageModel, template: str) → LLMChain¶
Create LLMChain from LLM and template.
generate(input_list: List[Dict[str, Any]], run_manager: Optional[CallbackManagerForChainRun] = None) → LLMResult¶
Generate LLM result from inputs.
invoke(input: Dict[str, Any], config: Optional[RunnableConfig] = None) → Dict[str, Any]¶
json(*, include: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, exclude: Optional[Union[AbstractSetIntStr, MappingIntStrAny]] = None, by_alias: bool = False, skip_defaults: Optional[bool] = None, exclude_unset: bool = False, exclude_defaults: bool = False, exclude_none: bool = False, encoder: Optional[Callable[[Any], Any]] = None, models_as_dict: bool = True, **dumps_kwargs: Any) → unicode¶
Generate a JSON representation of the model, include and exclude arguments as per dict().
encoder is an optional function to supply as default to json.dumps(), other arguments as per json.dumps().
classmethod parse_file(path: Union[str, Path], *, content_type: unicode = None, encoding: unicode = 'utf8', proto: Protocol = None, allow_pickle: bool = False) → Model¶
classmethod parse_obj(obj: Any) → Model¶
classmethod parse_raw(b: Union[str, bytes], *, content_type: unicode = None, encoding: unicode = 'utf8', proto: Protocol = None, allow_pickle: bool = False) → Model¶
predict(callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, **kwargs: Any) → str¶
Format prompt with kwargs and pass to LLM.
Parameters
callbacks – Callbacks to pass to LLMChain
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.ContextQAEvalChain.html
|
b1a8a0631ca2-9
|
Parameters
callbacks – Callbacks to pass to LLMChain
**kwargs – Keys to pass to prompt template.
Returns
Completion from LLM.
Example
completion = llm.predict(adjective="funny")
predict_and_parse(callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, **kwargs: Any) → Union[str, List[str], Dict[str, Any]]¶
Call predict and then parse the results.
prep_inputs(inputs: Union[Dict[str, Any], Any]) → Dict[str, str]¶
Validate and prepare chain inputs, including adding inputs from memory.
Parameters
inputs – Dictionary of raw inputs, or single input if chain expects
only one param. Should contain all inputs specified in
Chain.input_keys except for inputs that will be set by the chain’s
memory.
Returns
A dictionary of all inputs, including those added by the chain’s memory.
prep_outputs(inputs: Dict[str, str], outputs: Dict[str, str], return_only_outputs: bool = False) → Dict[str, str]¶
Validate and prepare chain outputs, and save info about this run to memory.
Parameters
inputs – Dictionary of chain inputs, including any inputs added by chain
memory.
outputs – Dictionary of initial chain outputs.
return_only_outputs – Whether to only return the chain outputs. If False,
inputs are also added to the final outputs.
Returns
A dict of the final chain outputs.
prep_prompts(input_list: List[Dict[str, Any]], run_manager: Optional[CallbackManagerForChainRun] = None) → Tuple[List[PromptValue], Optional[List[str]]]¶
Prepare prompts from inputs.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.ContextQAEvalChain.html
|
b1a8a0631ca2-10
|
Prepare prompts from inputs.
run(*args: Any, callbacks: Optional[Union[List[BaseCallbackHandler], BaseCallbackManager]] = None, tags: Optional[List[str]] = None, metadata: Optional[Dict[str, Any]] = None, **kwargs: Any) → Any¶
Convenience method for executing chain.
The main difference between this method and Chain.__call__ is that this
method expects inputs to be passed directly in as positional arguments or
keyword arguments, whereas Chain.__call__ expects a single input dictionary
with all the inputs
Parameters
*args – If the chain expects a single input, it can be passed in as the
sole positional argument.
callbacks – Callbacks to use for this chain run. These will be called in
addition to callbacks passed to the chain during construction, but only
these runtime callbacks will propagate to calls to other objects.
tags – List of string tags to pass to all callbacks. These will be passed in
addition to tags passed to the chain during construction, but only
these runtime tags will propagate to calls to other objects.
