litellm.providers.core — tech.unravel/litellm-clj 0.3.0-alpha

Core provider protocol and utilities for LiteLLM

authentication-error clj

(authentication-error provider)

Create an authentication error (DEPRECATED - use litellm.errors/authentication-error)

calculate-cost clj

(calculate-cost provider-name model prompt-tokens completion-tokens)

Calculate cost for a request/response

compare-providers clj

(compare-providers provider-names)

Compare multiple providers on various capabilities

create-standard-choice clj

(create-standard-choice index message finish-reason)

Create a standard choice format

create-standard-message clj

(create-standard-message role content & {:keys [tool-calls function-call]})

Create a standard message format

create-standard-response clj

(create-standard-response id model choices usage & {:keys [object created]})

Create a standard response format

create-standard-usage clj

(create-standard-usage prompt-tokens completion-tokens)

Create a standard usage format

default-provider-config clj

(default-provider-config provider-name)

Get default configuration for a provider

estimate-request-tokens clj

(estimate-request-tokens request)

Estimate token count for a request

estimate-tokens clj

(estimate-tokens text)

Rough estimation of token count for text

extract-model-name clj

(extract-model-name model)

Extract actual model name from model string (e.g., 'openai/gpt-4' -> 'gpt-4')

extract-provider-name clj

(extract-provider-name model)

Extract provider name from model string (e.g., 'openai/gpt-4' -> :openai)

find-providers-for-model clj

(find-providers-for-model provider-names model)

Find all providers that support a given model

get-cost-per-token clj multimethod

Get cost per token for a specific model

get-rate-limits clj multimethod

Get provider rate limits as a map

health-check clj multimethod

Perform health check, returns a future with boolean result

list-available-providers clj

(list-available-providers)

List all available providers by checking multimethod implementations

log-provider-status clj

(log-provider-status provider-name)

Log provider status for debugging

make-request clj multimethod

Make HTTP request to provider API, returns a future

make-streaming-request clj multimethod

Make streaming HTTP request to provider API, returns a core.async channel

merge-provider-config clj

(merge-provider-config provider-name user-config)

Merge user config with provider defaults

model-not-found-error clj

(model-not-found-error provider model)

Create a model not found error (DEPRECATED - use litellm.errors/model-not-found)

parse-model-string clj

(parse-model-string model)

Parse model string into provider (keyword) and model components

provider-available? clj

(provider-available? provider-name)

Check if a provider is available (has multimethod implementations)

provider-error clj

(provider-error provider message & {:keys [status code data]})

Create a provider-specific error (DEPRECATED - use litellm.errors/provider-error)

provider-status clj

(provider-status provider-name)

Get comprehensive provider status

provider-supports-model? clj

(provider-supports-model? provider-name model)

Check if provider supports a specific model

quota-exceeded-error clj

(quota-exceeded-error provider)

Create a quota exceeded error (DEPRECATED - use litellm.errors/quota-exceeded)

rate-limit-error clj

(rate-limit-error provider & {:keys [retry-after]})

Create a rate limit error (DEPRECATED - use litellm.errors/rate-limit)

supports-function-calling? clj multimethod

Check if provider supports function calling

supports-streaming? clj multimethod

Check if provider supports streaming responses

test-provider clj

(test-provider provider-name thread-pool telemetry config)

Test provider with a simple request

transform-messages-from-openai clj

(transform-messages-from-openai messages)

Transform messages from OpenAI format

transform-messages-to-openai clj

(transform-messages-to-openai messages)

Transform messages to OpenAI format

transform-request clj multimethod

Transform a standard request to provider-specific format

transform-response clj multimethod

Transform provider response to standard format

transform-streaming-chunk clj multimethod

Transform provider streaming chunk to standard format

transform-tools-to-openai clj

(transform-tools-to-openai tools)

Transform tools to OpenAI format

validate-provider-config clj

(validate-provider-config config)

Validate provider configuration

validate-request clj

(validate-request provider-name request)

Validate request against provider capabilities