Upstream spec sync (64bit#220)

* add project id in OpenAIConfig

* update audio tranlations: include json and verbose json functions

* add fine tuning list checkpoints endpoint

* update moderations comment

* assistants=v2 header

* add delete thread message api

* add vector stores api: create, retrive, list, update, delete

* VectorStores API group in client

* vector stores and vector store files

* vector store file batches API

* expose VectorStoreFileBatches

* Batches API

* update ApiError type to match spec

* updated from spec upto line 9601

* assistant api update; up to 10055

* update RunObject; up to 10327

* updates to Run and Thread objects and requests

* MessageObject updated

* message delta object

* run step delta object and related

* update step objects

* BatchRequestInput and BatchRequestOutput types

* example fix

* updated spec from upstream

* update README

* fix completions example

* fix completions streaming example

* fix models test

* fix AssistantsApiResponseFormatOption

* fix MessageObject

* have errors log in assistants example

* fix metadata type in batch.rs types

* fix metadata type in vector_store.rs type

Author: 64bit

async-openai/README.md

@@ -23,8 +23,10 @@
 
 - It's based on [OpenAI OpenAPI spec](https://github.com/openai/openai-openapi)
 - Current features:
-  - [x] Assistants (Beta)
+  - [x] Assistants v2
+  - [ ] Assistants v2 streaming
   - [x] Audio
+  - [x] Batch
   - [x] Chat
   - [x] Completions (Legacy)
   - [x] Embeddings
@@ -112,7 +114,7 @@ A good starting point would be to look at existing [open issues](https://github.
 
 To maintain quality of the project, a minimum of the following is a must for code contribution:
 - **Names & Documentation**: All struct names, field names and doc comments are from OpenAPI spec. Nested objects in spec without names leaves room for making appropriate name.
-- **Tested**: Examples are primary means of testing and should continue to work. For new features supporting example is required.
+- **Tested**:  For changes supporting test(s) and/or example is required. Existing examples, doc tests, unit tests, and integration tests should be made to work with the changes if applicable.
 - **Scope**: Keep scope limited to APIs available in official documents such as [API Reference](https://platform.openai.com/docs/api-reference) or [OpenAPI spec](https://github.com/openai/openai-openapi/). Other LLMs or AI Providers offer OpenAI-compatible APIs, yet they may not always have full parity. In such cases, the OpenAI spec takes precedence.
 - **Consistency**: Keep code style consistent across all the "APIs" that library exposes; it creates a great developer experience.
 

async-openai/src/audio.rs

@@ -6,7 +6,8 @@ use crate::{
     types::{
         CreateSpeechRequest, CreateSpeechResponse, CreateTranscriptionRequest,
         CreateTranscriptionResponseJson, CreateTranscriptionResponseVerboseJson,
-        CreateTranslationRequest, CreateTranslationResponse,
+        CreateTranslationRequest, CreateTranslationResponseJson,
+        CreateTranslationResponseVerboseJson,
     },
     Client,
 };
@@ -52,11 +53,19 @@ impl<'c, C: Config> Audio<'c, C> {
             .await
     }
 
-    /// Translates audio into into English.
+    /// Translates audio into English.
     pub async fn translate(
         &self,
         request: CreateTranslationRequest,
-    ) -> Result<CreateTranslationResponse, OpenAIError> {
+    ) -> Result<CreateTranslationResponseJson, OpenAIError> {
+        self.client.post_form("/audio/translations", request).await
+    }
+
+    /// Translates audio into English.
+    pub async fn translate_verbose_json(
+        &self,
+        request: CreateTranslationRequest,
+    ) -> Result<CreateTranslationResponseVerboseJson, OpenAIError> {
         self.client.post_form("/audio/translations", request).await
     }
 