**kwargs – If the chain expects multiple inputs, they can be passed in
directly as keyword arguments.
Returns
The chain output.
Example
# Suppose we have a single-input chain that takes a 'question' string:
chain.run("What's the temperature in Boise, Idaho?")
# -> "The temperature in Boise is..."
# Suppose we have a multi-input chain that takes a 'question' string
# and 'context' string:
question = "What's the temperature in Boise, Idaho?"
context = "Weather report for Boise, Idaho on 07/03/23..."
chain.run(question=question, context=context)
# -> "The temperature in Boise is..."
save(file_path: Union[Path, str]) → None¶
Save the chain.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.ContextQAEvalChain.html
|
b1a8a0631ca2-11
|
save(file_path: Union[Path, str]) → None¶
Save the chain.
Expects Chain._chain_type property to be implemented and for memory to benull.
Parameters
file_path – Path to file to save the chain to.
Example
chain.save(file_path="path/chain.yaml")
classmethod schema(by_alias: bool = True, ref_template: unicode = '#/definitions/{model}') → DictStrAny¶
classmethod schema_json(*, by_alias: bool = True, ref_template: unicode = '#/definitions/{model}', **dumps_kwargs: Any) → unicode¶
stream(input: Input, config: Optional[RunnableConfig] = None) → Iterator[Output]¶
to_json() → Union[SerializedConstructor, SerializedNotImplemented]¶
to_json_not_implemented() → SerializedNotImplemented¶
classmethod update_forward_refs(**localns: Any) → None¶
Try to update ForwardRefs on fields based on this Model, globalns and localns.
classmethod validate(value: Any) → Model¶
with_fallbacks(fallbacks: ~typing.Sequence[~langchain.schema.runnable.Runnable[~langchain.schema.runnable.Input, ~langchain.schema.runnable.Output]], *, exceptions_to_handle: ~typing.Tuple[~typing.Type[BaseException]] = (<class 'Exception'>,)) → RunnableWithFallbacks[Input, Output]¶
property evaluation_name: str¶
The name of the evaluation.
property lc_attributes: Dict¶
Return a list of attribute names that should be included in the
serialized kwargs. These attributes must be accepted by the
constructor.
property lc_namespace: List[str]¶
Return the namespace of the langchain object.
eg. [“langchain”, “llms”, “openai”]
property lc_secrets: Dict[str, str]¶
Return a map of constructor argument names to secret ids.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.ContextQAEvalChain.html
|
b1a8a0631ca2-12
|
Return a map of constructor argument names to secret ids.
eg. {“openai_api_key”: “OPENAI_API_KEY”}
property lc_serializable: bool¶
Return whether or not the class is serializable.
property requires_input: bool¶
Whether the chain requires an input string.
property requires_reference: bool¶
Whether the chain requires a reference string.
Examples using ContextQAEvalChain¶
Question Answering
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.qa.eval_chain.ContextQAEvalChain.html
|
1720aa23539b-0
|
langchain.evaluation.criteria.eval_chain.CriteriaEvalChain¶
class langchain.evaluation.criteria.eval_chain.CriteriaEvalChain[source]¶
Bases: StringEvaluator, LLMEvalChain, LLMChain
LLM Chain for evaluating runs against criteria.
Parameters
llm (BaseLanguageModel) – The language model to use for evaluation.
criteria (Union[Mapping[str, str]]) – The criteriaor rubric to evaluate the runs against. It can be a mapping of
criterion name to its sdescription, or a single criterion name.
prompt (Optional[BasePromptTemplate], default=None) – The prompt template to use for generating prompts. If not provided, a
default prompt template will be used based on the value of
requires_reference.
requires_reference (bool, default=False) – Whether the evaluation requires a reference text. If True, the
PROMPT_WITH_REFERENCES template will be used, which includes the
reference labels in the prompt. Otherwise, the PROMPT template will be
used, which is a reference-free prompt.
