async_openai/types/
assistant.rs

1use std::collections::HashMap;
2
3use derive_builder::Builder;
4use serde::{Deserialize, Serialize};
5
6use crate::error::OpenAIError;
7
8use super::{FunctionName, FunctionObject, ResponseFormat};
9
10#[derive(Clone, Serialize, Debug, Deserialize, PartialEq, Default)]
11pub struct AssistantToolCodeInterpreterResources {
12    ///A list of [file](https://platform.openai.com/docs/api-reference/files) IDs made available to the `code_interpreter`` tool. There can be a maximum of 20 files associated with the tool.
13    pub file_ids: Vec<String>, // maxItems: 20
14}
15
16#[derive(Clone, Serialize, Debug, Deserialize, PartialEq, Default)]
17pub struct AssistantToolFileSearchResources {
18    /// The ID of the [vector store](https://platform.openai.com/docs/api-reference/vector-stores/object) attached to this assistant. There can be a maximum of 1 vector store attached to the assistant.
19    pub vector_store_ids: Vec<String>,
20}
21
22#[derive(Clone, Serialize, Debug, Deserialize, PartialEq)]
23pub struct AssistantToolResources {
24    #[serde(skip_serializing_if = "Option::is_none")]
25    pub code_interpreter: Option<AssistantToolCodeInterpreterResources>,
26    #[serde(skip_serializing_if = "Option::is_none")]
27    pub file_search: Option<AssistantToolFileSearchResources>,
28}
29
30#[derive(Clone, Serialize, Debug, Deserialize, PartialEq)]
31pub struct CreateAssistantToolResources {
32    #[serde(skip_serializing_if = "Option::is_none")]
33    pub code_interpreter: Option<AssistantToolCodeInterpreterResources>,
34    #[serde(skip_serializing_if = "Option::is_none")]
35    pub file_search: Option<CreateAssistantToolFileSearchResources>,
36}
37
38#[derive(Clone, Serialize, Debug, Deserialize, PartialEq, Default)]
39pub struct CreateAssistantToolFileSearchResources {
40    ///  The [vector store](https://platform.openai.com/docs/api-reference/vector-stores/object) attached to this assistant. There can be a maximum of 1 vector store attached to the assistant.
41    pub vector_store_ids: Option<Vec<String>>,
42    /// A helper to create a [vector store](https://platform.openai.com/docs/api-reference/vector-stores/object) with file_ids and attach it to this assistant. There can be a maximum of 1 vector store attached to the assistant.
43    pub vector_stores: Option<Vec<AssistantVectorStore>>,
44}
45
46#[derive(Clone, Serialize, Debug, Deserialize, PartialEq, Default)]
47pub struct AssistantVectorStore {
48    /// A list of [file](https://platform.openai.com/docs/api-reference/files) IDs to add to the vector store. There can be a maximum of 10000 files in a vector store.
49    pub file_ids: Vec<String>,
50
51    /// The chunking strategy used to chunk the file(s). If not set, will use the `auto` strategy.
52    pub chunking_strategy: Option<AssistantVectorStoreChunkingStrategy>,
53
54    /// Set of 16 key-value pairs that can be attached to a vector store. This can be useful for storing additional information about the vector store in a structured format. Keys can be a maximum of 64 characters long and values can be a maxium of 512 characters long.
55    pub metadata: Option<HashMap<String, String>>,
56}
57
58#[derive(Clone, Serialize, Debug, Deserialize, PartialEq, Default)]
59#[serde(tag = "type")]
60pub enum AssistantVectorStoreChunkingStrategy {
61    /// The default strategy. This strategy currently uses a `max_chunk_size_tokens` of `800` and `chunk_overlap_tokens` of `400`.
62    #[default]
63    #[serde(rename = "auto")]
64    Auto,
65    #[serde(rename = "static")]
66    Static { r#static: StaticChunkingStrategy },
67}
68
69/// Static Chunking Strategy
70#[derive(Clone, Serialize, Debug, Deserialize, PartialEq, Default)]
71pub struct StaticChunkingStrategy {
72    /// The maximum number of tokens in each chunk. The default value is `800`. The minimum value is `100` and the maximum value is `4096`.
73    pub max_chunk_size_tokens: u16,
74    /// The number of tokens that overlap between chunks. The default value is `400`.
75    ///
76    /// Note that the overlap must not exceed half of `max_chunk_size_tokens`.
