locus-sdk 0.1.2

SDK-first STTP memory primitives and AI provider abstraction
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

use std::collections::HashMap;

use serde::{Deserialize, Serialize};

use crate::domain::ai::{AiTask, EmbedRequest, ScoreAvecRequest};

#[derive(Debug, Clone, Default, Serialize, Deserialize, PartialEq, Eq)]
pub struct ProviderModelProfile {
    pub semantic_model: Option<String>,
    pub avec_embedding_model: Option<String>,
    pub avec_scoring_model: Option<String>,
}

#[derive(Debug, Clone, Default, Serialize, Deserialize)]
pub struct AiRoutingConfig {
    pub default_provider_id: Option<String>,
    pub providers: HashMap<String, ProviderModelProfile>,
}

impl AiRoutingConfig {
    pub fn profile_for(&self, provider_id: &str) -> Option<&ProviderModelProfile> {
        self.providers.get(provider_id)
    }

    pub fn provider_for(&self, requested_provider_id: Option<&str>) -> Option<String> {
        requested_provider_id
            .filter(|value| !value.trim().is_empty())
            .map(ToString::to_string)
            .or_else(|| self.default_provider_id.clone())
    }

    pub fn model_for(&self, provider_id: Option<&str>, task: AiTask) -> Option<&str> {
        let provider_id = self.provider_for(provider_id)?;
        let profile = self.providers.get(&provider_id)?;

        match task {
            AiTask::SemanticEmbedding => profile.semantic_model.as_deref(),
            AiTask::AvecEmbedding => profile.avec_embedding_model.as_deref(),
            AiTask::AvecScoring => profile.avec_scoring_model.as_deref(),
        }
    }

    pub fn apply_to_embed_request(&self, request: &EmbedRequest) -> EmbedRequest {
        let provider_id = request
            .provider_id
            .as_deref()
            .map(ToString::to_string)
            .or_else(|| self.default_provider_id.clone());
        let model = request.model.clone().or_else(|| {
            self.model_for(provider_id.as_deref(), request.task)
                .map(ToString::to_string)
        });

        EmbedRequest {
            text: request.text.clone(),
            task: request.task,
            provider_id,
            model,
            policy: request.policy,
        }
    }

    pub fn apply_to_score_request(&self, request: &ScoreAvecRequest) -> ScoreAvecRequest {
        let provider_id = request
            .provider_id
            .as_deref()
            .map(ToString::to_string)
            .or_else(|| self.default_provider_id.clone());
        let model = request.model.clone().or_else(|| {
            self.model_for(provider_id.as_deref(), AiTask::AvecScoring)
                .map(ToString::to_string)
        });

        ScoreAvecRequest {
            text: request.text.clone(),
            provider_id,
            model,
            policy: request.policy,
        }
    }
}

#[cfg(test)]
mod tests {
    use super::{AiRoutingConfig, ProviderModelProfile};
    use crate::domain::ai::{AiTask, EmbedRequest, ProviderPolicy, ScoreAvecRequest};

    fn fixture_config() -> AiRoutingConfig {
        let mut config = AiRoutingConfig {
            default_provider_id: Some("genai".to_string()),
            providers: std::collections::HashMap::new(),
        };

        config.providers.insert(
            "genai".to_string(),
            ProviderModelProfile {
                semantic_model: Some("text-embedding-3-small".to_string()),
                avec_embedding_model: Some("text-embedding-3-large".to_string()),
                avec_scoring_model: Some("gpt-4o-mini".to_string()),
            },
        );

        config
    }

    #[test]
    fn apply_to_embed_request_fills_default_provider_and_model() {
        let config = fixture_config();
        let request = EmbedRequest {
            text: "hello".to_string(),
            task: AiTask::SemanticEmbedding,
            provider_id: None,
            model: None,
            policy: ProviderPolicy::Auto,
        };

        let resolved = config.apply_to_embed_request(&request);

        assert_eq!(resolved.provider_id.as_deref(), Some("genai"));
        assert_eq!(
            resolved.model.as_deref(),
            Some("text-embedding-3-small")
        );
    }

    #[test]
    fn apply_to_embed_request_keeps_explicit_model() {
        let config = fixture_config();
        let request = EmbedRequest {
            text: "hello".to_string(),
            task: AiTask::AvecEmbedding,
            provider_id: Some("genai".to_string()),
            model: Some("my-custom-model".to_string()),
            policy: ProviderPolicy::Preferred,
        };

        let resolved = config.apply_to_embed_request(&request);

        assert_eq!(resolved.model.as_deref(), Some("my-custom-model"));
    }

    #[test]
    fn apply_to_score_request_resolves_scoring_model() {
        let config = fixture_config();
        let request = ScoreAvecRequest {
            text: "score this".to_string(),
            provider_id: None,
            model: None,
            policy: ProviderPolicy::Auto,
        };

        let resolved = config.apply_to_score_request(&request);

        assert_eq!(resolved.provider_id.as_deref(), Some("genai"));
        assert_eq!(resolved.model.as_deref(), Some("gpt-4o-mini"));
    }
}