aha 0.2.5

aha model inference library, now supports Qwen(2.5VL/3/3VL/3.5/ASR/3Embedding/3Reranker), MiniCPM4, VoxCPM/1.5, DeepSeek-OCR/2, Hunyuan-OCR, PaddleOCR-VL/1.5, RMBG2.0, GLM(ASR-Nano-2512/OCR), Fun-ASR-Nano-2512, LFM(2/2.5/2VL/2.5VL)
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211

use serde::{Deserialize, Serialize};

#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
pub struct Usage {
    /// Number of tokens in the prompt.
    #[serde(skip_serializing_if = "Option::is_none")]
    pub prompt_tokens: Option<u32>,
    /// Number of tokens in the prompt.
    #[serde(skip_serializing_if = "Option::is_none")]
    pub prompt_secs: Option<f64>,
    /// Number of tokens in the completion.
    #[serde(skip_serializing_if = "Option::is_none")]
    pub completion_tokens: Option<u32>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub completion_secs: Option<f64>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub completion_per_token_secs: Option<f64>,
    #[serde(skip_serializing_if = "Option::is_none")]
    pub completion_tps: Option<f64>,
    /// Number of tokens in the entire response.
    pub total_tokens: u32,
    /// Breakdown of tokens used in the prompt.
    #[serde(skip_serializing_if = "Option::is_none")]
    pub prompt_tokens_details: Option<PromptTokensDetails>,
    /// Breakdown of tokens used in a completion.
    #[serde(skip_serializing_if = "Option::is_none")]
    pub completion_tokens_details: Option<CompletionTokensDetails>,
}

#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
pub struct InputTokensDetails {
    /// The number of tokens that were retrieved from the cache.
    pub cached_tokens: u32,
}

#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
pub struct OutputTokensDetails {
    /// The number of reasoning tokens.
    pub reasoning_tokens: u32,
}

#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
pub struct PromptTokensDetails {
    /// Audio input tokens present in the prompt.
    #[serde(skip_serializing_if = "Option::is_none")]
    pub audio_tokens: Option<u32>,
    /// Cached tokens present in the prompt.
    pub cached_tokens: u32,
}

#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
pub struct CompletionTokensDetails {
    /// Tokens generated by the model for reasoning.
    pub reasoning_tokens: u32,
    /// Audio input tokens generated by the model.
    #[serde(skip_serializing_if = "Option::is_none")]
    pub audio_tokens: Option<u32>,
    /// When using Predicted Outputs, the number of tokens in the prediction that appeared in the completion.
    #[serde(skip_serializing_if = "Option::is_none")]
    pub accepted_prediction_tokens: Option<u32>,
    /// When using Predicted Outputs, the number of tokens in the prediction that did not appear in the completion.
    #[serde(skip_serializing_if = "Option::is_none")]
    pub rejected_prediction_tokens: Option<u32>,
}

#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
pub struct ResponseWrapper<T> {
    pub data: T,
    pub headers: Headers,
}

#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
pub struct Headers {
    /// The maximum number of requests that are permitted before exhausting the rate limit.
    #[serde(rename = "x-ratelimit-limit-requests")]
    pub x_ratelimit_limit_requests: Option<u32>,
    /// The maximum number of tokens that are permitted before exhausting the rate limit.
    #[serde(rename = "x-ratelimit-limit-tokens")]
    pub x_ratelimit_limit_tokens: Option<u32>,
    /// The remaining number of requests that are permitted before exhausting the rate limit.
    #[serde(rename = "x-ratelimit-remaining-requests")]
    pub x_ratelimit_remaining_requests: Option<u32>,
    /// The remaining number of tokens that are permitted before exhausting the rate limit.
    #[serde(rename = "x-ratelimit-remaining-tokens")]
    pub x_ratelimit_remaining_tokens: Option<u32>,
    /// The time until the rate limit (based on requests) resets to its initial state.
    #[serde(rename = "x-ratelimit-reset-requests")]
    pub x_ratelimit_reset_requests: Option<String>,
    /// The time until the rate limit (based on tokens) resets to its initial state.
    #[serde(rename = "x-ratelimit-reset-tokens")]
    pub x_ratelimit_reset_tokens: Option<String>,
}