77    pub chunk_overlap_tokens: u16,
78}
79
80/// Represents an `assistant` that can call the model and use tools.
81#[derive(Clone, Serialize, Debug, Deserialize, PartialEq)]
82pub struct AssistantObject {
83    /// The identifier, which can be referenced in API endpoints.
84    pub id: String,
85    /// The object type, which is always `assistant`.
86    pub object: String,
87    /// The Unix timestamp (in seconds) for when the assistant was created.
88    pub created_at: i32,
89    /// The name of the assistant. The maximum length is 256 characters.
90    pub name: Option<String>,
91    /// The description of the assistant. The maximum length is 512 characters.
92    pub description: Option<String>,
93    /// ID of the model to use. You can use the [List models](https://platform.openai.com/docs/api-reference/models/list) API to see all of your available models, or see our [Model overview](https://platform.openai.com/docs/models) for descriptions of them.
94    pub model: String,
95    /// The system instructions that the assistant uses. The maximum length is 256,000 characters.
96    pub instructions: Option<String>,
97    /// A list of tool enabled on the assistant. There can be a maximum of 128 tools per assistant. Tools can be of types `code_interpreter`, `file_search`, or `function`.
98    #[serde(default)]
99    pub tools: Vec<AssistantTools>,
100    /// A set of resources that are used by the assistant's tools. The resources are specific to the type of tool. For example, the `code_interpreter` tool requires a list of file IDs, while the `file_search` tool requires a list of vector store IDs.
101    pub tool_resources: Option<AssistantToolResources>,
102    /// Set of 16 key-value pairs that can be attached to an object. This can be useful for storing additional information about the object in a structured format. Keys can be a maximum of 64 characters long and values can be a maximum of 512 characters long.
103    pub metadata: Option<HashMap<String, String>>,
104    /// 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.
105    pub temperature: Option<f32>,
106    /// 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.
107    /// We generally recommend altering this or temperature but not both.
108    pub top_p: Option<f32>,
109
110    pub response_format: Option<AssistantsApiResponseFormatOption>,
111}
112
113/// Specifies the format that the model must output. Compatible with [GPT-4o](https://platform.openai.com/docs/models/gpt-4o), [GPT-4 Turbo](https://platform.openai.com/docs/models/gpt-4-turbo-and-gpt-4), and all GPT-3.5 Turbo models since `gpt-3.5-turbo-1106`.
114///
115/// Setting to `{ "type": "json_schema", "json_schema": {...} }` enables Structured Outputs which guarantees the model will match your supplied JSON schema. Learn more in the [Structured Outputs guide](https://platform.openai.com/docs/guides/structured-outputs).
116///
117/// Setting to `{ "type": "json_object" }` enables JSON mode, which guarantees the message the model generates is valid JSON.
118///
119/// **Important:** when using JSON mode, you **must** also instruct the model to produce JSON yourself via a system or user message. Without this, the model may generate an unending stream of whitespace until the generation reaches the token limit, resulting in a long-running and seemingly "stuck" request. Also note that the message content may be partially cut off if `finish_reason="length"`, which indicates the generation exceeded `max_tokens` or the conversation exceeded the max context length.
120#[derive(Clone, Serialize, Debug, Deserialize, PartialEq, Default)]
121pub enum AssistantsApiResponseFormatOption {
122    #[default]
123    #[serde(rename = "auto")]
124    Auto,
125    #[serde(untagged)]
126    Format(ResponseFormat),
127}
128
129/// Retrieval tool
130#[derive(Clone, Serialize, Debug, Default, Deserialize, PartialEq)]
131pub struct AssistantToolsFileSearch {
132    /// Overrides for the file search tool.
133    #[serde(skip_serializing_if = "Option::is_none")]
134    pub file_search: Option<AssistantToolsFileSearchOverrides>,
135}
136
137#[derive(Clone, Serialize, Debug, Deserialize, PartialEq)]
138pub struct AssistantToolsFileSearchOverrides {
139    ///  The maximum number of results the file search tool should output. The default is 20 for gpt-4* models and 5 for gpt-3.5-turbo. This number should be between 1 and 50 inclusive.
140    ///
141    //// Note that the file search tool may output fewer than `max_num_results` results. See the [file search tool documentation](https://platform.openai.com/docs/assistants/tools/file-search/customizing-file-search-settings) for more information.
