OPEA API Service Spec (v0.9)¶

Authors:

This specification is used to define the Restful API of OPEA Mega Service for users to access, as long as the input and output definition of all OPEA Micro Services for developer to build OPEA Mega service.

Note

This API Service Specification is a work-in-progress and may be incomplete and contain errors.

OPEA Mega Service API ¶

OPEA Mega Service is the main entry user can access for a prebuilt GenAI application. Such GenAI application consists of single or several OPEA Micro Services chained as a DAG (Directed Acyclic Graph) and built as an execution workflow for developer to create complex applications.

List Services ¶

List all supported services by the OPEA Mega Service.

Request¶

Method	URL
GET	`/v1/list_service`

Response¶

Status

Response

200

{
   <service_name>: <service_description>
}

service_name (string)

The endpoints or URLs OPEA mega service is serving. For example, /v1/RAG.

Note some keywords such as /v1/audio/speech, /v1/audio/transcriptions, /v1/embeddings, /v1/chat/completions are reserved for openAI compatible Mega Service.

service_description (string)

The detail usage description user used to access the specified endpoints or urls OPEA mega service is serving, including the request and post format and details.

405

{"error": "Retrieve service name wrongly."}

List Configurable Parameters ¶

List all configurable parameters for users to control the behavior of the OPEA Mega Service.

Request¶

Method	URL
GET	`/v1/list_parameters`

Response¶

Status

Response

200

{
   <micro_service_name>:
   {
      <parameter_name>: data_type,
      . . .
   }
}

micro_service_name (string): The micro service name in OPEA mega service in which some parameters are configurable.
parameter_name (string): The configurable parameter name in OPEA mega service.
data_type (string): The supported data type: "string" or "integer".

For example: {"/v1/llm_generate": {"max_tokens": "integer"}}

405

{"error": "Retrieve configurable parameter wrongly."}

Embedding ¶

Optional. Only exists if a single OPEA microservice which exposes /micro_service/embedding interface is built as OPEA Mega service.

Request¶

Method	URL
POST	`/v1/list_parameters`

Type	Parameters	Values	Required	Description
POST	`input`	`string`	required	Input text to embed, encoded as a string or array of tokens. To embed multiple inputs in a single request, pass an array of strings or array of token arrays. The input must not exceed the max input tokens for the model (8192 tokens for text-embedding-ada-002), cannot be an empty string, and any array must be 2048 dimensions or less.
POST	`model`	`string`	deprecated	The ID of the model to use.
POST	`encoding_format`	`string`	required	The format to return the embeddings in. Can be either `"float"` or `"base64"`.
POST	`dimensions`	`integer`	optional	The number of dimensions the resulting output embeddings should have. Only supported in text-embedding-3 and later models.

Response¶

Status

Response

200

{
   "object": "list",
   "data": [{
      "object": "embedding",
      "embedding": [
         0.0023064255,
         ...
      ],
      "index": 0
   }],
   "model": "text-embedding-ada-002",
   "usage": {
      "prompt_tokens": 8,
      "total_tokens": 8
   },
}

embedding (float): The vector representation for given inputs.
index (integer): The index of the embedding in the list of embeddings.
parameter_name (string): The configurable parameter name in OPEA mega service.
data_type (string): The supported data type, "string" or "integer".

For example: {"llm": {"max_tokens": "integer"}}

405

{"error": "Retrieve configurable parameter wrongly."}

Chat ¶

Optional. . If a OPEA Mega service is built with this request url, it complies with below format.

Request¶

Method	URL
POST	`/v1/chat/completions`

Type	Parameters	Values	Required	Description
POST	`message`	`array`	required	A list of messages comprising the conversation. Refer to the detail format.
POST	`model`	`string`	deprecated	The ID of the model to use.
POST	`frequency_penalty`	`integer`	optional	Number between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model’s likelihood to repeat the same line verbatim.
POST	`logit_bias`	`map`	optional	Modify the likelihood of specified tokens appearing in the completion. Accepts a JSON object that maps tokens (specified by their token ID in the tokenizer) to an associated bias value from -100 to 100. Mathematically, the bias is added to the logits generated by the model prior to sampling. The exact effect will vary per model, but values between -1 and 1 should decrease or increase likelihood of selection; values like -100 or 100 should result in a ban or exclusive selection of the relevant token.
POST	`logprobs`	`bool`	optional
POST	`top_logprobs`	`integer`	optional
POST	`max_tokens`	`integer`	optional
POST	`n`	`integer`	optional
POST	`presence_penalty`	`float`	optional
POST	`response_format`	`object`	optional
POST	`seed`	`integer`	optional
POST	`stop`	`string`	optional
POST	`stream`	`bool`	optional
POST	`stream_options`	`object`	optional
POST	`temperature`	`float`	optional
POST	`top_p`	`float`	optional
POST	`tools`	`array`	optional
POST	`tool_choice`	`string`	optional
POST	`user`	`string`	optional