#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
pub struct SimpleListParameters {
    /// Identifier for the last object from the previous pagination request.
    pub after: Option<String>,
    /// Number of objects to retrieve.
    pub limit: Option<u32>,
}

#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
pub struct ListParameters {
    /// A limit on the number of objects to be returned. Limit can range between 1 and 100, and the default is 20.
    #[serde(skip_serializing_if = "Option::is_none")]
    pub limit: Option<u32>,
    /// Sort order by the created_at timestamp of the objects. asc for ascending order and desc for descending order.
    #[serde(skip_serializing_if = "Option::is_none")]
    pub order: Option<String>,
    /// A cursor for use in pagination. after is an object ID that defines your place in the list.
    #[serde(skip_serializing_if = "Option::is_none")]
    pub after: Option<String>,
    /// A cursor for use in pagination. before is an object ID that defines your place in the list.
    #[serde(skip_serializing_if = "Option::is_none")]
    pub before: Option<String>,
}

#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
pub struct ListResponse<T> {
    // The object type, which is always "list".
    pub object: String,
    /// The list ob objects.
    pub data: Vec<T>,
    /// The ID of the first objects in the list.
    pub first_id: Option<String>,
    /// The ID of the last objects in the list.
    pub last_id: Option<String>,
    /// Indicates whether there are more objects to retrieve.
    pub has_more: bool,
}

#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
pub struct DeletedObject {
    /// ID of the deleted object.
    pub id: String,
    /// The object type.
    pub object: String,
    /// Indicates whether the file was successfully deleted.
    pub deleted: bool,
}

#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
pub struct LastError {
    /// One of 'server_error' or 'rate_limit_exceeded'.
    pub code: LastErrorCode,
    /// A human-readable description of the error.
    pub message: String,
}

#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
#[serde(rename_all = "snake_case")]
pub enum LastErrorCode {
    ServerError,
    RateLimitExceeded,
}

#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
#[serde(rename_all = "snake_case")]
pub enum FinishReason {
    /// API returned complete message, or a message terminated by one of the stop sequences provided via the stop parameter.
    #[serde(rename = "stop", alias = "STOP")]
    StopSequenceReached,
    /// Incomplete model output due to max_tokens parameter or token limit.
    #[serde(rename = "length", alias = "MAX_TOKENS")]
    TokenLimitReached,
    /// Omitted content due to a flag from our content filters.
    #[serde(
        rename = "content_filter",
        alias = "SAFETY",
        alias = "SPII",
        alias = "PROHIBITED_CONTENT",
        alias = "BLOCKLIST",
        alias = "RECITATION"
    )]
    ContentFilterFlagged,
    /// The model decided to call one or more tools.
    ToolCalls,
    /// The model reached a natural stopping point. [Claude]
    EndTurn,
    /// The finish reason is unspecified. [Gemini]
    #[serde(rename = "FINISH_REASON_UNSPECIFIED	")]
    Unspecified,
    #[serde(rename = "MALFORMED_FUNCTION_CALL")]
    MalformedFunctionCall,
    #[serde(rename = "OTHER")]
    Other,
}

#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
#[serde(rename_all = "snake_case")]
pub enum ReasoningEffort {
    High,
    Medium,
    Low,
    Minimal,
}

#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
#[serde(rename_all = "snake_case")]
pub enum WebSearchContextSize {
    Low,
    Medium,
    Large,
}

#[derive(Serialize, Deserialize, Debug, Clone, PartialEq)]
#[serde(untagged)]
pub enum StopToken {
    String(String),
    Array(Vec<String>),
}