**kwargs (Any) – Additional keyword arguments to pass to the LLMChain constructor.
Returns
An instance of the CriteriaEvalChain class.
Return type
CriteriaEvalChain
Examples
>>> from langchain.chat_models import ChatAnthropic
>>> from langchain.evaluation.criteria import CriteriaEvalChain
>>> llm = ChatAnthropic(temperature=0)
>>> criteria = {"my-custom-criterion": "Is the submission the most amazing ever?"}
>>> evaluator = CriteriaEvalChain.from_llm(llm=llm, criteria=criteria)
>>> evaluator.evaluate_strings(prediction="Imagine an ice cream flavor for the color aquamarine", input="Tell me an idea")
{
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.criteria.eval_chain.CriteriaEvalChain.html
|
1720aa23539b-1
|
{
'reasoning': 'Here is my step-by-step reasoning for the given criteria:\n\nThe criterion is: "Is the submission the most amazing ever?" This is a subjective criterion and open to interpretation. The submission suggests an aquamarine-colored ice cream flavor which is creative but may or may not be considered the most amazing idea ever conceived. There are many possible amazing ideas and this one ice cream flavor suggestion may or may not rise to that level for every person. \n\nN',
'value': 'N',
'score': 0,
}
>>> from langchain.chat_models import ChatOpenAI
>>> from langchain.evaluation.criteria import LabeledCriteriaEvalChain
>>> llm = ChatOpenAI(model="gpt-4", temperature=0)
>>> criteria = "correctness"
>>> evaluator = LabeledCriteriaEvalChain.from_llm(
... llm=llm,
... criteria=criteria,
... )
>>> evaluator.evaluate_strings(
... prediction="The answer is 4",
... input="How many apples are there?",
... reference="There are 3 apples",
... )
{
'score': 0,
'reasoning': 'The criterion for this task is the correctness of the submission. The submission states that there are 4 apples, but the reference indicates that there are actually 3 apples. Therefore, the submission is not correct, accurate, or factual according to the given criterion.\n\nN',
'value': 'N',
}
param callback_manager: Optional[BaseCallbackManager] = None¶
Deprecated, use callbacks instead.
param callbacks: Callbacks = None¶
Optional list of callback handlers (or callback manager). Defaults to None.
Callback handlers are called throughout the lifecycle of a call to a chain,
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.criteria.eval_chain.CriteriaEvalChain.html
|
1720aa23539b-2
|
Callback handlers are called throughout the lifecycle of a call to a chain,
starting with on_chain_start, ending with on_chain_end or on_chain_error.
Each custom chain can optionally call additional callback methods, see Callback docs
for full details.
param criterion_name: str [Required]¶
The name of the criterion being evaluated.
param llm: BaseLanguageModel [Required]¶
Language model to call.
param llm_kwargs: dict [Optional]¶
param memory: Optional[BaseMemory] = None¶
Optional memory object. Defaults to None.
Memory is a class that gets called at the start
and at the end of every chain. At the start, memory loads variables and passes
them along in the chain. At the end, it saves any returned variables.
There are many different types of memory - please see memory docs
for the full catalog.
param metadata: Optional[Dict[str, Any]] = None¶
Optional metadata associated with the chain. Defaults to None.
This metadata will be associated with each call to this chain,
and passed as arguments to the handlers defined in callbacks.
You can use these to eg identify a specific instance of a chain with its use case.
param output_parser: BaseOutputParser [Optional]¶
The parser to use to map the output to a structured result.
param prompt: BasePromptTemplate [Required]¶
Prompt object to use.
param return_final_only: bool = True¶
Whether to return only the final parsed result. Defaults to True.
If false, will return a bunch of extra information about the generation.
param tags: Optional[List[str]] = None¶
Optional list of tags associated with the chain. Defaults to None.
These tags will be associated with each call to this chain,
and passed as arguments to the handlers defined in callbacks.
|
https://api.python.langchain.com/en/latest/evaluation/langchain.evaluation.criteria.eval_chain.CriteriaEvalChain.html
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.