async-openai/src/batches.rs

@@ -0,0 +1,49 @@
+use serde::Serialize;
+
+use crate::{
+    config::Config,
+    error::OpenAIError,
+    types::{Batch, BatchRequest, ListBatchesResponse},
+    Client,
+};
+
+/// Create large batches of API requests for asynchronous processing. The Batch API returns completions within 24 hours for a 50% discount.
+///
+/// Related guide: [Batch](https://platform.openai.com/docs/guides/batch)
+pub struct Batches<'c, C: Config> {
+    client: &'c Client<C>,
+}
+
+impl<'c, C: Config> Batches<'c, C> {
+    pub fn new(client: &'c Client<C>) -> Self {
+        Self { client }
+    }
+
+    /// Creates and executes a batch from an uploaded file of requests
+    pub async fn create(&self, request: BatchRequest) -> Result<Batch, OpenAIError> {
+        self.client.post("/batches", request).await
+    }
+
+    /// List your organization's batches.
+    pub async fn list<Q>(&self, query: &Q) -> Result<ListBatchesResponse, OpenAIError>
+    where
+        Q: Serialize + ?Sized,
+    {
+        self.client.get_with_query("/batches", query).await
+    }
+
+    /// Retrieves a batch.
+    pub async fn retrieve(&self, batch_id: &str) -> Result<Batch, OpenAIError> {
+        self.client.get(&format!("/batches/{batch_id}")).await
+    }
+
+    /// Cancels an in-progress batch.
+    pub async fn cancel(&self, batch_id: &str) -> Result<Batch, OpenAIError> {
+        self.client
+            .post(
+                &format!("/batches/{batch_id}/cancel"),
+                serde_json::json!({}),
+            )
+            .await
+    }
+}

async-openai/src/client.rs

@@ -11,7 +11,8 @@ use crate::{
     file::Files,
     image::Images,
     moderation::Moderations,
-    Assistants, Audio, Chat, Completions, Embeddings, FineTuning, Models, Threads,
+    Assistants, Audio, Batches, Chat, Completions, Embeddings, FineTuning, Models, Threads,
+    VectorStores,
 };
 
 #[derive(Debug, Clone)]
@@ -128,6 +129,16 @@ impl<C: Config> Client<C> {
         Threads::new(self)
     }
 
+    /// To call [VectorStores] group related APIs using this client.
+    pub fn vector_stores(&self) -> VectorStores<C> {
+        VectorStores::new(self)
+    }
+
+    /// To call [Batches] group related APIs using this client.
+    pub fn batches(&self) -> Batches<C> {
+        Batches::new(self)
+    }
+
     pub fn config(&self) -> &C {
         &self.config
     }

async-openai/src/config.rs

@@ -5,8 +5,10 @@ use serde::Deserialize;
 
 /// Default v1 API base url
 pub const OPENAI_API_BASE: &str = "https://api.openai.com/v1";
-/// Name for organization header
+/// Organization header
 pub const OPENAI_ORGANIZATION_HEADER: &str = "OpenAI-Organization";
+/// Project header
+pub const OPENAI_PROJECT_HEADER: &str = "OpenAI-Project";
 
 /// Calls to the Assistants API require that you pass a Beta header
 pub const OPENAI_BETA_HEADER: &str = "OpenAI-Beta";
@@ -30,6 +32,7 @@ pub struct OpenAIConfig {
     api_base: String,
     api_key: Secret<String>,
     org_id: String,
+    project_id: String,
 }
 
 impl Default for OpenAIConfig {
@@ -40,6 +43,7 @@ impl Default for OpenAIConfig {
                 .unwrap_or_else(|_| "".to_string())
                 .into(),
             org_id: Default::default(),
+            project_id: Default::default(),
         }
     }
 }
@@ -56,6 +60,12 @@ impl OpenAIConfig {
         self
     }
 
+    /// Non default project id
+    pub fn with_project_id<S: Into<String>>(mut self, project_id: S) -> Self {
+        self.project_id = project_id.into();
+        self
+    }
+
     /// To use a different API key different from default OPENAI_API_KEY env var
     pub fn with_api_key<S: Into<String>>(mut self, api_key: S) -> Self {
         self.api_key = Secret::from(api_key.into());
@@ -83,6 +93,13 @@ impl Config for OpenAIConfig {
             );
         }
 
+        if !self.project_id.is_empty() {
+            headers.insert(
+                OPENAI_PROJECT_HEADER,
+                self.project_id.as_str().parse().unwrap(),
+            );
+        }
+
         headers.insert(
             AUTHORIZATION,
             format!("Bearer {}", self.api_key.expose_secret())
@@ -93,7 +110,7 @@ impl Config for OpenAIConfig {
 
