8ae5a3b61a
Co-authored-by: Antonio <antonio@zed.dev> Resurrected this from some assistant work I did in Spring of 2023. - [x] Resurrect streaming responses - [x] Use streaming responses to enable AI via Zed's servers by default (but preserve API key option for now) - [x] Simplify protobuf - [x] Proxy to OpenAI on zed.dev - [x] Proxy to Gemini on zed.dev - [x] Improve UX for switching between openAI and google models - We current disallow cycling when setting a custom model, but we need a better solution to keep OpenAI models available while testing the google ones - [x] Show remaining tokens correctly for Google models - [x] Remove semantic index - [x] Delete `ai` crate - [x] Cloud front so we can ban abuse - [x] Rate-limiting - [x] Fix panic when using inline assistant - [x] Double check the upgraded `AssistantSettings` are backwards-compatible - [x] Add hosted LLM interaction behind a `language-models` feature flag. Release Notes: - We are temporarily removing the semantic index in order to redesign it from scratch. --------- Co-authored-by: Antonio <antonio@zed.dev> Co-authored-by: Antonio Scandurra <me@as-cii.com> Co-authored-by: Thorsten <thorsten@zed.dev> Co-authored-by: Max <max@zed.dev>
75 lines
2.7 KiB
Rust
75 lines
2.7 KiB
Rust
use anyhow::{anyhow, Result};
|
|
use rpc::proto;
|
|
|
|
pub fn language_model_request_to_open_ai(
|
|
request: proto::CompleteWithLanguageModel,
|
|
) -> Result<open_ai::Request> {
|
|
Ok(open_ai::Request {
|
|
model: open_ai::Model::from_id(&request.model).unwrap_or(open_ai::Model::FourTurbo),
|
|
messages: request
|
|
.messages
|
|
.into_iter()
|
|
.map(|message| {
|
|
let role = proto::LanguageModelRole::from_i32(message.role)
|
|
.ok_or_else(|| anyhow!("invalid role {}", message.role))?;
|
|
Ok(open_ai::RequestMessage {
|
|
role: match role {
|
|
proto::LanguageModelRole::LanguageModelUser => open_ai::Role::User,
|
|
proto::LanguageModelRole::LanguageModelAssistant => {
|
|
open_ai::Role::Assistant
|
|
}
|
|
proto::LanguageModelRole::LanguageModelSystem => open_ai::Role::System,
|
|
},
|
|
content: message.content,
|
|
})
|
|
})
|
|
.collect::<Result<Vec<open_ai::RequestMessage>>>()?,
|
|
stream: true,
|
|
stop: request.stop,
|
|
temperature: request.temperature,
|
|
})
|
|
}
|
|
|
|
pub fn language_model_request_to_google_ai(
|
|
request: proto::CompleteWithLanguageModel,
|
|
) -> Result<google_ai::GenerateContentRequest> {
|
|
Ok(google_ai::GenerateContentRequest {
|
|
contents: request
|
|
.messages
|
|
.into_iter()
|
|
.map(language_model_request_message_to_google_ai)
|
|
.collect::<Result<Vec<_>>>()?,
|
|
generation_config: None,
|
|
safety_settings: None,
|
|
})
|
|
}
|
|
|
|
pub fn language_model_request_message_to_google_ai(
|
|
message: proto::LanguageModelRequestMessage,
|
|
) -> Result<google_ai::Content> {
|
|
let role = proto::LanguageModelRole::from_i32(message.role)
|
|
.ok_or_else(|| anyhow!("invalid role {}", message.role))?;
|
|
|
|
Ok(google_ai::Content {
|
|
parts: vec![google_ai::Part::TextPart(google_ai::TextPart {
|
|
text: message.content,
|
|
})],
|
|
role: match role {
|
|
proto::LanguageModelRole::LanguageModelUser => google_ai::Role::User,
|
|
proto::LanguageModelRole::LanguageModelAssistant => google_ai::Role::Model,
|
|
proto::LanguageModelRole::LanguageModelSystem => google_ai::Role::User,
|
|
},
|
|
})
|
|
}
|
|
|
|
pub fn count_tokens_request_to_google_ai(
|
|
request: proto::CountTokensWithLanguageModel,
|
|
) -> Result<google_ai::CountTokensRequest> {
|
|
Ok(google_ai::CountTokensRequest {
|
|
contents: request
|
|
.messages
|
|
.into_iter()
|
|
.map(language_model_request_message_to_google_ai)
|
|
.collect::<Result<Vec<_>>>()?,
|
|
})
|
|
}
|