142    pub max_num_results: Option<u8>,
143    pub ranking_options: Option<FileSearchRankingOptions>,
144}
145
146#[derive(Clone, Serialize, Debug, Deserialize, PartialEq)]
147pub enum FileSearchRanker {
148    #[serde(rename = "auto")]
149    Auto,
150    #[serde(rename = "default_2024_08_21")]
151    Default2024_08_21,
152}
153
154/// The ranking options for the file search. If not specified, the file search tool will use the `auto` ranker and a score_threshold of 0.
155///
156/// See the [file search tool documentation](https://platform.openai.com/docs/assistants/tools/file-search#customizing-file-search-settings) for more information.
157#[derive(Debug, Clone, Serialize, Deserialize, PartialEq)]
158pub struct FileSearchRankingOptions {
159    /// The ranker to use for the file search. If not specified will use the `auto` ranker.
160    #[serde(skip_serializing_if = "Option::is_none")]
161    pub ranker: Option<FileSearchRanker>,
162
163    /// The score threshold for the file search. All values must be a floating point number between 0 and 1.
164    pub score_threshold: f32,
165}
166
167/// Function tool
168#[derive(Clone, Serialize, Debug, Default, Deserialize, PartialEq)]
169pub struct AssistantToolsFunction {
170    pub function: FunctionObject,
171}
172
173#[derive(Clone, Serialize, Debug, Deserialize, PartialEq)]
174#[serde(tag = "type")]
175#[serde(rename_all = "snake_case")]
176pub enum AssistantTools {
177    CodeInterpreter,
178    FileSearch(AssistantToolsFileSearch),
179    Function(AssistantToolsFunction),
180}
181
182#[derive(Clone, Serialize, Default, Debug, Deserialize, Builder, PartialEq)]
183#[builder(name = "CreateAssistantRequestArgs")]
184#[builder(pattern = "mutable")]
185#[builder(setter(into, strip_option), default)]
186#[builder(derive(Debug))]
187#[builder(build_fn(error = "OpenAIError"))]
188pub struct CreateAssistantRequest {
189    /// ID of the model to use. You can use the [List models](https://platform.openai.com/docs/api-reference/models/list) API to see all of your available models, or see our [Model overview](https://platform.openai.com/docs/models/overview) for descriptions of them.
190    pub model: String,
191
192    /// The name of the assistant. The maximum length is 256 characters.
193    #[serde(skip_serializing_if = "Option::is_none")]
194    pub name: Option<String>,
195
196    /// The description of the assistant. The maximum length is 512 characters.
197    #[serde(skip_serializing_if = "Option::is_none")]
198    pub description: Option<String>,
199
200    /// The system instructions that the assistant uses. The maximum length is 256,000 characters.
201    #[serde(skip_serializing_if = "Option::is_none")]
202    pub instructions: Option<String>,
203
204    /// A list of tool enabled on the assistant. There can be a maximum of 128 tools per assistant. Tools can be of types `code_interpreter`, `file_search`, or `function`.
205    #[serde(skip_serializing_if = "Option::is_none")]
206    pub tools: Option<Vec<AssistantTools>>,
207
208    /// A set of resources that are used by the assistant's tools. The resources are specific to the type of tool. For example, the `code_interpreter` tool requires a list of file IDs, while the `file_search` tool requires a list of vector store IDs.
209    #[serde(skip_serializing_if = "Option::is_none")]
210    pub tool_resources: Option<CreateAssistantToolResources>,
211
212    /// Set of 16 key-value pairs that can be attached to an object. This can be useful for storing additional information about the object in a structured format. Keys can be a maximum of 64 characters long and values can be a maximum of 512 characters long.
213    #[serde(skip_serializing_if = "Option::is_none")]
214    pub metadata: Option<HashMap<String, String>>,
215
216    /// 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.
217    #[serde(skip_serializing_if = "Option::is_none")]
218    pub temperature: Option<f32>,
219
220    /// 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.
221    ///
222    /// We generally recommend altering this or temperature but not both.
223    #[serde(skip_serializing_if = "Option::is_none")]
224    pub top_p: Option<f32>,
225
226    #[serde(skip_serializing_if = "Option::is_none")]
227    pub response_format: Option<AssistantsApiResponseFormatOption>,
228}
229
230#[derive(Clone, Serialize, Default, Debug, Deserialize, Builder, PartialEq)]
231#[builder(name = "ModifyAssistantRequestArgs")]
232#[builder(pattern = "mutable")]
233#[builder(setter(into, strip_option), default)]
234#[builder(derive(Debug))]
235#[builder(build_fn(error = "OpenAIError"))]
236pub struct ModifyAssistantRequest {
237    /// ID of the model to use. You can use the [List models](https://platform.openai.com/docs/api-reference/models/list) API to see all of your available models, or see our [Model overview](https://platform.openai.com/docs/models/overview) for descriptions of them.