Response¶

Status

Response

200

{
   "id": "chatcmpl-123",
   "object": "chat.completion",
   "created": 1677652288,
   "model": "gpt-3.5-turbo-0125",
   "system_fingerprint": "fp_44709d6fcb",

   "choices": [{
      "index": 0,
      "object": "embedding",
      "message": {
         "role": "assistant",
         "content": "\n\nHello there, how may I assist you today?",
      },
      "logprobs": null,
      "finish_reason": "stop",
   }],

   "usage": {
      "prompt_tokens": 9,
      "completion_tokens": 12,
      "total_tokens": 21
   },
}

id (string): A unique identifier for the chat completion.
choices (array): A list of chat completion choices. Can be more than one if n is greater than 1.
created (integer): The Unix timestamp (in seconds) of when the chat completion was created.
model (string): The model used for the chat completion.
system_fingerprint (string): This fingerprint represents the backend configuration that the model runs with. Can be used in conjunction with the seed request parameter to understand when backend changes have been made that might impact determinism.
object (string): The object type, which is always "chat.completion".
usage (object): Usage statistics for the completion request.

Other Operations ¶

Check the usage description returned in List Services to know what other operations are supported by this OPEA Mega Service.

OPEA Micro Service API ¶

OPEA Micro Service is the building block of constructing any GenAI applications. The API in OPEA micro service is used by developers to construct OPEA Mega Service like a DAG chain and is invisible for end user.

Embedding Micro Service ¶

The micro service is used to generate a vector representation of a given input.

Request¶

Method	URL
POST	`/v1/embeddings`

Type	Parameters	Values	Required	Description
POST	`input`	`string`	required	Input text to embed, encoded as a string or array of tokens. To embed multiple inputs in a single request, pass an array of strings or array of token arrays. The input must not exceed the max input tokens for the model (8192 tokens for text-embedding-ada-002), cannot be an empty string, and any array must be 2048 dimensions or less
POST	`model`	`string`	required	The ID of the model to use.
POST	`encoding_format`	`string`	optional	The format to return the embeddings in. Can be either `"float"` or `"base64"`. Devault to `"float"`.
POST	`dimensions`	`integer`	optional	The number of dimensions the resulting output embeddings should have.
POST	`user`	`string`	optional	A unique identifier representing your end-user, which can help OpenAI to monitor and detect abuse.

Response¶

Status

Response

200

{
   "object": "list",
   "data": [{
      "object": "embedding",
      "embedding": [
         0.0023064255,
         -0.009327292,
         . . . (1536 floats total for ada-002)
         -0.0028842222,
      ],
      "index": 0
   }],
   "model": "text-embedding-ada-002",
   "usage": {
      "prompt_tokens": 8,
      "total_tokens": 8
   },
}

embedding (list of float): The vector representation for given inputs.

405

{"error": "The request of getting embedding vector fails."}

LLM Generation Micro Service ¶

The micro service is used to provide LLM generation service.