         // hack for Assistants APIs
         // Calls to the Assistants API require that you pass a Beta header
-        headers.insert(OPENAI_BETA_HEADER, "assistants=v1".parse().unwrap());
+        headers.insert(OPENAI_BETA_HEADER, "assistants=v2".parse().unwrap());
 
         headers
     }

async-openai/src/error.rs

@@ -32,8 +32,8 @@ pub enum OpenAIError {
 pub struct ApiError {
     pub message: String,
     pub r#type: Option<String>,
-    pub param: Option<serde_json::Value>,
-    pub code: Option<serde_json::Value>,
+    pub param: Option<String>,
+    pub code: Option<String>,
 }
 
 /// Wrapper to deserialize the error object nested in "error" JSON key

async-openai/src/fine_tuning.rs

@@ -4,8 +4,8 @@ use crate::{
     config::Config,
     error::OpenAIError,
     types::{
-        CreateFineTuningJobRequest, FineTuningJob, ListFineTuningJobEventsResponse,
-        ListPaginatedFineTuningJobsResponse,
+        CreateFineTuningJobRequest, FineTuningJob, ListFineTuningJobCheckpointsResponse,
+        ListFineTuningJobEventsResponse, ListPaginatedFineTuningJobsResponse,
     },
     Client,
 };
@@ -80,4 +80,20 @@ impl<'c, C: Config> FineTuning<'c, C> {
             )
             .await
     }
+
+    pub async fn list_checkpoints<Q>(
+        &self,
+        fine_tuning_job_id: &str,
+        query: &Q,
+    ) -> Result<ListFineTuningJobCheckpointsResponse, OpenAIError>
+    where
+        Q: Serialize + ?Sized,
+    {
+        self.client
+            .get_with_query(
+                format!("/fine_tuning/jobs/{fine_tuning_job_id}/checkpoints").as_str(),
+                query,
+            )
+            .await
+    }
 }

async-openai/src/lib.rs

@@ -79,6 +79,7 @@
 mod assistant_files;
 mod assistants;
 mod audio;
+mod batches;
 mod chat;
 mod client;
 mod completion;
@@ -98,10 +99,14 @@ mod steps;
 mod threads;
 pub mod types;
 mod util;
+mod vector_store_file_batches;
+mod vector_store_files;
+mod vector_stores;
 
 pub use assistant_files::AssistantFiles;
 pub use assistants::Assistants;
 pub use audio::Audio;
+pub use batches::Batches;
 pub use chat::Chat;
 pub use client::Client;
 pub use completion::Completions;
@@ -116,3 +121,6 @@ pub use moderation::Moderations;
 pub use runs::Runs;
 pub use steps::Steps;
 pub use threads::Threads;
+pub use vector_store_file_batches::VectorStoreFileBatches;
+pub use vector_store_files::VectorStoreFiles;
+pub use vector_stores::VectorStores;

async-openai/src/messages.rs

@@ -3,7 +3,10 @@ use serde::Serialize;
 use crate::{
     config::Config,
     error::OpenAIError,
-    types::{CreateMessageRequest, ListMessagesResponse, MessageObject, ModifyMessageRequest},
+    types::{
+        CreateMessageRequest, DeleteMessageResponse, ListMessagesResponse, MessageObject,
+        ModifyMessageRequest,
+    },
     Client, MessageFiles,
 };
 
@@ -70,4 +73,13 @@ impl<'c, C: Config> Messages<'c, C> {
             .get_with_query(&format!("/threads/{}/messages", self.thread_id), query)
             .await
     }
+
+    pub async fn delete(&self, message_id: &str) -> Result<DeleteMessageResponse, OpenAIError> {
+        self.client
+            .delete(&format!(
+                "/threads/{}/messages/{message_id}",
+                self.thread_id
+            ))
+            .await
+    }
 }

async-openai/src/moderation.rs

@@ -5,7 +5,7 @@ use crate::{
     Client,
 };
 
-/// Given a input text, outputs if the model classifies it as violating OpenAI's content policy.
+/// Given some input text, outputs if the model classifies it as potentially harmful across several categories.
 ///
 /// Related guide: [Moderations](https://platform.openai.com/docs/guides/moderation)
 pub struct Moderations<'c, C: Config> {
@@ -17,7 +17,7 @@ impl<'c, C: Config> Moderations<'c, C> {
         Self { client }
     }
 
-    /// Classifies if text violates OpenAI's Content Policy
+    /// Classifies if text is potentially harmful.
     pub async fn create(
         &self,
         request: CreateModerationRequest,