238    #[serde(skip_serializing_if = "Option::is_none")]
239    pub model: Option<String>,
240
241    /// The name of the assistant. The maximum length is 256 characters.
242    #[serde(skip_serializing_if = "Option::is_none")]
243    pub name: Option<String>,
244
245    /// The description of the assistant. The maximum length is 512 characters.
246    #[serde(skip_serializing_if = "Option::is_none")]
247    pub description: Option<String>,
248
249    /// The system instructions that the assistant uses. The maximum length is 256,000 characters.
250    #[serde(skip_serializing_if = "Option::is_none")]
251    pub instructions: Option<String>,
252
253    /// A list of tool enabled on the assistant. There can be a maximum of 128 tools per assistant. Tools can be of types `code_interpreter`, `file_search`, or `function`.
254    #[serde(skip_serializing_if = "Option::is_none")]
255    pub tools: Option<Vec<AssistantTools>>,
256
257    /// A set of resources that are used by the assistant's tools. The resources are specific to the type of tool. For example, the `code_interpreter` tool requires a list of file IDs, while the `file_search` tool requires a list of vector store IDs.
258    #[serde(skip_serializing_if = "Option::is_none")]
259    pub tool_resources: Option<AssistantToolResources>,
260    /// Set of 16 key-value pairs that can be attached to an object. This can be useful for storing additional information about the object in a structured format. Keys can be a maximum of 64 characters long and values can be a maxium of 512 characters long.
261    #[serde(skip_serializing_if = "Option::is_none")]
262    pub metadata: Option<HashMap<String, String>>,
263
264    /// 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.
265    #[serde(skip_serializing_if = "Option::is_none")]
266    pub temperature: Option<f32>,
267
268    /// 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.
269    ///
270    /// We generally recommend altering this or temperature but not both.
271    #[serde(skip_serializing_if = "Option::is_none")]
272    pub top_p: Option<f32>,
273
274    #[serde(skip_serializing_if = "Option::is_none")]
275    pub response_format: Option<AssistantsApiResponseFormatOption>,
276}
277
278#[derive(Clone, Serialize, Default, Debug, Deserialize, PartialEq)]
279pub struct DeleteAssistantResponse {
280    pub id: String,
281    pub deleted: bool,
282    pub object: String,
283}
284
285#[derive(Clone, Serialize, Default, Debug, Deserialize, PartialEq)]
286pub struct ListAssistantsResponse {
287    pub object: String,
288    pub data: Vec<AssistantObject>,
289    pub first_id: Option<String>,
290    pub last_id: Option<String>,
291    pub has_more: bool,
292}
293
294/// Controls which (if any) tool is called by the model.
295/// `none` means the model will not call any tools and instead generates a message.
296/// `auto` is the default value and means the model can pick between generating a message or calling one or more tools.
297/// `required` means the model must call one or more tools before responding to the user.
298/// Specifying a particular tool like `{"type": "file_search"}` or `{"type": "function", "function": {"name": "my_function"}}` forces the model to call that tool.
299#[derive(Clone, Serialize, Default, Debug, Deserialize, PartialEq)]
300#[serde(rename_all = "lowercase")]
301pub enum AssistantsApiToolChoiceOption {
302    #[default]
303    None,
304    Auto,
305    Required,
306    #[serde(untagged)]
307    Named(AssistantsNamedToolChoice),
308}
309
310/// Specifies a tool the model should use. Use to force the model to call a specific tool.
311#[derive(Clone, Serialize, Default, Debug, Deserialize, PartialEq)]
312pub struct AssistantsNamedToolChoice {
313    /// The type of the tool. If type is `function`, the function name must be set
314    pub r#type: AssistantToolType,
315
316    pub function: Option<FunctionName>,
317}
318
319#[derive(Clone, Serialize, Default, Debug, Deserialize, PartialEq)]
320#[serde(rename_all = "snake_case")]
321pub enum AssistantToolType {
322    #[default]
323    Function,
324    CodeInterpreter,
325    FileSearch,
326}
async_openai/types/assistant.rs

async_openai/types/
assistant.rs