Request¶

Method	URL
POST	`/v1/chat/completions`

Type	Parameters	Values	Required	Description
POST	`message`	`array`	required	A list of messages comprising the conversation so far. Example Python code.
POST	`model`	`string`	required	The ID of the model to use. See the model endpoint compatibility table for details on which models work with the Chat API.
POST	`frequency_penalty`	`float`	optional	Number between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model’s likelihood to repeat the same line verbatim.
POST	`logit_bias`	`map`	optional	Modify the likelihood of specified tokens appearing in the completion.Accepts a JSON object that maps tokens (specified by their token ID in the tokenizer) to an associated bias value from -100 to 100. Mathematically, the bias is added to the logits generated by the model prior to sampling. The exact effect will vary per model, but values between -1 and 1 should decrease or increase likelihood of selection; values like -100 or 100 should result in a ban or exclusive selection of the relevant token.
POST	`logprobs`	`bool`	optional	Whether to return log probabilities of the output tokens or not. If true, returns the log probabilities of each output token returned in the content of message.
POST	`top_logprobs`	`integer`	optional	An integer between 0 and 20 specifying the number of most likely tokens to return at each token position, each with an associated log probability. `logprobs` must be set to true if this parameter is used.
POST	`max_tokens`	`integer`	optional	The maximum number of tokens that can be generated in the chat completion.The total length of input tokens and generated tokens is limited by the model’s context length. Example Python code for counting tokens.
POST	`n`	`integer`	optional	How many chat completion choices to generate for each input message. Note that you will be charged based on the number of generated tokens across all of the choices. Keep n as 1 to minimize costs.
POST	`presence_penalty`	`float`	optional
POST	`response_format`	`object`	optional
POST	`seed`	`integer`	optional	This feature is in Beta. If specified, our system will make a best effort to sample deterministically, such that repeated requests with the same seed and parameters should return the same result. Determinism is not guaranteed, and you should refer to the `system_fingerprint` response parameter to monitor changes in the backend.
POST	`service_tier`	`string`	optional	Specifies the latency tier to use for processing the request. This parameter is relevant for customers subscribed to the scale tier service. If set to `"auto"`, the system will utilize scale tier credits until they are exhausted. If set to `"default"`, the request will be processed using the default service tier with a lower uptime SLA and no latency guarentee. When this parameter is set, the response body will include the `service_tier` utilized.
POST	`stop`	`string`	optional	Up to 4 sequences where the API will stop generating further tokens.
POST	`stream`	`bool`	optional	If set, partial message deltas will be sent, like in ChatGPT. Tokens will be sent as data-only server-sent events as they become available, with the stream terminated by a data: `[DONE]` message. Example Python code.
POST	`stream_options`	`object`	optional	Options for streaming response. Only set this when you set `"stream": "true"`.
POST	`temperature`	`float`	optional	What sampling temperature to use, between 0 and 2. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic. We generally recommend altering this or `top_p` but not both.
POST	`top_p`	`float`	optional	An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with `top_p` probability mass. So 0.1 means only the tokens comprising the top 10% probability mass are considered. We generally recommend altering this or `temperature` but not both.
POST	`tools`	`array`	optional	A list of tools the model may call. Currently, only functions are supported as a tool. Use this to provide a list of functions the model may generate JSON inputs for. A max of 128 functions are supported.
POST	`tool_choice`	`string`	optional	Controls which (if any) tool is called by the model. `"none"` means the model will not call any tool and instead generates a message. `"auto"` means the model can pick between generating a message or calling one or more tools. required means the model must call one or more tools. Specifying a particular tool via `{"type": "function", "function": {"name": "my_function"}}` forces the model to call that tool. `"none"` is the default when no tools are present. `"auto"` is the default if tools are present.

Response¶

Status

Response

200

{
   "id": "chatcmpl-123",
   "object": "chat.completion",
   "created": 1677652288,
   "model": "gpt-4o-mini",
   "system_fingerprint": "fp_44709d6fcb",
   "choices": [{
      "index": 0,
      "object": "embedding",
      "message": {
         "role": "assistant",
         "content": "\n\nHello there, how may I assist you today?",
      },
      "logprobs": null,
      "finish_reason": "stop",
   }],

   "usage": {
      "prompt_tokens": 9,
      "completion_tokens": 12,
      "total_tokens": 21
   },
}

405

{"error": "The request of LLM generation fails."}

ASR Micro Service ¶

The micro service is used to provide audio to text service.

Request¶

Method	URL
POST	`/v1/asr`

Type	Parameters	Values	Required	Description
POST	`url`	`docarray.AudioUrl`	optional	The link to the audio.
POST	`model_name_or_path`	`string`	optional	The model used to do audio-to-text translation.
POST	`Language`	`string`	optional	The language that model prefer to detect. Default is `"auto"`.

Response¶

Status	Response
200	{ "text": string }
405	`{"error": "The request of ASR fails."}`

RAG Retrieval Micro Service ¶

The micro service is used to provide RAG retrieval service. It’s usually after embedding micro sevice and before RAG reranking micro service to build a RAG Mega service.

Request¶

Method	URL
POST	`/v1/rag_retrieval`

Type	Parameters	Values	Required	Description
POST	`text`	`string`	required	The input string to query.
POST	`embedding`	`list of float`	required	The list of float for text as vector representation.

Response¶

Status	Response
200	{ "retrieved_docs": list of string, "initial_query": string, "json_encoders": [{ "text": "I am the agent of chatbot. What can I do for you?", }, ... ] }
405	`{"error": "The request of ASR fails."}`

RAG Reranking Micro Service ¶

The micro service is used to provide RAG reranking service. It’s usually after RAG retrieval and before LLM generation micro service.

Request¶

Method	URL
POST	`/v1/rag_reranking`

Type	Parameters	Values	Required	Description
POST	`retrieved_docs`	`list of string`	required	The docs to be retreived.
POST	`initial_query`	`string`	required	The string to query.
POST	`json_encoders`	`list of float`	required	The json encoder used.

Response¶

Status	Response
200	{ "query": string, "doc": [{ "text": "I am the agent of chatbot. What can I do for you?", }, ... ] }
405	`{"error": "The request of ASR fails."}`